vvEPA
United States
Environmental Protection
Agency
EPA-600-R-16-236Fa
December 2016
www.epa.gov/hfstudy
Hydraulic Fracturing for Oil and Gas:
Impacts from the Hydraulic Fracturing
Water Cycle on Drinking Water
Resources in the United States
Office of Research and Development
Washington, DC
-------
-------
EPA-600-R-16-236Fa
December 2016
www. epa. gov /hfstudy
Hydraulic Fracturing for Oil and Gas:
Impacts from the Hydraulic
Fracturing Water Cycle on Drinking
Water Resources in the United States
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
-------
This page is intentionally left blank.
-------
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: USEPA (U.S. Environmental Protection Agency). 2016. Hydraulic Fracturing for
Oil and Gas: Impacts from the Hydraulic Fracturing Water Cycle on Drinking Water Resources in
the United States. Office of Research and Development, Washington, DC. EPA-600-R-16-236Fa.
i
-------
Contents
List of Tables ix
List of Figures xii
List of Text Boxes xvii
List of Acronyms/Abbreviations xix
Preface xxiv
Authors, Contributors, and Reviewers xxv
Acknowledgements xxxi
Executive Summary ES-1
Drinking Water Resources in the United States ES-4
Hydraulic Fracturing for Oil and Gas in the United States ES-5
Approach: The Hydraulic Fracturing Water Cycle ES-9
Water Acquisition ES-12
Water Acquisition Conclusions ES-18
Chemical Mixing ES-18
Chemical Mixing Conclusions ES-26
Well Injection ES-26
Well Injection Conclusions ES-32
Produced Water Handling ES-33
Produced Water Handling Conclusions ES-37
Wastewater Disposal and Reuse ES-38
Wastewater Disposal and Reuse Conclusions ES-42
Chemicals in the Hydraulic Fracturing Water Cycle ES-42
Chemicals in the Hydraulic Fracturing Water Cycle Conclusions ES-44
Data Gaps and Uncertainties ES-44
Report Conclusions ES-46
Chapter 1. Introduction 1-1
1.1 Background 1-3
1.2 Goals 1-4
1.3 Scope 1-4
1.4 Approach 1-9
1.4.1 EPA Hydraulic Fracturing Study Publications 1-9
1.4.2 Literature and Data Search Strategy 1-10
1.4.3 Literature and Data Evaluation Strategy 1-11
1.4.4 Quality Assurance and Peer Review 1-11
1.5 Organization 1-12
1.6 Intended Use 1-13
ii
-------
Chapter 2. Drinking Water Resources in the United States 2-1
Abstract 2-1
2.1 Introduction 2-3
2.2 Ground and Surface Water Resources 2-3
2.2.1 Groundwater Resources 2-5
2.2.2 Surface Water Resources 2-7
2.3 Current Drinking Water Sources 2-9
2.3.1 Factors Affecting How Water Becomes a Drinking Water Source 2-11
2.4 Future Drinking Water Sources 2-13
2.5 Proximity of Drinking Water Resources to Hydraulic Fracturing Operations 2-14
2.5.1 Lateral Distance between Public Water System Sources and Hydraulic Fracturing 2-14
2.5.2 Vertical Distance between Drinking Water Resources and Hydraulic Fracturing 2-16
2.6 Conclusions 2-18
Chapter 3. Hydraulic Fracturing for Oil and Gas in the United States 3-1
Abstract 3-1
3.1 Introduction 3-3
3.2 What is Hydraulic Fracturing? 3-3
3.3 Hydraulic Fracturing and the Life of a Well 3-11
3.3.1 Site Preparation and Well Construction 3-12
3.3.2 Hydraulic Fracturing 3-18
3.3.3 Fluid Recovery, Handling, and Disposal or Reuse 3-23
3.3.4 Oil and Gas Production 3-24
3.3.5 Site and Well Closure 3-25
3.4 How Widespread is Hydraulic Fracturing? 3-25
3.4.1 Number of Wells Fractured per Year 3-29
3.4.2 Hydraulic Fracturing Rates 3-31
3.5 Trends and Outlook for the Future 3-32
3.5.1 Natural Gas 3-33
3.5.2 Oil 3-36
3.6 Conclusions 3-38
Chapter 4. Water Acquisition 4-1
Abstract 4-1
4.1 Introduction 4-3
4.2 Types of Water Used 4-5
4.2.1 Source 4-5
4.2.2 Quality 4-8
4.2.3 Provisioning 4-10
4.3 Water Use Per Well 4-10
4.3.1 Hydraulic Fracturing Water Use in the Life Cycle of Oil and Gas 4-10
4.3.2 National Estimates and Variability in Water Use Per Well for Hydraulic Fracturing 4-11
iii
-------
4.4 Hydraulic Fracturing Water Use and Consumption at the National, State, and County
Scale 4-13
4.4.1 National and State Scale 4-13
4.4.2 County Scale 4-15
4.5 Potential for Impacts by Location 4-21
4.5.1 Texas 4-21
4.5.2 Colorado and Wyoming 4-31
4.5.3 Pennsylvania, West Virginia, and Ohio 4-35
4.5.4 North Dakota and Montana 4-40
4.5.5 Arkansas and Louisiana 4-42
4.6 Chapter Synthesis 4-45
4.6.1 Major Findings 4-45
4.6.2 Factors Affecting Frequency or Severity of Impacts 4-47
4.6.3 Uncertainties 4-49
4.6.4 Conclusions 4-50
Chapter 5. Chemical Mixing 5-1
Abstract 5-1
5.1 Introduction 5-3
5.2 Chemical Mixing Process 5-4
5.3 Overview of Hydraulic Fracturing Fluids 5-8
5.3.1 Water-Based Fracturing Fluids 5-12
5.3.2 Alternative Fracturing Fluids 5-13
5.3.3 Tracers 5-14
5.3.4 Proppants 5-16
5.3.5 Example Hydraulic Fracturing Fluids 5-16
5.4 Frequency and Volume of Hydraulic Fracturing Chemical Use 5-17
5.4.1 National Frequency of Use of Hydraulic Fracturing Chemicals 5-20
5.4.2 Nationwide Oil versus Gas 5-24
5.4.3 State-by-State Frequency of Use of Hydraulic Fracturing Chemicals 5-25
5.4.4 Volume of Chemical Use 5-26
5.4.5 Chemical Composition of Hydraulic Fracturing Fluids and Additives 5-28
5.5 Chemical Management and Spill Potential 5-31
5.5.1 Storage 5-33
5.5.2 Hoses and Lines 5-37
5.5.3 Blender 5-38
5.5.4 Manifold 5-39
5.5.5 High-Pressure Fracturing Pumps 5-39
5.5.6 Surface Wellhead for Fracture Stimulation 5-39
5.6 Overview of Chemical Spills Data 5-41
5.6.1 EPA Analysis of Spills Associated with Hydraulic Fracturing 5-41
5.6.2 Estimated Spill Rate and Other Spill Reports and Data 5-45
5.7 Spill Prevention, Containment, and Mitigation 5-46
iv
-------
5.8 Fate and Transport of Spilled Chemicals 5-47
5.8.1 Potential Paths 5-49
5.8.2 Physicochemical Properties of Organic Hydraulic Fracturing Chemicals 5-50
5.8.3 Mobility of Organic Hydraulic Fracturing Chemicals 5-52
5.8.4 Transformation Processes 5-56
5.8.5 Fate and Transport of Chemical Mixtures 5-57
5.8.6 Site and Environmental Conditions 5-58
5.8.7 Peer-Reviewed Literature on the Fate and Transport of Hydraulic Fracturing
Fluid Spills 5-58
5.8.8 Potential and Documented Fate and Transport of Documented Spills 5-59
5.8.9 Challenges with Unmonitored and Undetected Chemicals 5-62
5.9 Trends in the Use of Hydraulic Fracturing Chemicals 5-63
5.10 Synthesis 5-64
5.10.1 Summary of Findings 5-64
5.10.2 Factors Affecting the Frequency or Severity of Impacts 5-66
5.10.3 Uncertainties 5-67
5.10.4 Conclusions 5-69
Chapter 6. Well Injection 6-1
Abstract 6-1
6.1 Introduction 6-3
6.2 Fluid Migration Pathways Within and Along the Production Well 6-5
6.2.1 Overview of Well Construction 6-5
6.2.2 Factors that can Affect Fluid Movement to Drinking Water Resources 6-16
6.3 Fluid Migration Associated with Induced Fractures within Subsurface Formations 6-38
6.3.1 Overview of Subsurface Fracture Growth 6-40
6.3.2 Migration of Fluids through Pathways Related to Fractures/Formations 6-44
6.4 Synthesis 6-69
6.4.1 Summary of Findings 6-70
6.4.2 Factors Affecting Frequency or Severity of Impacts 6-73
6.4.3 Uncertainties 6-75
6.4.4 Conclusions 6-77
Chapter 7. Produced Water Handling 7-1
Abstract 7-1
7.1 Introduction 7-3
7.1.1 Definitions 7-4
7.2 Volume of Hydraulic Fracturing Flowback and Produced Water 7-5
7.2.1 Flowback of Injected Hydraulic Fracturing Fluid 7-6
7.2.2 Produced Water Volumes 7-8
7.3 Chemical Composition of Produced Water 7-11
7.3.1 Determination of Produced Water Composition 7-11
7.3.2 Factors Influencing Produced Water Composition 7-12
7.3.3 Produced Water Composition During the Flowback Period 7-12
v
-------
7.3.4 Produced Water Composition 7-16
7.3.5 Spatial Trends in Produced Water Composition 7-24
7.4 Spill and Release Impacts on Drinking Water Resources 7-25
7.4.1 Produced Water Handling and Spill Potential 7-25
7.4.2 Spills of Produced Water 7-26
7.5 Roadway Transport of Produced Water 7-40
7.6 Synthesis 7-41
7.6.1 Summary of Findings 7-41
7.6.2 Factors Affecting the Frequency or Severity of Impacts 7-43
7.6.3 Uncertainties 7-43
7.6.4 Conclusions 7-44
Chapter 8. Wastewater Disposal and Reuse 8-1
Abstract 8-1
8.1 Introduction 8-3
8.2 Volumes of Hydraulic Fracturing Wastewater 8-4
8.2.1 National Level Estimate 8-6
8.2.2 Regional/State Level Estimates 8-6
8.2.3 Estimation Methodologies and Challenges 8-8
8.3 Wastewater Characteristics 8-11
8.3.1 Wastewater 8-11
8.3.2 Constituents in Residuals 8-13
8.4 Wastewater Management Practices and Their Potential Impacts on Drinking Water
Resources 8-14
8.4.1 Underground Injection 8-23
8.4.2 Publicly Owned Treatment Works 8-27
8.4.3 Centralized Waste Treatment Facilities 8-28
8.4.4 Wastewater Reuse for Hydraulic Fracturing 8-35
8.4.5 Storage and Disposal Pits and Impoundments 8-39
8.4.6 Other Management Practices and Issues 8-46
8.4.7 Management of Solid and Liquid Residuals 8-51
8.5 Potential Impacts of Hydraulic Fracturing Wastewater Constituents on Drinking Water
Resources 8-54
8.5.1 Bromide, Iodide, and Chloride 8-54
8.5.2 Radionuclides 8-58
8.5.3 Metals 8-64
8.5.4 Volatile Organic Compounds 8-65
8.5.5 Semi-Volatile Organic Compounds 8-65
8.5.6 Oil and Grease 8-66
8.6 Synthesis 8-66
8.6.1 Summary of Findings 8-66
8.6.2 Factors Affecting the Frequency or Severity of Impacts 8-70
8.6.3 Uncertainties 8-73
vi
-------
8.6.4 Conclusions
8-75
Chapter 9. Identification and Hazard Evaluation of Chemicals across the Hydraulic
Fracturing Water Cycle 9-1
Abstract 9-1
9.1 Introduction 9-3
9.2 Overview: Hydraulic Fracturing and Potential Impacts on Human Health 9-5
9.3 Identification of Chemicals Associated with the Hydraulic Fracturing Water Cycle 9-8
9.3.1 Chemicals Used in Hydraulic Fracturing Fluids 9-9
9.3.2 Chemicals Detected in Produced Water 9-11
9.4 Toxicological and Physicochemical Properties of Hydraulic Fracturing Chemicals 9-11
9.4.1 Reference Values (RfVs), Oral Slope Factors (OSFs), and Qualitative Cancer
Classifications 9-13
9.4.2 Estimating Toxicity Using Quantitative Structure Activity Relationship (QSAR)
Modeling 9-17
9.4.3 Chemical Data Available from EPA's Aggregated Computations Toxicology Resource
(ACToR) Database 9-19
9.4.4 Additional Tools for Hazard Evaluation 9-20
9.4.5 Physicochemical Properties 9-21
9.4.6 Summary of Available Toxicological and Physicochemical Information for Hydraulic
Fracturing Chemicals 9-21
9.5 Hazard Identification of Hydraulic Fracturing Chemicals 9-22
9.5.1 Chemicals Used in Hydraulic Fracturing Fluids 9-23
9.5.2 Organic Chemicals in Produced Water 9-30
9.5.3 Inorganic Chemicals and TENORM in Produced Water 9-38
9.5.4 Organochlorine Pesticides and Polychlorinated Biphenyls (PCBs) in Produced Water .9-44
9.5.5 Methane in Stray Gas 9-46
9.5.6 Disinfection Byproducts (DBPs) Formed from Wastewater Constituents 9-47
9.5.7 Chemicals Detected in Multiple Stages of the Hydraulic Fracturing Water Cycle 9-48
9.6 Hazard Evaluation of Selected Subsets of Hydraulic Fracturing Chemicals Using Multi-Criteria
Decision Analysis (MCDA): Integrating Toxicity, Occurrence, and Physicochemical
Data 9-51
9.6.1 Overview of the MCDA Framework for Hazard Evaluation 9-53
9.6.2 Selection of Chemicals for Hazard Evaluation in the MCDA Framework 9-53
9.6.3 Calculation of MCDA Scores 9-57
9.6.4 Total Hazard Potential Score 9-58
9.6.5 MCDA Results 9-59
9.6.6 Limitations and Uncertainty of the MCDA Framework 9-77
9.6.7 Application of the MCDA Framework for Preliminary Hazard Evaluation 9-78
9.7 Synthesis 9-79
9.7.1 Summary of Findings 9-79
9.7.2 Factors Affecting the Frequency or Severity of Impacts 9-81
9.7.3 Uncertainties 9-82
9.7.4 Conclusions 9-83
vii
-------
9.8 Annex 9-84
9.8.1 Calculation of Physicochemical Property Scores (MCDA Hazard Evaluation) 9-84
9.8.2 Example of MCDA Score Calculation 9-85
Chapter 10. Synthesis 10-1
Introduction 10-3
10.1 Factors Affecting the Frequency or Severity of Impacts 10-4
10.1.1 Water Acquisition 10-4
10.1.2 Chemical Mixing and Produced Water Handling 10-8
10.1.3 Well Injection 10-13
10.1.4 Wastewater Disposal and Reuse 10-21
10.1.5 Summary 10-23
10.2 Uncertainties and Data Gaps 10-24
10.3 Use of this Assessment 10-28
Chapter 11. References 11-1
viii
-------
List of Tables
Table ES-1. Water use per hydraulically fractured well between January 2011 and February 2013 ES-13
Table ES-2. Chemicals reported in 10% or more of disclosures in FracFocus 1.0 ES-20
Table ES-3. Available chronic oral reference values for hydraulic fracturing chemicals reported
in 10% or more of disclosures in FracFocus 1.0 ES-43
Table 1-1. The five factors and accompanying criteria used to evaluate literature and data cited
in this assessment 1-11
Table 2-1. Summary of drinking water sources in the United States in 2010 2-10
Table 3-1. Estimated number of new wells hydraulically fractured nationally by year from
various sources 3-29
Table 4-1. Estimated proportions of hydraulic fracturing source water from surface water and
groundwater 4-6
Table 4-2. Percentage of injected water volume that comes from reused hydraulic fracturing
wastewater in various states, basins, and plays 4-7
Table 4-3. Average annual hydraulic fracturing water use and consumption in 2011 and 2012
compared to total annual water use and consumption in 2010, by county 4-16
Table 4-4. Estimated brackish water use as a percentage of total hydraulic fracturing water use
in the main hydraulic fracturing areas of Texas, 2011 4-23
Table 5-1. Examples of common additives, their function, and the most frequently used
chemicals reported to FracFocus for these additives 5-11
Table 5-2. Classes and specifically identified examples of tracers used in hydraulic fracturing
fluids 5-14
Table 5-3. Chemicals identified in the EPA FracFocus 1.0 project database in 10% or more
disclosures, with the percent of disclosures for which each chemical is reported as an ingredient
in an additive and the top four reported additives for which the chemical is used 5-21
Table 5-4. Example list of chemicals and chemical volumes used in hydraulic fracturing 5-26
Table 5-5. Fluid and additive composition by maximum mass percent 5-29
Table 5-6. Examples of typical hydraulic fracturing equipment and its function 5-32
Table 5-7. The 20 chemicals reported most frequently nationwide for hydraulic fracturing based
on the EPA FracFocus 1.0 project database, with EPI Suite™ physicochemical parameters where
available, and estimated mean and median volumes of those chemicals where density was
available 5-54
Table 6-1. Failure rates of vertical wells in the Wattenberg field, Colorado 6-20
Table 6-2. Results of studies of PA DEP violation data that examined mechanical integrity failure
rates 6-30
Table 6-3. Comparing the approximate depth and thickness of selected U.S. shale gas plays and
coalbed methane basins 6-46
Table 6-4. Modeling parameters and scenarios investigated by Reagan et al. (2015) 6-57
Table 7-1. Data from one company's operations indicating approximate total water use and
approximate produced water volumes within 10 days after completion of wells 7-6
ix
-------
Table 7-2. Additional short-, medium-, and long-term produced water estimates 7-7
Table 7-3. Flowback water characteristics for wells in unconventional reservoirs 7-8
Table 7-4. Long-term produced water generation rates (gal/day per well) for wells in
unconventional reservoirs 7-9
Table 7-5. Compiled minimum and maximum concentrations for various geochemical
constituents in produced water from shale gas, tight gas, and CBM produced water 7-17
Table 7-6. Examples of compounds identified in produced water that can be components of
hydraulic fracturing fluid 7-22
Table 7-7. Summary of produced water release volumes 7-37
Table 8-1. Estimated volumes (millions of gal) of wastewater based on state data for selected
years and numbers of wells producing fluid 8-9
Table 8-2. Estimated percentages of wastewater managed by practice and by state 8-16
Table 8-3. Management practices for wastewater from unconventional oil and gas resources 8-17
Table 8-4. Distribution of active Class IID wells across the United States 8-24
Table 8-5. Number, by state, of CWT facilities that have accepted or plan to accept wastewater
from unconventional oil and gas activities 8-30
Table 8-6. Estimated percentages of reuse of hydraulic fracturing wastewater 8-36
Table 9-1. Sources of selected RfVs, OSFs, and qualitative cancer classifications 9-15
Table 9-2. Chemicals reported to FracFocus 1.0 from January 1, 2011 to February 28, 2013 in
10% or more disclosures, with the percent of disclosures for which each chemical is reported.
Chronic oral RfVs, TOPKAT LOAEL estimates, and availability of ACToR data are shown when
available 9-25
Table 9-3. List of OSFs and qualitative cancer classifications available for all carcinogenic
chemicals reported to FracFocus 1.0 from January 1, 2011 to February 28, 2013 in 10% or more
disclosures 9-28
Table 9-4. List of a subset of organic chemicals that have been detected in produced water, with
respective chronic oral RfVs, TOPKAT LOAEL estimates, and availability of ACToR data shown
when available 9-32
Table 9-5. List of OSFs and qualitative cancer classifications available for a subset of organic
chemicals that have been reported in produced water 9-36
Table 9-6. List of inorganics and TENORM reported in produced water, and respective chronic
oral RfVs and OSFs when available 9-39
Table 9-7. List of qualitative cancer classifications available for inorganics and NORM that were
reported in produced water 9-42
Table 9-8. List of organochlorine pesticides and PCBs that were reported in produced water, and
their respective chronic oral RfVs, TOPKAT LOAEL estimates, and availability of data in EPA's
ACToR database 9-44
Table 9-9. List of OSFs and qualitative cancer classifications available for organochlorine
pesticides reported in produced water 9-46
Table 9-10. List of 45 chemicals on EPA's list that were used in hydraulic fracturing fluids and
detected in produced water and have an RfV or OSF available 9-48
x
-------
Table 9-11. Thresholds used for developing the Toxicity Score, Occurrence Score, and
Physicochemical Properties Score in this MCDA framework 9-59
Table 9-12. Data on the selected subset of chemicals in hydraulic fracturing fluids used for input
into a noncancer MCDA 9-62
Table 9-13. Data on the selected subset of chemicals in hydraulic fracturing fluids used for input
into a cancer MCDA 9-70
Table 9-14. Data on the selected subset of chemicals detected in produced water used for input
into a noncancer MCDA 9-72
Table 9-15. Data on the selected subset of chemicals detected in produced water used for input
into a cancer MCDA 9-76
Table 10-1. Literature estimates of mechanical integrity failure rates resulting in contamination
of groundwater or failure of all well barriers, potentially exposing the groundwater 10-15
xi
-------
List of Figures
Figure ES-1. General timeline and summary of activities at a hydraulically fractured oil or gas
production well ES-6
Figure ES-2. Locations of approximately 275,000 wells that were drilled and likely hydraulically
fractured between 2000 and 2013 ES-8
Figure ES-3. The five stages of the hydraulic fracturing water cycle ES-10
Figure ES-4. Water budgets illustrative of hydraulic fracturing water management practices in
the Marcellus Shale in the Susquehanna River Basin between approximately 2008 and 2013 and
the Barnett Shale in Texas between approximately 2011 and 2013 ES-14
Figure ES-5. Generalized depiction of factors that influence whether spilled hydraulic fracturing
fluids or additives reach drinking water resources, including spill characteristics, environmental
fate and transport, and spill response activities ES-25
Figure ES-6. Potential pathways for fluid movement in a cemented well ES-29
Figure ES-7. Examples of different subsurface environments in which hydraulic fracturing takes
place ES-31
Figure ES-8. Changes in wastewater management practices over time in the Marcellus Shale area
of Pennsylvania ES-41
Figure 1-1. Conceptualized view of the stages of the hydraulic fracturing water cycle 1-5
Figure 2-1. Geographic variability in drinking water sources for public water systems 2-10
Figure 2-2. The location of public water system sources having hydraulically fractured wells
within 1 mile 2-15
Figure 2-3. Separation distance between drinking water resources and hydraulically fractured
intervals in wells 2-17
Figure 3-1. Conceptual illustration of the types of oil and gas reservoirs and production wells
used in hydraulic fracturing 3-6
Figure 3-2. Major shale gas and oil plays in the contiguous United States 3-8
Figure 3-3. Major tight gas plays in the contiguous United States 3-9
Figure 3-4. Coalbed methane fields and coal basins in the contiguous United States 3-10
Figure 3-5. General timeline and summary of activities that take place during the preparation
and through the operations of an oil or gas well site at which hydraulic fracturing is used 3-12
Figure 3-6. Surface water being pumped for oil and gas development 3-14
Figure 3-7. Illustration of well construction showing different types of casing and cement 3-15
Figure 3-8. Sections of well casing ready for installation at a well site in Colorado 3-16
Figure 3-9. Aerial photograph of two hydraulic fracturing well sites and a service road in
Springville Township, Pennsylvania 3-17
Figure 3-10. Aerial photograph of hydraulic fracturing well sites near Williston, North Dakota 3-18
Figure 3-11. Well site with equipment (and pits in the background) in preparation for hydraulic
fracturing in Troy, Pennsylvania 3-19
Figure 3-12. Three wellheads on a multi-well pad connected to the piping used for hydraulic
fracturing injection 3-20
xii
-------
Figure 3-13. Water tanks (blue, foreground) lined up for hydraulic fracturing at a well site in
central Arkansas 3-22
Figure 3-14. A pit on the site of a hydraulic fracturing operation in central Arkansas 3-23
Figure 3-15. Locations of the approximately 275,000 wells drilled and hydraulically fractured
between 2000 and 2013 3-26
Figure 3-16. Landsat photo showing hydraulic fracturing well sites near Frierson, Louisiana 3-27
Figure 3-17. Landsat photo showing hydraulic fracturing well sites near Pinedale, Wyoming 3-28
Figure 3-18. Primary U.S. energy production by source, 1950 to 2015 3-32
Figure 3-19. U.S. production of oil (left) and gas (right) from hydraulically fractured wells from
2000 to 2015 3-33
Figure 3-20. Location of horizontal wells that began producing oil or natural gas in 2000, 2005,
and 2012 3-34
Figure 3-21. Natural gas prices and drilling activity, United States, 1988 to 2015 3-35
Figure 3-22. Historic and projected natural gas production by source (trillion cubic feet) 3-35
Figure 3-23. Production from U.S. shale gas plays, 2000-2014 3-36
Figure 3-24. Crude oil prices and drilling activity, United States, 1988 to 2015 3-37
Figure 3-25. Historic and projected oil production by source (million barrels per day) 3-37
Figure 3-26. Production from U.S. tight oil plays, 2000-2014 3-38
Figure 4-1. Median water volume per hydraulically fractured well nationally, expressed by well
type and completion year 4-12
Figure 4-2. Average annual hydraulic fracturing water use in 2011 and 2012 by county 4-18
Figure 4-3. (a) Average annual hydraulic fracturing water use in 2011 and 2012 compared to
total annual water use in 2010, by county, expressed as a percentage; (b) Average annual
hydraulic fracturing water consumption in 2011 and 2012 compared to total annual water
consumption in 2010, by county, expressed as a percentage 4-19
Figure 4-4. Locations of wells in the EPA FracFocus 1.0 project database, with respect to U.S. EIA
shale plays and basins 4-22
Figure 4-5. Major U.S. EIA shale plays and basins for Texas 4-22
Figure 4-6. Average annual hydraulic fracturing water use in 2011 and 2012 compared to (a)
fresh water available and (b) total water (fresh, brackish, and wastewater) available, by county,
expressed as a percentage 4-26
Figure 4-7. (a) Estimated annual surface water runoff from the USGS; (b) Reliance on
groundwater as indicated by the ratio of groundwater pumping to stream flow and pumping 4-27
Figure 4-8. Percentage of weeks in drought between 2000 and 2013 by county 4-29
Figure 4-9. Major U.S. EIA shale plays and basins for Colorado and Wyoming 4-31
Figure 4-10. Major U.S. EIA shale plays and basins for Pennsylvania, West Virginia, and Ohio 4-35
Figure 4-11. Major U.S. EIA shale plays and basins for North Dakota and Montana 4-40
Figure 4-12. Major U.S. EIA shale plays and basins for Arkansas and Louisiana 4-43
xiii
-------
Figure 5-1. Representative hydraulic fracturing site showing equipment used on-site during the
chemical mixing process 5-5
Figure 5-2. Overview of a chemical mixing process of the hydraulic fracturing water cycle 5-6
Figure 5-3. Example hydraulic fracturing fluid decision tree for gas and oil wells 5-9
Figure 5-4. Example hydraulic fracturing fluids 5-17
Figure 5-5. Estimated median volumes for 74 chemicals reported in at least 100 disclosures in
the FracFocus 1.0 project database for use in hydraulic fracturing from January 1, 2011 to
February 28, 2013 5-28
Figure 5-6. Typical hydraulic fracturing equipment layout 5-33
Figure 5-7. Metal and high-density polyethylene (HDPE) additive units 5-35
Figure 5-8. Hoses and lines at a site in Arkansas 5-37
Figure 5-9. Multiple fracture heads 5-40
Figure 5-10. Percent distribution of the causes of spills 5-42
Figure 5-11. Percent distribution of the sources of spills 5-43
Figure 5-12. Distribution of the number of spills for different ranges of spill volumes 5-43
Figure 5-13. Total volume of fluids spilled from different sources 5-44
Figure 5-14. Number of spills by environmental receptor 5-45
Figure 5-15. Fate and transport schematic for a spilled hydraulic fracturing fluid 5-48
Figure 5-16. Histograms of physicochemical properties of organic chemicals used in the
hydraulic fracturing process 5-52
Figure 5-17. Fate and Transport Spill Example: Case 1 5-60
Figure 5-18. Fate and Transport Spill Example: Case 2 5-61
Figure 5-19 Fate and Transport Spill Example: Case 3 5-62
Figure 6-1. Schematic cross-section of general types of oil and gas resources and the
orientations of production wells used in hydraulic fracturing 6-6
Figure 6-2. Overview of well construction 6-8
Figure 6-3. The various stresses to which the casing will be exposed 6-10
Figure 6-4. Potential pathways for fluid movement in a cemented wellbore 6-17
Figure 6-5. Hydraulic fracture planes (represented as ovals), with respect to the principal
subsurface compressive stresses: Sv (the vertical stress), Sh (the maximum horizontal stress),
and Sh (the minimum horizontal stress) 6-41
Figure 6-6. Vertical distances in the subsurface separating drinking water resources and
hydraulic fracturing depths 6-45
Figure 6-7. Conceptualized depiction of potential pathways for fluid movement out of the
production zone: (a) induced fracture overgrowth into over- or underlying formations; (b)
induced fractures intersecting natural fractures; and (c) induced fractures intersecting a
permeable fault 6-53
Figure 6-8. Induced fractures intersecting an offset well (in a production zone, as shown, or in
overlying formations into which fracture growth may have occurred) 6-59
xiv
-------
Figure 6-9. Well communication (a frac hit) 6-60
Figure 7-1. Generalized examples of produced water flow from five formations 7-10
Figure 7-2. Typical produced water volume for a coal bed methane well in the western United
States 7-11
Figure 7-3. TDS concentrations measured through time for injected fluid (at 0 days), and
produced water samples from four Marcellus Shale gas wells in three southwest Pennsylvania
counties 7-13
Figure 7-4. Total radium and TDS concentrations measured through time for injected (day 0),
and produced water samples Greene County, PA, Marcellus Shale gas wells 7-14
Figure 7-5. (a) Increasing chloride (CI) and (b) decreasing DOC concentrations measured
through time for samples from three Marcellus Shale gas wells on a single well pad in Greene
County, PA 7-15
Figure 7-6. Data on radium 226 (open symbols) and total radium (filled symbols) for Marcellus
Shale wells (leftmost three columns) and other formations (rightmost three columns) 7-21
Figure 7-7. Produced water spill rates (spills per active wells) for North Dakota from 2001 to
2015 (Appendix Section E.5) 7-31
Figure 7-8. Number of produced water releases in North Dakota by cause for 2014 and 2015
(Appendix Section E.5) 7-32
Figure 7-9. Distribution of spill causes in Oklahoma, pre-high volume hydraulic fracturing years
of 1993-2003 (left) and in the EPA study of spills on production pads (right) 7-33
Figure 7-10. Distribution of spill sources in Oklahoma, pre-high volume hydraulic fracturing
years of 1993-2003 (left) and in the EPA study of spills on production pads (right) 7-33
Figure 7-11. Volumes of 2015 North Dakota salt water releases by cause (leftmost 13 boxes in
red), and all causes (last box in blue) 7-34
Figure 7-12. Volumes of produced water spills reported by the EPA for 2006 to 2012 by cause
(the five left most boxes in red), source (the second five boxes in yellow), and all spills (blue) 7-35
Figure 7-13. Median, mean, and maximum produced water spill volumes for North Dakota from
2001 to 2015 7-36
Figure 7-14. Schematic view of transport processes occurring during releases of produced
water 7-40
Figure 8-1. Wastewater (i.e., produced water and fracturing fluid waste) and produced gas
volumes from unconventional (as defined by PA DEP) wells in Pennsylvania from January 2010
through June 2016 8-5
Figure 8-2. Wastewater quantities in the western United States (billions of gal per year) 8-7
Figure 8-3. Schematic of wastewater management strategies 8-14
Figure 8-4. Percentages of total unconventional wastewater (as defined by PA DEP) managed via
various practices for the second half of 2009 through the first half of 2014 8-21
Figure 8-5. Management of wastewater in Colorado in regions where hydraulic fracturing is
being performed 8-22
Figure 8-6. Oil and gas wastewater volumes discharged to POTWs from 2001-2011 in the
Marcellus Shale. ("Conventional" is indicated by the authors as non-Marcellus wells and
described as vertically drilled to shallower depths in more porous formations.) 8-27
xv
-------
Figure 8-7. Map showing Pennsylvania surface water designated as potable water supplies and
upstream CWTs 8-32
Figure 8-8. Lined evaporation pit in the Battle Creek Field (Montana) 8-42
Figure 9-1. Fate and transport schematic for a hydraulic fracturing-related spill or release 9-6
Figure 9-2. Percentage of hydraulic fracturing-related chemicals (out of 1,606 total) with at least
one data point in each ACToR data class 9-20
Figure 9-3. Overall representation of the selected toxicological, physicochemical, and occurrence
data available for the 1,606 hydraulic fracturing-related chemicals identified by the EPA 9-22
Figure 9-4. Availability of toxicity data (chronic oral RfVs/OSFs, TOPKAT LOAEL estimates, and
relevant data on ACToR) for subsets of chemicals used at various frequencies in hydraulic
fracturing fluids, as determined based on the number of disclosures in the EPA FracFocus 1.0
project database 9-29
Figure 9-5. Overview of the MCDA framework for hazard evaluation 9-53
Figure 9-6. The subsets of chemicals selected for hazard evaluation using the noncancer MCDA
framework included 42 chemicals used in hydraulic fracturing fluids and 29 chemicals detected
in produced water 9-55
Figure 9-7. The subsets of chemicals selected for hazard evaluation using the cancer MCDA
framework included 10 chemicals used in hydraulic fracturing fluids, and 7 chemicals detected
in produced water 9-56
Figure 9-8. Noncancer MCDA results for 42 chemicals used in hydraulic fracturing fluids
(national analysis), showing the Toxicity Score, Occurrence Score, and Physicochemical
Properties Score for each chemical 9-65
Figure 9-9. Noncancer MCDA results for 36 chemicals used in hydraulic fracturing fluids in
Texas (state-specific analysis), showing the Toxicity Score, Occurrence Score, and
Physicochemical Properties Score for each chemical 9-66
Figure 9-10. Noncancer MCDA results for 20 chemicals used in hydraulic fracturing fluids in
Pennsylvania (state-specific analysis), showing the Toxicity Score, Occurrence Score, and
Physicochemical Properties Score for each chemical 9-67
Figure 9-11. Noncancer MCDA results for 21 chemicals used in hydraulic fracturing fluids in
North Dakota (state-specific analysis), showing the Toxicity Score, Occurrence Score, and
Physicochemical Properties Score for each chemical 9-68
Figure 9-12. Cancer MCDA results for 10 chemicals used in hydraulic fracturing fluids, showing
the Toxicity Score, Occurrence Score, and Physicochemical Properties Score for each chemical 9-71
Figure 9-13. Noncancer MCDA results for a subset of 29 chemicals detected in produced water,
showing the Toxicity Score, Occurrence Score, and Physicochemical Properties Score for each
chemical 9-74
Figure 9-14. Cancer MCDA results for 7 chemicals detected in produced water, showing the
Toxicity Score, Occurrence Score, and Physicochemical Properties Score for each chemical 9-77
Figure 10-1. Water budgets representative of practices in (top) the Marcellus Shale in the
Susquehanna River Basin in Pennsylvania and (bottom) the Barnett Shale in Texas 10-7
Figure 10-2. Fate and transport schematic for a spill of chemicals, hydraulic fracturing fluid, or
produced water 10-11
Figure 10-3. Separation in measured depth between drinking water resources and hydraulically
fractured intervals in wells 10-20
xvi
-------
List of Text Boxes
Text Box ES-1. Drinking Water Resources ES-5
Text Box ES-2. Hydraulically Fractured Oil and Gas Production Wells ES-7
Text Box ES-3. The EPA's Study of the Potential Impacts of Hydraulic Fracturing for Oil and Gas on
Drinking Water Resources ES-11
Text Box ES-4. FracFocus Chemical Disclosure Registry ES-13
Text Box ES-5. County-Level Water Use for Hydraulic Fracturing ES-16
Text Box ES-6. Examples of Hydraulic Fracturing Fluids ES-19
Text Box ES-7. Chemical Mixing Equipment ES-23
Text Box ES-8. Fracture Growth ES-28
Text Box ES-9. Produced Water from Hydraulically Fractured Oil and Gas Production Wells ES-34
Text Box ES-10. On-Site Storage of Produced Water ES-36
Text Box ES-11. Hydraulic Fracturing Wastewater Management ES-39
Text Box 1-1. Regulatory Protection for Drinking Water Resources 1-7
Text Box 2-1. The Hydrologic Cycle 2-4
Text Box 2-2. El Paso's Use of Higher Salinity Water for Drinking Water 2-11
Text Box 3-1. Hydraulic Fracturing: Not New, but Different and Still Changing 3-4
Text Box 3-2. "Conventional" Versus "Unconventional." 3-7
Text Box 4-1. Using the EPA's FracFocus 1.0 Project Database to Estimate Water Use for
Hydraulic Fracturing 4-20
Text Box 4-2. Hydraulic Fracturing Water Use as a Percentage of Water Availability Estimates 4-25
Text Box 4-3. Case Study: Water Profile of the Eagle Ford Play, Texas 4-30
Text Box 4-4. Case Study: Impact of Water Acquisition for Hydraulic Fracturing on Local Water
Availability in the Upper Colorado River Basin 4-34
Text Box 4-5. Case Study: Impact of Water Acquisition for Hydraulic Fracturing on Local Water
Availability in the Susquehanna River Basin 4-38
Text Box 5-1. The FracFocus Registry and EPA FracFocus Report 5-18
Text Box 5-2. Confidential Business Information (CBI) 5-20
Text Box 5-3. Spills from Storage Units 5-34
Text Box 5-4. Spill from Additive (Crosslinker) Storage Tote 5-34
Text Box 5-5. Spill of Acid from Storage Container 5-36
Text Box 5-6. Spill of Gel Slurry during Mixing 5-36
Text Box 5-7. Spill of Hydraulic Fracturing Fluid from Blender 5-38
Text Box 5-8. Spill of Fluid from Fracture Pump 5-39
Text Box 5-9. Spill from Frac Head Failure 5-40
xvii
-------
Text Box 5-10. EPA Review of State and Industry Spill Data: Characterization of Hydraulic
Fracturing-Related Spills 5-41
Text Box 6-1. The Well File Review 6-9
Text Box 6-2. Dimock, Pennsylvania 6-11
Text Box 6-3. Stray Gas Migration 6-23
Text Box 6-4. Parker County, Texas 6-26
Text Box 6-5. Pavillion, Wyoming 6-47
Text Box 6-6. Monitoring at the Greene County, Pennsylvania, Hydraulic Fracturing Test Site 6-54
Text Box 6-7. Well Communication at a Horizontal Well near Innisfail, Alberta, Canada 6-63
Text Box 8-1. Temporal Trends in Wastewater Management - Experience of Pennsylvania 8-19
Text Box 8-2. Regulations Affecting Wastewater Management 8-21
Text Box 8-3. Wastewater Treatment Processes 8-28
Text Box 9-1. Applying Toxicological Data for Human Health Risk Assessment 9-5
Text Box 9-2. The EPA's List of Chemicals Identified in Hydraulic Fracturing Fluids and/or
Produced Water 9-9
Text Box 9-3. Toxicity Values for Hydraulic Fracturing-Related Chemicals 9-12
Text Box 10-1. Hydraulic Fracturing and Groundwater Quality Monitoring in California 10-25
Text Box 10-2. Causal Assessment and Hydraulic Fracturing Water Cycle Activities 10-27
xviii
-------
List of Acronyms/Abbreviations
Acronvm
Definition
2BE
2-butoxyethanol
ACToR
Aggregated Computational Toxicology Resource database
AME
Acton Mickelson Environmental, Inc.
AMEC
AMEC Environment & Infrastructure, Inc.
ANRC
Arkansas Natural Resources Commission
AO
administrative order
AOGC
Arkansas Oil and Gas Commission
API
American Petroleum Institute
ATSDR
Agency for Toxic Substance and Disease Registry
AWWA
American Water Works Association
BLM
Bureau of Land Management
BTEX
benzene, toluene, ethylbenzene, and xylenes
CARES
Casella Altela Regional Environmental Services
CASRN
chemical abstract services registration number
CBI
confidential business information
CBM
coalbed methane
CCST
California Council on Science and Technology
CDC
Centers for Disease Control and Prevention
CDWR
Colorado Division of Water Resources
CFR
Code of Federal Regulations
CICAD
Concise International Chemical Assessment Document
CM
chemical mixing
CMV
commercial motor vehicle
COGCC
Colorado Oil and Gas Conservation Commission
COWDF
Commercial Oil Field Waste Disposal Facilities
CWA
Clean Water Act
CWCB
Colorado Water Conservation Board
CWT
centralized waste treatment
CWTF
centralized water treatment facility
DBNM
dibromochloronitromethane
DBP
disinfection byproduct
DecaBDE
decabromodipheyl ether
DfE
Design for the Environment
DI
Drilling Info, Inc.
DMA
dimethylamine
DMR
Discharge Monitoring Report
DNR
Department of Natural Resources
DO
dissolved oxygen
DOC
dissolved organic carbon
xix
-------
DOE
U.S. Department of Energy
DOGGR
California Department of Conservation's Division of Oil, Gas & Geothermal Resources
DOI
U.S. Department of the Interior
DO]
U.S. Department of Justice
DOT
U.S. Department of Transportation
DRB
Delaware River Basin
DRO
diesel range organics
DWSHA
Drinking Water Standards and Health Advisories
ECHA
European Chemicals Agency
EERC
Energy and Environmental Research Center, University of North Dakota
EIA
U.S. Energy Information Administration
EPA
U.S. Environmental Protection Agency
EPAOW
U.S. Environmental Protection Agency, Office of Water
EPI
estimation programs interface
EPWU
El Paso Water Utility
ERCB
Energy Resource Conservation Board
ERG
Eastern Research Group
ESN
Environmental Services Network
ESOD
erythrocyte Cu, Zn-superoxide dismutase
EWI
Energy Water Initiative
FDA
U.S. Food and Drug Administration
FOIA
Freedom of Information Act
FRS
fluids recovery services
GES
Groundwater & Environmental Services, Inc.
GHGRP
Greenhouse Gas Reporting Program
GNB
Government of New Brunswick
GRAS
generally recognized as safe
GRO
gasoline range organics
GTI
Gas Technology Institute
GWPC
Ground Water Protection Council
HBCD
hexabromocyclododecane
HDPE
high-density polyethylene
HF
hydraulic fracturing
HHBP
Human Health Benchmarks for Pesticides
HISA
Highly Influential Scientific Assessment
HPG
hydroxypropylguar
HTS
high throughput screening
HUC
hydrological unit code
IAEA
International Atomic Energy Agency
IARC
International Agency for Research on Cancer
IHS
Information Handling Services
IOGCC
Interstate Oil and Gas Compact Commission
XX
-------
IPCC Intergovernmental Panel on Climate Change
IPCS International Programme on Chemical Safety
IRIS Integrated Risk Information System
IUPAC International Union of Pure and Applied Chemistry
KWO Kansas Water Office
LDEQ Louisiana Department of Environmental Quality
LOAEL lowest observed adverse effect level
MCDA multicriteria decision analysis
MCL maximum contaminant level
MCLG maximum containment level goal
MCOR Marcellus Center for Outreach and Research
MGD million gallons per day
MIT mechanical integrity test
MMCF million cubic feet
MRL minimum risk level
MSC Marcellus Shale Coalition
MT GWIC Montana Ground Water Information Center
MTBE methyl tert-butyl ether
MVR mechanical vapor recompression
NAS National Academy of Sciences
NDDMR North Dakota Department of Mineral Resources
NDDOH North Dakota Department of Health
NDMA N-nitrosodimethylamine
NDPES National Pollution Discharge Elimination System
NDSWC North Dakota State Water Commission
NETL National Energy Technology Laboratory
NGO non-governmental organization
NIH National Institutes of Health
NM OCD New Mexico Oil Conservation Division
NM OSE New Mexico Office of the State Engineer
NOAEL no observed adverse effect level
NORM naturally occurring radioactive material
NPC National Petroleum Council
NPDES National Pollution Discharge Elimination System
NPDWR National Primary Drinking Water Regulations
NRC National Resource Council
NTP U.S. National Toxicology Program
NYSDEC New York State Department of Environmental Conservation
O&G oil and gas
ODNR Ohio Department of Natural Resources
DMRM Division of Mineral Resources Management
OECD The Organisation for Economic Co-operation and Development
xxi
-------
OEPA
Ohio Environmental Protection Agency
OMB
Office of Management and Budget
ORB
Ohio River Basin
ORD
Office of Research and Development
OSF
oral slope factor
OSHA
Occupational Safety & Health Administration
OSWER
Office of Solid Water and Emergency Response
OWRB
Oklahoma Water Resources Board
PA DCNR
Pennsylvania Department of Conservation and Natural Resources
PADEP
Pennsylvania Department of Environmental Protection
PAH
polycyclic aromatic hydrocarbon
PCB
polychlorinated biphenyl
PFBC
Pennsylvania Fish and Boat Commission
PDL
positive determination letter
PMF
Positive Matrix Factorization
PMN
pre-manufacturing notices
POD
point-of-departure
POTW
publicly owned treatment work
PPRTV
provisional peer-reviewed toxicity value
PVC
polyvinyl chloride
PWS
public water system
PWSA
Pittsburgh Water and Sewer Authority
QA
quality assurance
QAPP
quality assurance project plan
QC
quality control
QSAR
Quantitative Structure Activity Relationship
RAHC
reasonably anticipated to be a human carcinogen
RBC
red blood cells
RfD
reference dose
RfV
reference value
RO
reverse osmosis
SAB
Science Advisory Board
SAIC
Science Applications International Corporation
SAR
sodium adsorption ratio
SCN
thiocyanates
SDWA
Safe Drinking Water Act
SDWIS
Safe Drinking Water Information System
SEECO
Southern Electrical Equipment Company
SGEIS
supplemented generic environmental impact statement
SHS MSC
statewide health standards for medium-specific concentrations
SMCL
secondary maximum contaminant level
SPE
Society of Petroleum Engineers
xxii
-------
SRB
Susquehanna River basin
SRBC
Susquehanna River Basin Commission
STO
Statoil
STRONGER
State review of oil and natural gas environmental regulations
SVOC
semi-volatile organic compounds
SWE
Southwestern Energy
TARM
TerrAqua Resource Management
TBA
tert-butyl alcohol
TDI
tolerable daily intake
TDS
total dissolved solids
TENORM
technologically enhanced naturally occurring radioactive material
THM
trihalomethane
TIPRO
Texas Independent Producers and Royalty Owners Association
TMDL
total maximum daily load
TOC
total organic carbon
TOPKAT
Toxicity Prediction by Komputer Assisted Technology
TPH
total petroleum hydrocarbons
TPHWG
Total Petroleum Hydrocarbon Criteria Working Group
TSS
total suspended solids
TTC
Threshold of Toxicological Concern
TTHM
total trihalomethane
TWDB
Texas Water Development Board
TXRRC
Texas Railroad Commission
UCRB
Upper Colorado River basin
UIC
underground injection control
UOG
unconventional oil and gas
USGAO
U.S. Government Accountability Office
USGS
U.S. Geological Survey
UWS
Universal Well Services
VES
viscoelastic surfactant
VOC
volatile organic compounds
WAWSA
Western Area Water Supply Authority
WFR
Well File Review
WHO
World Health Organization
WOE
weight of evidence
WRF
Water Research Foundation
WVDEP
West Virginia Department of Environmental Protection
WWTP
wastewater treatment plant
WYOGCC
Wyoming Oil and Gas Conservation Commission
xxiii
-------
Preface
Hydraulic fracturing is a technique used to increase oil and gas production from underground oil-
or gas-bearing rock formations. Since the mid-2000s, the combination of hydraulic fracturing and
directional drilling has become widespread, raising concerns about the potential impacts of
hydraulic fracturing on drinking water resources. This concern is the focus of this report.
In 2010, the U.S. Environmental Protection Agency (EPA) initiated a study of the potential impacts
of hydraulic fracturing activities on drinking water resources. The EPA defined the scope of its
study to focus on the acquisition, use, disposal, and reuse of water used for hydraulic fracturing—
what we call the hydraulic fracturing water cycle. This was done in recognition that concerns raised
about potential impacts were not limited to the relatively short-term act of fracturing rock, but can
include impacts related to other activities associated with hydraulic fracturing.
The EPA's study included the development of multiple research projects using the following
research approaches: the analysis of existing data, scenario and modeling evaluations, laboratory
studies, toxicological assessments, and five case studies. Throughout the study, the EPA engaged
with stakeholders, including industry, the states, tribal nations, academia, and others, for input on
the scope, approach, and initial results. To date, the study has resulted in the publication of multiple
peer-reviewed scientific products, including 13 EPA technical reports and 14 journal articles.
This report represents the capstone product of the EPA's hydraulic fracturing drinking water study.
It captures the state-of-the-science concerning drinking water impacts from activities in the
hydraulic fracturing activities water cycle and integrates the results of the EPA's study of the
subject with approximately 1,200 other publications and sources of information. The goals of this
report were to assess the potential for activities in the hydraulic fracturing water cycle to impact
the quality or quantity of drinking water resources and to identify factors that affect the frequency
or severity of those impacts.
This report is a science document and does not present or evaluate policy options or make policy
recommendations. A draft of this report was reviewed by the EPA's independent Science Advisory
Board (SAB). Reflecting the complexity of the subject, the expert ad hoc panel formed by the SAB
was the largest ever convened for the review of a scientific product Combined with over 100,000
comments submitted by members of the public, SAB comments helped the EPA to refine, clarify,
and better support the final conclusions presented in this report.
The release of this final assessment report marks the completion of the EPA's hydraulic fracturing
drinking water study. The study has already prompted increased dialogue among industry, the
states, tribal nations, the public, and others concerning how drinking water resources can be better
protected in areas where hydraulic fracturing is occurring or being considered. However, there are
data gaps and uncertainties limiting our understanding of the impacts of hydraulic fracturing
activities on drinking water resources. As additional data become available, and with continued
dialogue among stakeholders, our understanding of the potential impacts of hydraulic fracturing on
drinking water resources will improve.
xxiv
-------
Authors, Contributors, and Reviewers
Authors of the Final Assessment (EPA 600-R-16-236Fa and EPA 600-R-16-236Fb)
Susan Burden, USEPA-Office of Research and Development, Washington, DC
Megan M. Fleming, USEPA-Office of Research and Development, Washington, DC
Jeffrey Frithsen, USEPA-Office of Research and Development, Washington, DC
Linda Hills, The Cadmus Group, Inc., Helena, MT
Kenneth Klewicki, The Cadmus Group, Inc., Arlington, VA
Christopher D. Knightes, USEPA-Office of Research and Development, Athens, GA
Sandie Koenig, The Cadmus Group, Inc., Helena, MT
Jonathan Koplos, The Cadmus Group, Inc., Waltham, MA
Stephen D. LeDuc, USEPA-Office of Research and Development, Washington, DC
Caroline E. Ridley, USEPA-Office of Research and Development, Washington, DC
Shari Ring, The Cadmus Group, Inc., Arlington, VA
Sarah Solomon, Student Services Contractor, USEPA-Office of Research and Development, Washington, DC
John Stanek, USEPA-Office of Research and Development, Research Triangle Park, NC
Mary Ellen Tuccillo, The Cadmus Group, Inc., Waltham, MA
Jim Weaver, USEPA-Office of Research and Development, Ada, OK
Anna Weber, The Cadmus Group, Inc., Arlington, VA
Nathan Wiser, USEPA-Office of Research and Development, Denver, CO
Erin Yost, USEPA-Office of Research and Development, Research Triangle Park, NC
Authors of the June 2015 External Review Draft (EPA-600-R-15-047a)
William Bates, USEPA-Office of Water, Washington, DC
Glen Boyd, The Cadmus Group, Inc., Seattle, WA
Jeanne Briskin, USEPA-Office of Research and Development, Washington, DC
Lyle Burgoon, USEPA-Office of Research and Development, Research Triangle Park, NC; currently with
US-ACOE, Research Triangle Park, NC
Susan Burden, USEPA-Office of Research and Development, Washington, DC
Christopher M. Clark, USEPA-Office of Research and Development, Washington, DC
Maryam Cluff, Student Services Contractor, USEPA-Office of Research and Development, Washington, DC
Rebecca Daiss, USEPA-Office of Research and Development, Washington, DC
Jill Dean, USEPA-Office of Water, Washington, DC
Inci Demirkanli, The Cadmus Group, Inc., Arlington, VA
Megan M. Fleming, USEPA-Office of Research and Development, Washington, DC
Jeffrey Frithsen, USEPA-Office of Research and Development, Washington, DC
Linda Hills, The Cadmus Group, Inc., Helena, MT
Kenneth Klewicki, The Cadmus Group, Inc., Arlington, VA
Christopher D. Knightes, USEPA-Office of Research and Development, Athens, GA
Sandie Koenig, The Cadmus Group, Inc., Helena, MT
Jonathan Koplos, The Cadmus Group, Inc., Waltham, MA
Stephen D. LeDuc, USEPA-Office of Research and Development, Washington, DC
XXV
-------
Claudia Meza-Cuadra, Student Services Contractor, USEPA-Office of Research and Development,
Washington, DC
Brent Ranalli, The Cadmus Group, Inc., Waltham, MA
Caroline E. Ridley, USEPA-Office of Research and Development, Washington, DC
Shari Ring, The Cadmus Group, Inc., Arlington, VA
Alison Singer, Student Services Contractor, USEPA-Office of Research and Development, Washington, DC
John Stanek, USEPA-Office of Research and Development, Research Triangle Park, NC
M. Jason Todd, USEPA-Office of Research and Development, Washington, DC
Mary Ellen Tuccillo, The Cadmus Group, Inc., Waltham, MA
Jim Weaver, USEPA-Office of Research and Development, Ada, OK
Anna Weber, The Cadmus Group, Inc., Arlington, VA
Larke Williams, USEPA-Office of Research and Development, Washington, DC; currently with the US
State Department, Washington, DC
Liabeth Yohannes, Student Services Contractor, USEPA-Office of Research and Development,
Washington, DC
Erin Yost, USEPA-Office of Research and Development, Research Triangle Park, NC
Contributors
Maryam Akhavan, The Cadmus Group, Inc., Arlington, VA
Natalie Auer, The Cadmus Group, Inc., Arlington, VA
Kevin Blackwood, Student Services Contractor, USEPA-Office of Research and Development, Ada, OK
Alison Cullity, The Cadmus Group, Rollinsford, NH
Rob Dewoskin, USEPA-Office of Research and Development, Research Triangle Park, NC
Krissy Downing, The Cadmus Group, Inc., Seattle, WA
Christopher Impellitteri, USEPA-Office of Research and Development, Cincinnati, OH
Will Jobs, The Cadmus Group, Inc., Waltham, MA
Richard Judson, USEPA-Office of Research and Development, Research Triangle Park, NC
Erina Keefe, The Cadmus Group, Waltham, MA
Ava Lazor, The Cadmus Group, Inc., Arlington, VA
Matt Landis, USEPA-Office of Research and Development, Research Triangle Park, NC
Ralph Ludwig, USEPA-Office of Research and Development, Ada, OK
John Martin, The Cadmus Group, Inc., Waltham, MA
Ashley McElmury, Student Services Contractor, USEPA-Office of Research and Development, Ada, OK
Gary Norris, USEPA-Office of Research and Development, Research Triangle Park, NC
Kay Pinley, Senior Environmental Employment Program, USEPA-Office of Research and Development,
Ada, OK
Jesse Pritts, USEPA-Office of Water, Washington, DC
Ann Richard, USEPA-Office of Research and Development, Research Triangle Park, NC
Ana Rosner, The Cadmus Group, Inc., Waltham, MA
Susan Sharkey, USEPA-Office of Research and Development, Washington, DC
Jessica Wilhelm, Student Services Contractor, USEPA-Office of Research and Development, Ada, OK
Holly Wooten, The Cadmus Group, Inc., Arlington, VA
Jie Xu, Student Services Contractor, USEPA-Office of Research and Development, Ada, OK
xxvi
-------
U.S. Environmental Protection Agency Science Advisory Board
Joseph Arvai, University of Michigan, Ann Arbor, MI
Kiros T. Berhane, University of Southern California, Los Angeles, CA
Sylvie M. Brouder, Purdue University, West Lafayette, IN
Ingrid Burke, University of Wyoming, Laramie, WY
Thomas Carpenter, Designated Federal Officer, U.S. Environmental Protection Agency, Science Advisory
Board, Washington, DC
Ana V. Diez Roux, Drexel University, Philadelphia, PA
Michael Dourson, University of Cincinnati, Cincinnati, OH
Joel J. Ducoste, North Carolina State University, Raleigh, NC
David A. Dzombak, Carnegie Mellon Unviersity, Pittsburgh, PA
Elaine M. Faustman, University of Washington, Seattle, WA
Susan P. Felter, Proctor & Gamble, Mason, OH
R. William Field, University of Iowa, Iowa City, IA
H. Christopher Frey, North Carolina State University, Raleigh, NC
Steven Hamburg, Environmental Defense Fund, Boston, MA
Cynthia M. Harris, Florida A&M University, Tallahassee, FL
Robert J. Johnston, Clark University, Worcester, MA
Kimberly L. Jones, Howard University, Washington, DC
Catherine J. Karr, University of Washington, Seattle, WA
Madhu Khanna, University of Illinois at Urbana-Champaign, Urbana, IL
Francine Laden, Brigham and Women's Hospital and Harvard Medical School, Boston, MA
Lois Lehman-McKeeman, Bristol-Myers Squibb, Princeton, NJ
Robert E. Mace, Texas Water Development Board, Austin, TX
Mary Sue Marty, The Dow Chemical Company, Midland, MI
Denise Mauzerall, Princeton University, Princeton, NJ
Kristina D. Mena, University of Texas Health Science Center at Houston, El Paso, TX
Surabi Menon, ClimateWorks Foundation, San Francisco, CA
James R. Mihelcic, University of South Florida, Tampa, FL
Keith H. Moo-Young, Washington State University, Tri-Cities, Richland, WA
Kari Nadeau, Stanford University School of Medicine, Stanford, CA
James Opaluch, University of Rhode Island, Kingston, RI
Thomas F. Parkerton, ExxonMobil Biomedical Science, Houston, TX
Richard L. Poirot, Independent Consultant, Burlington, VT
Kenneth M. Portier, American Cancer Society, Atlanta, GA
Kenneth Ramos, University of Arizona, Tucson, AZ
David B. Richardson, University of North Carolina, Chapel Hill, NC
Tara L. Sabo-Attwood, University of Florida, Gainesville, FL
William Schlesinger, Cry Institute of Ecosystem Studies, Millbrook, NY
Gina Solomon, California Environmental Protection Agency, Sacramento, CA
Daniel 0. Stram, University of Southern California, Los Angeles, CA
Peter S. Thorne (Chair), University of Iowa, Iowa City, IA
xxvii
-------
Jay Turner, Washington University, St. Louis, MO
Edwin van Wijngaarden, University of Rochester, Rochester, NY
Jeanne M. VanBriesen, Carnegie Mellon University, Pittsburgh, PA
John Vena, Medical University of South Carolina, Charleston, SC
Elke Weber, Columbia Business School, New York, NY
Charles Werth, University of Texas at Austin, Austin, TX
Peter J. Wilcoxen, Syracuse University, Syracuse, NY
Robyn S. Wilson, The Ohio State University, Columbus, OH
U.S. Environmental Protection Agency Science Advisory Board Hydraulic Fracturing
Research Advisory Panel
Stephen W. Almond, Fritz Industries, Inc., Houston, TX
E. Scott Bair, The Ohio State University, Columbus, OH
Peter Bloomfield, North Carolina State University, Raleigh, NC
Steven R. Bohlen, State of California Department of Conservation, Sacramento, CA
Elizabeth W. Boyer, Pennsylvania State University, University Park, PA
Susan L. Brantley, Pennsylvania State University, University Park, PA
James V. Bruckner, University of Georgia, Athens, GA
Thomas L. Davis, Colorado School of Mines, Golden, CO
Joseph J. DeGeorge, Merck Research Laboratories, Lansdale, PA
Joel Ducoste, North Carolina State University, Raleigh, NC
Shari Dunn-Norman, Missouri University of Science and Technology, Rolla, MO
David A. Dzombak (Chair), Carnegie Mellon University, Pittsburgh, PA
Katherine Bennett Ensor, Rice University, Houston, TX
Elaine M. Faustman, University of Washington, Seattle, WA
John V. Fontana, Vista GeoScience LLC, Golden, CO
Daniel J. Goode, U.S. Geological Survey, Exton, PA
Edward Hanlon, Designated Federal Officer, U.S. Environmental Protection Agency, Science Advisory
Board Staff, Washington, DC
Bruce D. Honeyman, Colorado School of Mines, Golden, CO
Walter R. Hufford, Talisman Energy USA Inc. - REPSOL, Warrendale, PA
Richard F. Jack, Thermo Fisher Scientific Inc., San Jose, CA
Dawn S. Kaback, Amec Foster Wheeler, Denver, CO
Abby A. Li, Exponent Inc., San Francisco, CA
Dean N. Malouta, White Mountain Energy Consulting, LLC, Houston, TX
Cass T. Miller, University of North Carolina, Chapel Hill, NC
Laura J. Pyrak-Nolte, Purdue University, West Lafayette, IN
Stephen Randtke, University of Kansas, Lawrence, KS
Joseph N. Ryan, University of Colorado-Boulder, Boulder, CO
James E. Saiers, Yale University, New Haven, CT
Azra N. Tutuncu, Colorado School of Mines, Golden, CO
Paul K. Westerhoff, Arizona State University, Tempe, AZ
Thomas M. Young, University of California-Davis, Davis, CA
xxviii
-------
U.S. Environmental Protection Agency Internal Technical Reviewers
Lisa Biddle, Office of Water, Washington, DC
Britta Bierwagen, Office of Research and Development, Washington, DC
Frank Brock, Region 2, New York, NY
Thomas Burke, Office of Research and Development, Washington, DC
Kyle Carey, Office of Water, Washington, DC
Mark Corrales, Office of Policy, Washington, DC
Brian D'Amico, Office of Water, Washington, DC
Kathleen Deener, Office of Research and Development, Washington, DC
Tim Elkins, Region 5, Chicago, IL
Malcolm Field, Office of Research and Development, Washington, DC
Erin Floto, Region 2, New York, NY
Robert Ford, Office of Research and Development, Cincinnati, OH
Greg Fritz, Office of Chemical Safety and Pollution Prevention, Washington, DC
Andrew Gillespie, Office of Research and Development, Research Triangle Park, NC
Janet Goodwin, Office of Water, Washington, DC
Bradley Grams, Region 5, Chicago, IL
Holly Green, Office of Water, Washington, DC
Richard Hall, Region 4, Atlanta, GA
Mary Hanley, Administrator's Office, Washington, DC
Mohamed Hantush, Office of Research and Development, Cincinnati, OH
Jana Harvill, Region 6, Dallas, TX
Fred Hauchman, Office of Science Policy, Washington, DC
Kurt Hildebrandt, Region 7, Lenexa, KS
Charles Hillenbrand, Region 2, New York, NY
Mark W. Howard, Office of Solid Waste and Emergency Response, Washington, DC
Junqi Huang, Office of Research and Development, Ada, OK
Stephen Jann, Region 5, Chicago, IL
Thomas Johnson, Office of Research and Development, Washington, DC
Jeff Jollie, Office of Water, Washington, DC
Robert Kavlock, Office of Research and Development, Washington, DC
James Kenney, Office of Enforcement and Compliance Assurance, Washington, DC
Kristin Keteles, Region 8, Denver, CO
Bruce Kobelski, Office of Water, Washington, DC
Stephen Kraemer, Office of Research and Development, Athens, GA
Paul Lewis, Office of Chemical Safety and Pollution Prevention, Washington, DC
Chris Lister, Region 6, Dallas, TX
Barbara Martinez, ORISE Fellow to USEPA-Office of Research and Development, Washington, DC
Mike Mattheisen, Office of Chemical Safety and Pollution Prevention, Washington, DC
Damon McElroy, Region 6, Dallas, TX
Karen Milam, Office of Water, Washington, DC
Keara Moore, Office of Water, Washington, DC
xxix
-------
Nathan Mottl, Office of Chemical Safely and Pollution Prevention, Washington, DC
Greg Oberley, Region 8, Denver, CO
Mike Overbay, Region 6, Dallas, TX
Pooja Parikh, Office of General Council, Washington, DC
Dale Perry, Administrator's Office, Washington, DC
Tricia Pfeiffer, Region 8, Denver, CO
Steve Piatt, Region 3, Philadelphia, PA
Dave Rectenwald, Region 3, Philadelphia, PA
Meredith Russell, Office of Water, Washington, DC
Daniel Ryan, Region 3, Philadelphia, PA
Greg Schweer, Office of Chemical Safety and Pollution Prevention, Washington, DC
Brian Smith, Region 4, Atlanta, GA
Kelly Smith, Office of Research and Development, Cincinnati, OH
Steve Souders, Office of Solid Waste and Emergency Response, Washington, DC
Kate Sullivan, Office of Research and Development, Athens, GA
Kevin Teichman, Office of Research and Development, Washington, DC
Chuck Tinsley, Region 8, Denver, CO
Scott Wilson, Office of Water, Washington, DC
Jose Zambrana, Office of Research and Development, Washington, DC
xxx
-------
Acknowledgements
The development of this assessment involved individuals from across the Agency, many of which
have been listed as authors, contributors or reviewers, or acknowledged below. We want to
specifically highlight and acknowledge Jeanne Briskin for her role with making the EPA's Hydraulic
Fracturing Drinking Water Study a success. Jeanne was responsible for the development of the
2011 Study Plan and for coordinating the implementation of the projects outlined in that Study
Plan. Jeanne also played a large role with reaching out to stakeholders for input that informed the
development of the reports and publications resulting from Study Plan projects, and the
development of this assessment.
Other individuals who have made this assessment report possible and have not been previously
mentioned include: Jessica Agatstein, Adam Banasiak, Tom Beneke, Amy Bergdale, Ann Calamai,
Amy Clark, Brian Devir, Dayna Gibbons, Chris Grulke, Seth Haines, H. Jason Harmon, Cheryl Itkin,
Maureen Johnson, Randy Lamdin, Audrey Levine, Jordan Macknick, Kelsey Maloney, Lisa Matthews,
Angela McFadden, Connie Meacham, Marc Morandi, Jean-Philippe Nicot, Jennifer Orme-Zavaleta,
Nancy Parrotta, Robert Puis, Bridget R. Scanlon, Ayn Schmit, Cynthia Sonich-Mullen, Vicki Soto,
Inthirany Thillainadarajah, Vincent Tidwe 11, Martha Walters, and Steve Watkins.
Contract support was provided by The Cadmus Group, Inc. under contracts EP-C-08-015 and EP-C-
15-022 and by Neptune & Co., Inc. under contract EP-C-13-022. Authors and contributors included
student service contractors to USEPA: Kevin Blackwood (Contract EP-13-C-000133); Maryam Cluff
(Contract EP-13-H-000438); Ashley McElmury (Contract EP-12-C-000025); Claudia Meza-Cuadra
(Contract EP-13-H-000054); Alison Singer (Contract EP-13-H-000474); Sarah Solomon (Contract
EP-D-15-003); Jessica Wilhelm (Contract EP-D-15-003); Jie Xu (Contract EP-13-C-00120); Liabeth
Yohannes (Contract EP-14-H-000455). Kay Pinley was supported under the Senior Environmental
Employment Program under agreement CQ-835363 with NCCBA.
xxxi
-------
This page is intentionally left blank.
xxxii
-------
Executive Summary
Executive Summary
ES-l
-------
Executive Summary
This page is intentionally left blank.
ES-2
-------
Executive Summary
Executive Summary
People rely on clean and plentiful water resources to meet their basic needs, including drinking,
bathing, and cooking. In the early 2000s, members of the public began to raise concerns about
potential impacts on their drinking water from hydraulic fracturing at nearby oil and gas
production wells. In response to these concerns, Congress urged the U.S. Environmental Protection
Agency (EPA) to study the relationship between hydraulic fracturing for oil and gas and drinking
water in the United States.
The goals of the study were to assess the potential for activities in the hydraulic fracturing water
cycle to impact the quality or quantity of drinking water resources and to identify factors that affect
the frequency or severity of those impacts. To achieve these goals, the EPA conducted independent
research, engaged stakeholders through technical workshops and roundtables, and reviewed
approximately 1,200 cited sources of data and information. The data and information gathered
through these efforts served as the basis for this report, which represents the culmination of the
EPA's study of the potential impacts of hydraulic fracturing for oil and gas on drinking water
resources.
The hydraulic fracturing water cycle describes the use of water in hydraulic fracturing, from water
withdrawals to make hydraulic fracturing fluids, through the mixing and injection of hydraulic
fracturing fluids in oil and gas production wells, to the collection and disposal or reuse of produced
water. These activities can impact drinking water resources under some circumstances. Impacts
can range in frequency and severity, depending on the combination of hydraulic fracturing water
cycle activities and local- or regional-scale factors. The following combinations of activities and
factors are more likely than others to result in more frequent or more severe impacts:
• Water withdrawals for hydraulic fracturing in times or areas of low water availability,
particularly in areas with limited or declining groundwater resources;
• Spills during the management of hydraulic fracturing fluids and chemicals or produced
water that result in large volumes or high concentrations of chemicals reaching
groundwater resources;
• Injection of hydraulic fracturing fluids into wells with inadequate mechanical integrity,
allowing gases or liquids to move to groundwater resources;
• Injection of hydraulic fracturing fluids directly into groundwater resources;
• Discharge of inadequately treated hydraulic fracturing wastewater to surface water
resources; and
• Disposal or storage of hydraulic fracturing wastewater in unlined pits, resulting in
contamination of groundwater resources.
The above conclusions are based on cases of identified impacts and other data, information, and
analyses presented in this report. Cases of impacts were identified for all stages of the hydraulic
fracturing water cycle. Identified impacts generally occurred near hydraulically fractured oil and
ES-3
-------
Executive Summary
gas production wells and ranged in severity, from temporary changes in water quality to
contamination that made private drinking water wells unusable.
The available data and information allowed us to qualitatively describe factors that affect the
frequency or severity of impacts at the local level. However, significant data gaps and uncertainties
in the available data prevented us from calculating or estimating the national frequency of impacts
on drinking water resources from activities in the hydraulic fracturing water cycle. The data gaps
and uncertainties described in this report also precluded a full characterization of the severity of
impacts.
The scientific information in this report can help inform decisions by federal, state, tribal, and local
officials; industry; and communities. In the short-term, attention could be focused on the
combinations of activities and factors outlined above. In the longer-term, attention could be focused
on reducing the data gaps and uncertainties identified in this report Through these efforts, current
and future drinking water resources can be better protected in areas where hydraulic fracturing is
occurring or being considered.
Drinking Water Resources in the United States
In this report, drinking water resources are defined as any water that now serves, or in the future
could serve, as a source of drinking water for public or private use. This includes both surface water
resources and groundwater resources (Text Box ES-1). In 2010, approximately 58% of the total
volume of water withdrawn for public and non-public water supplies came from surface water
resources and approximately 42% came from groundwater resources (Maupin etal.. 2014).1 Most
people (86% of the population) in the United States relied on public water supplies for their
drinking water in 2010, and approximately 14% of the population obtained drinking water from
non-public water supplies. Non-public water supplies are often private water wells that supply
drinking water to a residence.
Future access to high-quality drinking water in the United States will likely be affected by changes
in climate and water use. Since 2000, about 30% of the total area of the contiguous United States
has experienced moderate drought conditions and about 20% has experienced severe drought
conditions. Declines in surface water resources have led to increased withdrawals and net
depletions of groundwater in some areas. As a result, non-fresh water resources (e.g., wastewater
from sewage treatment plants, brackish groundwater and surface water, and seawater) are
increasingly treated and used to meet drinking water demand.
Natural processes and human activities can affect the quality and quantity of current and future
drinking water resources. This report focuses on the potential for activities in the hydraulic
fracturing water cycle to impact drinking water resources; other processes or activities are not
discussed.
1 Public water systems provide water for human consumption from surface or groundwater through pipes or other
infrastructure to at least 15 service connections or serve an average of at least 25 people for at least 60 days a year. Non-
public water systems have fewer than 15 service connections and serve fewer than 25 individuals.
ES-4
-------
Executive Summaiy
Text Box ES-1. Drinking Water Resources.
In this report, drinking water resources are considered to be any water that now serves, or in the future could
serve, as a source of drinking water for public or private use. This includes both surface water bodies and
underground rock formations that contain water.
Surface water resources include water bodies located on the surface of the Earth. Rivers, springs, lakes, and
reservoirs are examples of surface water resources. Water quality and quantity are often considered when
determining whether a surface water resource could be used as a drinking water resource.
Groundwater resources are underground rock formations that contain water. Groundwater resources are found at
different depths nearly everywhere in the United States. Resource depth, water quality, and water yield are often
considered when determining whether a groundwater resource could be used as a drinking water resource.
Hydraulic Fracturing for Oil and Gas in the United States
Hydraulic fracturing is frequently used to enhance oil and gas production from underground rock
formations and is one of many activities that occur during the life of an oil and gas production well
(Figure ES-1). During hydraulic fracturing, hydraulic fracturing fluid is injected down an oil or gas
production well and into the targeted rock formation under pressures great enough to fracture the
oil- and gas-bearing rock.1 The hydraulic fracturing fluid usually carries proppant (typically sand)
into the newly-created fractures to keep the fractures "propped" open. After hydraulic fracturing,
oil, gas, and other fluids flow through the fractures and up the production well to the surface, where
they are collected and managed.
1 The targeted rock formation (sometimes called the 'target zone" or "production zone") is the portion of a subsurface
rock formation that contains the oil or gas to be extracted.
ES-5
-------
Executive Summary
Site assessment Well drilling Hydraulic
and preparation and construction fracturing
Produced water
handling and
wastewater
disposal/reuse
Oil and gas
production
Site and well
closure
Figure ES-1. General timeline and summary of activities at a hydraulically fractured oil or gas
production well.
Hydraulically fractured oil and gas production wells have significantly contributed to the surge in
domestic oil and gas production, accounting for slightly more than 50% of oil production and nearly
70% of gas production in 2015 (EIA. 2016c. d). The surge occurred when hydraulic fracturing was
combined with directional drilling technologies around 2000. Directional drilling allows oil and gas
production wells to be drilled horizontally or directionally along the targeted rock formation,
exposing more of the oil- or gas-bearing rock formation to the production well. When combined
with directional drilling technologies, hydraulic fracturing expanded oil and gas production to oil-
and gas-bearing rock formations previously considered uneconomical. Although hydraulic
fracturing is commonly associated with oil and gas production from deep, horizontal wells drilled
into shale (e.g., the Marcellus Shale in Pennsylvania or the Bakken Shale in North Dakota), it has
been used in a variety of oil and gas production wells (Text Box ES-2) and other types of oil- or gas-
bearing rock (e.g., sandstone, carbonate, and coal).
Approximately 1 million wells have been hydraulically fractured since the technique was first
developed in the late 1940s fGallegos and Varela. 2015: IOGCC. 20021. Roughly one third of those
wells were hydraulically fractured between 2000 and approximately 2014. Wells hydraulically
fractured between 2000 and 2013 were located in pockets of activity across the United States
(Figure ES-2). Based on several different data compilations, we estimate that 25,000 to 30,000 new
wells were drilled and hydraulically fractured in the United States each year between 2011 and
2014, in addition to existing wells that were hydraulically fractured to increase production.1
Following the decline in oil and gas prices, the number of new wells drilled and hydraulically
fractured appears to have decreased, with about 20,000 new wells drilled and hydraulically
fractured in 2015.
1 See Table 3-1 in Chapter 3.
ES-6
-------
Executive Summaiy
Text Box ES-2. Hydraulically Fractured Oil arid Gas Production Wells.
Hydraulically fractured oil and gas production wells come in different shapes and sizes. They can have different
depths, orientations, and construction characteristics. They can include new wells (i.e., wells that are hydraulically
fractured soon after construction) and old wells (i.e., wells that are hydraulically fractured after producing oil and
gas for some time).
Well Depth
Wells can be relatively shallow or relatively deep, depending
on the depth of the targeted rock formation.
Production Well
Ground Surface
Targeted Rock Formation
Milam County, Texas
Well depth = 685feet
San Augustine County, Texas
Well depth = 19,349 feet
Targeted Rock Formation
Well depths and locations from FracFocus.org.
Well Orientation
Wells can be vertical, horizontal, or deviated.
Vertical
Horizontal
Deviated
Well Construction Characteristics
Wells are typically constructed using multiple layers of casing and cement. The subsurface environment, state and federal
regulations, and industry experience and practices influence the number and placement of casing and cement.
Ground Surface
Protected
Groundwater
Casing >
Cement »-
Targeted Rock
Formation
•Conductor*
»Surface •
Drilled Hole ¦
- Production-
P
I
Conductor-
- Surface-
¦ Intermediate-
¦ Production-
Well diagrams are not to scale.
Conductor, surface, and production casings
Conductor, surface, intermediate, and
production casings
Oil and Gas Production Well Dictionary
Casing
Cement
Conductor casing
Intermediate casing
Production casing
Surface casing
Targeted rock formation
Steel pipe that extends from the ground surface to the bottom of the drilled hole
A slurry that hardens around the outside of the casing; cement fills the space between casings or
between a casing and the drilled hole and provides support for the casing
Casing that prevents the in-fill of dirt and rock in the uppermost few feet of drilled hole
Casing that seals off intermediate rock formations that may have different pressures than
deeper or shallower rock formations
Casing that transports fluids up and down the well
Casing that seals off groundwater resources that are identified as drinking water or useable
The part of a rock formation that contains the oil and/or gas to be extracted
ES-7
-------
Hydraulically Fractured Well Locations
~ 2000-2013 Oil and Gas Wells
Source data credits: Drillinglnfo, Inc.
Basemap Credits: U.S. Census Bureau, Esri, DeLorme, GEBCO, NOAA
NGDC, and other contributors
Figure ES-2. Locations of approximately 275,000 wells that were drilled and likely
hydraulically fractured between 2000 and 2013.
Data from Drillinglnfo (2014a).
Executive Summary
Hydraulically fractured oil and gas production wells can be located near or within sources of
drinking water. Between 2000 and 2013, approximately 3,900 public water systems were
estimated to have had at least one hydraulically fractured well within 1 mile of their water source;
these public water systems served more than 8.6 million people year-round in 2013. An additional
3.6 million people were estimated to have obtained drinking water from non-public water supplies
in counties with at least one hydraulically fractured well.1 Underground, hydraulic fracturing can
occur in close vertical proximity to drinking water resources. In some parts of the United States
(e.g., the Powder River Basin in Montana and Wyoming), there is no vertical distance between the
top of the hydraulically fractured oil- or gas-bearing rock formation and the bottom of treatable
water, as determined by data from state oil and gas agencies and state geological survey data.2 In
other parts of the country (e.g., the Eagle Ford Shale in Texas], there can be thousands of feet of
1 This estimate only includes counties in which 30% or more of the population (i.e., two or more times the national
average] relied on non-public water supplies in 2010. See Section 2.5 in Chapter 2.
2 In these cases, water that is naturally found in the oil- and gas-bearing rock formation meets the definition of drinking
water in some parts of the basin. See Section 6.3.2 in Chapter 6.
ES-8
-------
Executive Summary
rock that separate treatable water from the hydraulically fractured oil- or gas-bearing rock
formation. When hydraulically fractured oil and gas production wells are located near or within
drinking water resources, there is a greater potential for activities in the hydraulic fracturing water
cycle to impact those resources.
Approach: The Hydraulic Fracturing Water Cycle
The EPA studied the relationship between hydraulic fracturing for oil and gas and drinking water
resources using the hydraulic fracturing water cycle (Figure ES-3). The hydraulic fracturing water
cycle has five stages; each stage is defined by an activity involving water that supports hydraulic
fracturing. The stages and activities of the hydraulic fracturing water cycle include:
• Water Acquisition: the withdrawal of groundwater or surface water to make hydraulic
fracturing fluids;
• Chemical Mixing: the mixing of a base fluid (typically water), proppant, and additives at
the well site to create hydraulic fracturing fluids;1
• Well Injection: the injection and movement of hydraulic fracturing fluids through the oil
and gas production well and in the targeted rock formation;
• Produced Water Handling: the on-site collection and handling of water that returns to
the surface after hydraulic fracturing and the transportation of that water for disposal or
reuse;2 and
• Wastewater Disposal and Reuse: the disposal and reuse of hydraulic fracturing
wastewater.3
Potential impacts on drinking water resources from the above activities are considered in this
report. We do not address other concerns that have been raised by stakeholders about hydraulic
fracturing (e.g., potential air quality impacts or induced seismicity) or other oil and gas exploration
and production activities (e.g., environmental impacts from site selection and development), as
these were not included in the scope of the study. Additionally, this report is not a human health
risk assessment; it does not identify populations exposed to hydraulic fracturing-related chemicals,
and it does not estimate the extent of exposure or estimate the incidence of human health impacts.
1A base fluid is the fluid into which proppants and additives are mixed to make a hydraulic fracturing fluid; water is an
example of a base fluid. Additives are chemicals or mixtures of chemicals that are added to the base fluid to change its
properties.
2 "Produced water" is defined in this report as water that flows from and through oil and gas wells to the surface as a by-
product of oil and gas production.
3 "Hydraulic fracturing wastewater" is defined in this report as produced water from hydraulically fractured oil and gas
wells that is being managed using practices that include, but are not limited to, injection in Class II wells, reuse in other
hydraulic fracturing operations, and various aboveground disposal practices. The term "wastewater" is being used as a
general description of certain waters and is not intended to constitute a term of art for legal or regulatory purposes. Class
II wells are used to inject wastewater associated with oil and gas production underground and are regulated under the
Underground Injection Control Program of the Safe Drinking Water Act.
ES-9
-------
Executive Summary
Well Injection
Chemical Mixing
Produced Water Handling
Water Acquisition
Wastewater Disposal and Reuse
Figure ES-3. The five stages of the hydraulic fracturing water cycle.
The stages (shown in the insets) identify activities involving water that support hydraulic fracturing for oil and gas.
Activities may take place in the same watershed or different watersheds and close to or far from drinking water
resources. Thin arrows in the insets depict the movement of water and chemicals. Specific activities in the
"Wastewater Disposal and Reuse" inset include (a) disposal of wastewater through underground injection, (b)
wastewater treatment followed by reuse in other hydraulic fracturing operations or discharge to surface waters,
and (c) disposal through evaporation or percolation pits.
Each stage of the hydraulic fracturing water cycle was assessed to identify [1] the potential for
impacts on drinking water resources and (2) factors that affect the frequency or severity of impacts.
Specific definitions used in this report are provided below:
• An impact is any change in the quality or quantity of drinking water resources, regardless
of severity, that results from an activity in the hydraulic fracturing water cycle.
• A factor is a feature of hydraulic fracturing operations or an environmental condition that
affects the frequency or severity of impacts.
• Frequency is the number of impacts per a given unit (e.g., geographic area, unit of time,
number of hydraulically fractured wells, or number of water bodies).
• Severity is the magnitude of change in the quality or quantity of a drinking water resource
as measured by a given metric (e.g., duration, spatial extent, or contaminant
concentration).
ES-10
-------
Executive Summary
Factors affecting the frequency or severity of impacts were identified because they describe
conditions under which impacts are more or less likely to occur and because they could inform the
development of future strategies and actions to prevent or reduce impacts. Although no attempt
was made to identify or evaluate best practices, ways to reduce the frequency or severity of impacts
from activities in the hydraulic fracturing water cycle are described in this report when they were
reported in the scientific literature. Laws, regulations, and policies also exist to protect drinking
water resources, but a comprehensive summary and broad evaluation of current or proposed
regulations and policies was beyond the scope of this report.
Relevant scientific literature and data were evaluated for each stage of the hydraulic fracturing
water cycle. Literature included articles published in science and engineering journals, federal and
state government reports, non-governmental organization reports, and industry publications. Data
sources included federal- and state-collected data sets, databases maintained by federal and state
government agencies, other publicly available data, and industry data provided to the EPA.1 The
relevant literature and data complement research conducted by the EPA under its Plan to Study the
Potential Impacts of Hydraulic Fracturing on Drinking Water Resources (Text Box ES-3).
Text Box ES-3. The EPA's Study of the Potential Impacts of Hydraulic Fracturing for Oil and
Gas on Drinking Water Resources.
The EPA's study is the first national study of the potential impacts of hydraulic fracturing for oil and gas on drinking
water resources. It included independent research projects conducted by EPA scientists and contractors and a
state-of-the-science assessment of available data and information on the relationship between hydraulic fracturing
and drinking water resources (i.e., this report).
Public Comments
Public Meetings
Scientific
Literature
Science
Advisory Board
Existing Data
Science Advisory Board
Technical Workshops
and Roundtables
Existing Data
This Report
EPA Research Projects
Study of the Potential Impacts of Hydraulic Fracturing for Oil and Gas on Drinking Water Resources
Scientific Literature
Throughout the study, the EPA consulted with the Agency's independent Science Advisory Board (SAB) on the
scope of the study and the progress made on the research projects. The SAB also conducted a peer review of both
the Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources (U.S. EPA, 2011d;
referred to as the Study Plan in this report) and a draft of this report.
Stakeholder engagement also played an important role in the development and implementation of the study.
While developing the scope of the study, the EPA held public meetings to get input from stakeholders on the study
scope and design. While conducting the study, the EPA requested information from the public and engaged with
technical, subject-matter experts on topics relevant to the study in a series of technical workshops and
roundtables. For more information on the EPA's study, including the role of the SAB and stakeholders, visit
www.epa.gov/hfstudy.
1 Industry data was provided to the EPA in response to two separate information requests to oil and gas service
companies and oil and gas production well operators. Some of these data were claimed as confidential business
information under the Toxic Substances Control Act and were treated as such in this report.
ES-11
-------
Executive Summary
A draft of this report underwent peer review by the EPA's Science Advisory Board (SAB). The SAB is
an independent federal advisory committee that often conducts peer reviews of high-profile
scientific matters relevant to the EPA. Members of the SAB and ad hoc panels formed under the
auspices of the SAB are nominated by the public and selected based on factors such as technical
expertise, knowledge, experience, and absence of any real or perceived conflicts of interest. Peer
review comments provided by the SAB and public comments submitted to the SAB during their
peer review, including comments on major conclusions and technical content, were carefully
considered in the development of this final document.
A summary of the activities in the hydraulic fracturing water cycle and their potential to impact
drinking water resources is provided below, including what is known about human health hazards
associated with chemicals identified across all stages of the hydraulic fracturing water cycle.
Additional details are available in the full report
Water Acquisition
Activity: The withdrawal of groundwater or surface water to make hydraulic fracturing fluids.
Relationship to Drinking Water Resources: Groundwater and surface water resources that
provide water for hydraulic fracturing fluids can also provide drinking water for public or non-
public water supplies.
Water is the major component of nearly all hydraulic fracturing fluids, typically making up 90-97%
of the total fluid volume injected into a well. The median volume of water used, per well, for
hydraulic fracturing was approximately 1.5 million gallons (5.7 million liters) between January
2011 and February 2013, as reported in FracFocus 1.0 (Text Box ES-4). There was wide variation in
the water volumes reported per well, with 10th and 90th percentiles of 74,000 gallons (280,000
liters) and 6 million gallons (23 million liters) per well, respectively. There was also variation in
water use per well within and among states (Table ES-1). This variation likely results from several
factors, including the type of well, the fracture design, and the type of hydraulic fracturing fluid
used. An analysis of hydraulic fracturing fluid data from Gallegos etal. f20151 indicates that water
volumes used per well have increased over time as more horizontal wells have been drilled.
Water used for hydraulic fracturing is typically fresh water taken from available groundwater
and/or surface water resources located near hydraulically fractured oil and gas production wells.
Water sources can vary across the United States, depending on regional or local water availability;
laws, regulations, and policies; and water management practices. Hydraulic fracturing operations in
the humid eastern United States generally rely on surface water resources, whereas operations in
the arid and semi-arid western United States generally rely on groundwater or surface water.
Geographic differences in water use for hydraulic fracturing are illustrated in Figure ES-4, which
shows that most of the water used for hydraulic fracturing in the Marcellus Shale region of the
Susquehanna River Basin came from surface water resources between approximately 2008 and
2013. In comparison, less than half of the water used for hydraulic fracturing in the Barnett Shale
region of Texas came from surface water resources between approximately 2011 and 2013.
ES-12
-------
Executive Summary
Text Box ES-4. FracFocus Chemical Disclosure Registry.
The FracFocus Chemical Disclosure Registry is a publicly-accessible website (www.fracfocus.org) managed by the
Ground Water Protection Council (GWPC) and the Interstate Oil and Gas Compact Commission (IOGCC). Oil and gas
production well operators can disclose information at this website about water and chemicals used in hydraulic
fracturing fluids at individual wells. In many states where oil and gas production occurs, well operators are
required to disclose to FracFocus well-specific information on water and chemical use during hydraulic fracturing.
The GWPC and the IOGCC provided the EPA with over 39,000 PDF disclosures submitted by well operators to
FracFocus (version 1.0) before March 1, 2013. Data in the disclosures were extracted and compiled in a project
database, which was used to conduct analyses on water and chemical use for hydraulic fracturing. Analyses were
conducted on over 38,000 unique disclosures for wells located in 20 states that were hydraulically fractured
between January 1, 2011, and February 28, 2013.
Despite the challenge of adapting a dataset originally created for local use and single-PDF viewing to answer
broader questions, the project database created by the EPA provided substantial insight into water and chemical
use for hydraulic fracturing. The project database represents the data reported to FracFocus 1.0 rather than all
hydraulic fracturing that occurred in the United States during the study time period. The project database is an
incomplete picture of all hydraulic fracturing due to voluntary reporting in some states for certain time periods (in
the absence of state reporting requirements), the omission of information on confidential chemicals from
disclosures, and invalid or erroneous information in the original disclosures or created during the development of
the database. The development of FracFocus 2.0, which became the exclusive reporting mechanism in June 2013,
was intended to increase the quality, completeness, and consistency of the data submitted by providing
dropdown menus, warning and error messages during submission, and automatic formatting of certain fields. The
GWPC has announced additional changes and upgrades for FracFocus 3.0 to enhance data searchability, increase
system security, provide greater data accuracy, and further increase data transparency.
Table ES-1. Water use per hydraulically fractured well between January 2011 and February 2013.
Medians and percentiles were calculated from data submitted to FracFocus 1.0 (Appendix B).
State
Number of FracFocus
1.0 disclosures
Median volume per
well (gallons)
10th percentile
(gallons)
90th percentile
(gallons)
Arkansas
1,423
5,259,965
3,234,963
7,121,249
California
711
76,818
21,462
285,306
Colorado
4,898
463,462
147,353
3,092,024
Kansas
121
1,453,788
10,836
2,227,926
Louisiana
966
5,077,863
1,812,099
7,945,630
Montana
207
1,455,757
367,326
2,997,552
New Mexico
1,145
175,241
35,638
1,871,666
North Dakota
2,109
2,022,380
969,380
3,313,482
Ohio
146
3,887,499
2,885,568
5,571,027
Oklahoma
1,783
2,591,778
1,260,906
7,402,230
Pennsylvania
2,445
4,184,936
2,313,649
6,615,981
Texas
16,882
1,420,613
58,709
6,115,195
Utah
1,406
302,075
76,286
769,360
West Virginia
273
5,012,238
3,170,210
7,297,080
Wyoming
1,405
322,793
5,727
1,837,602
ES-13
-------
Executive Summary
Hydraulic fracturing wastewater and other lower-quality water can also be used in hydraulic
fracturing fluids to offset the need for fresh water, although the proportion of injected fluid that is
reused hydraulic fracturing wastewater varies by location (Figure ES-4).1 Overall, the proportion of
(a) Marcellus Shale,
Susquehanna River Basin
4.1-4,6 million gallons
injected
420,000-1,3 million gallons
produced
79%
Surface Water ¦ Groundwater
Reused hydraulic fracturing wastewater
Reuse in hydraulic fracturing
¦ Class II well
'Less than approximately 1% is treated at facilities that are
either permitted to discharge to surface water or whose
discharge status is uncertain.
Most of the injected fluid stays in the subsurface; produced
water volumes aver 10 years are approximately 10-30% of
t/ie injected fluid volume.
(b) Barnett Shale, Texas
3,9-4.5 million gallons
injected
3.9-4.5 million gallons
produced
Surface Water ¦ Groundwater
Reused hydraulic fracturing wastewater
Reuse in hydraulic fracturing
¦ Class II well
Produced water volumes over three years can be
approximately the same as the injected fluid volume.
Figure ES-4, Water budgets illustrative of hydraulic fracturing water management practices in
the Marcellus Shale in the Susquehanna River Basin between approximately 200S and 2013
and the Barnett Shale in Texas between approximately 2011 and 2013.
Class II wells are used to inject wastewater associated with oil and gas production underground and are regulated
under the Underground Injection Control Program of the Safe Drinking Water Act. Data sources are described in
Figure 10-1 in Chapter 10.
1 Reused hydraulic fracturing wastewater as a percentage of injected fluid differs from the percentage of produced water
that is managed through reuse in other hydraulic fracturing operations. For example, in the Marcellus Shale region of the
Susquehanna River Basin, approximately 14% of injected fluid was reused hydraulic fracturing wastewater, while
approximately 90% of produced water was managed through reuse in other hydraulic fracturing operations (Figure ES-
4a],
ES-14
-------
Executive Summary
water used in hydraulic fracturing that comes from reused hydraulic fracturing wastewater
appears to be low. In a survey of literature values from 10 states, basins, or plays, the median
percentage of the injected fluid volume that came from reused hydraulic fracturing wastewater was
5% between approximately 2008 and 2014.1 There was an increase in the reuse of hydraulic
fracturing wastewater as a percentage of the injected hydraulic fracturing fluid in both
Pennsylvania and West Virginia between approximately 2008 and 2014. This increase is likely due
to the limited availability of Class II wells, which are commonly used to dispose of oil and gas
wastewater, and the costs of trucking wastewater to Ohio, where Class II wells are more prevalent.2
Class II wells are also prevalent in Texas, and the reuse of wastewater in hydraulic fracturing fluids
in the Barnett Shale appears to be lower than in the Marcellus Shale (Figure ES-4).
Because the same water resource can be used to support hydraulic fracturing and to provide
drinking water, withdrawals for hydraulic fracturing can directly impact drinking water resources
by changing the quantity or quality of the remaining water. Although every water withdrawal
affects water quantity, we focused on water withdrawals that have the potential to significantly
impact drinking water resources by limiting the availability of drinking water or altering its quality.
Water withdrawals for a single hydraulically fractured oil and gas production well are not expected
to significantly impact drinking water resources, because the volume of water needed to
hydraulically fracture a single well is unlikely to limit the availability of drinking water or alter its
quality. If, however, multiple oil and gas production wells are located within an area, the total
volume of water needed to hydraulically fracture all of the wells has the potential to be a significant
portion of the water available and impacts on drinking water resources can occur.
To assess whether hydraulic fracturing operations are a relatively large or small user of water, we
compared water use for hydraulic fracturing to total water use at the county level (Text Box ES-5).
In most counties studied, the average annual water volumes reported in FracFocus 1.0 were
generally less than 1% of total water use. This suggests that hydraulic fracturing operations
represented a relatively small user of water in most counties. There were exceptions, however.
Average annual water volumes reported in FracFocus 1.0 were 10% or more of total water use in
26 of the 401 counties studied, 30% or more in nine counties, and 50% or more in four counties.3 In
these counties, hydraulic fracturing operations represented a relatively large user of water.
The above results suggest that hydraulic fracturing operations can significantly increase the volume
of water withdrawn in particular areas. Increased water withdrawals can result in significant
impacts on drinking water resources if there is insufficient water available in the area to
accommodate all users. To assess the potential for these impacts, we compared hydraulic fracturing
water use to estimates of water availability at the county level.4 In most counties studied, average
1 See Section 4.2 in Chapter 4.
2 See Chapter 8 for additional information on Class II wells.
3 Hydraulic fracturing water consumption estimates followed the same general pattern as the water use estimates
presented here, but with slightly larger percentages in each category (Section 4.4 in Chapter 4].
4 County-level water availability estimates were derived from the Tidwell etal. ("20131 estimates of water availability for
siting new thermoelectric power plants (see Text Box 4-2 in Chapter 4 for details]. The county-level water availability
estimates used in this report represent the portion of water available to new users within a county.
ES-15
-------
Executive Summary
Text Box ES-5. County-Level Water Use for Hydraulic Fracturing.
To assess whether hydraulic fracturing operations are a relatively large or small user of water, the average annual
water use for hydraulic fracturing in 2011 and 2012 was compared, at the county-level, to total water use in 2010.
For most counties studied, average annual water volumes reported for individual counties in FracFocus 1.0 were
less than 1% of total water use in those counties. But in some counties, hydraulic fracturing operations reported in
FracFocus 1.0 represented a relatively large user of water.
Examples of Water Use in Two Counties: Wilson County, Texas, and Mountrail County, North Dakota
Wilson County, Texas
44 wells reported In FracFocus 1.0
7,844
2010 Total Water Uset
154
1,872
106
85
Hydraulic
Fracturing*
Total t
4,833
Industrial use was 12 rrvJlian go/fans
Public Supply Irrigation
Domestic ¦ Livestock
¦ Industrial ¦ Mining
Depending on local water availability, hydraulic fracturing
water withdrawals may be less likely to significantly impact
drinking water resources under this kind of scenario.
Mountrail County, North Dakota
508 wells reported in FracFocus 1.0
1,248
449
I
2010 Total Water Uset
179
288
438
183
135
Hydraulic
Fracturing*
Totalt
Public Supply Irrigation
Domestic ¦ Livestock
I Industrial ¦ Mining
Depending on local water availability, hydraulic fracturing
water withdrawals may be more likely to significantly impact
drinking water resources under this kind of scenario.
•Hydraulic fracturing water use is a function of the water use per well and the total number of wells hydraulically fractured within a county. Average annual
water use for hydraulic fracturing was calculated at the county-level using data reported in FracFocus 1.0 in 2011 and 2012 (Appendix B).
~Tre U.S. Geological Survey compiles national water use estimates every five years in the National Water Census. Total water use at the county-level was
obtained from the most recent census, which was conducted in 2010 (Maupin et al., 2014).
2010 Total Water Use Categories
Public supply Water withdrawn by public and private water suppliers that provide water to at (east 25 people or
have a minimum of 15 connections
Domestic Self-supplied water withdrawals for indoor household purposes such as drinking, food preparation,
bathing, washing clothes and dishes, flushing toilets, and outdoor purposes such as watering lawns
and gardens
Industrial Water used for fabrication, processing, washing, and cooling
Irrigation Water that is applied by an irrigation system to assist crop and pasture growth or to maintain
vegetation on recreational lands (e.g., parks and golf courses)
Livestock Water used for livestock watering, feedlots, dairy operations, and other on-farm needs
Mining Water used for the extraction of naturally-occurring minerals, including solids (e.g., coal, sand, gravel,
and other ores), liquids {e.g., crude petroleum), and gases (e.g., natural gas)
annual water volumes reported for hydraulic fracturing were less than 1% of the estimated annual
volume of readily-available fresh water. However, average annual water volumes reported for
hydraulic fracturing were greater than the estimated annual volume of readily-available fresh
water in 17 counties in Texas. This analysis suggests that there was enough water available
annually to support the level of hydraulic fracturing reported to FracFocus 1.0 in most, but not all,
ES-16
-------
Executive Summary
areas of the country. This observation does not preclude the possibility of local impacts in other
areas of the country, nor does it indicate that local impacts have occurred or will occur in the 17
counties in Texas. To better understand whether local impacts have occurred, and the factors that
affect those impacts, local-level studies, such as the ones described below, are needed.
Local impacts on drinking water quantity have occurred in areas with increased hydraulic
fracturing activity. In 2011, for example, drinking water wells in an area overlying the Haynesville
Shale ran out of water due to higher than normal groundwater withdrawals and drought (LA
Ground Water Resources Commission. 2012). Water withdrawals for hydraulic fracturing
contributed to these conditions, along with other water users and the lack of precipitation.
Groundwater impacts have also been reported in Texas. In a detailed case study, Scanlon et al.
f2014bl estimated that groundwater levels in approximately 6% of the area studied dropped by
100 feet (31 meters) to 200 feet (61 meters) or more after hydraulic fracturing activity increased in
2009.
In contrast, studies in the Upper Colorado and Susquehanna River basins found minimal impacts on
drinking water resources from hydraulic fracturing. In the Upper Colorado River Basin, the EPA
found that high-quality water produced from oil and gas wells in the Piceance tight sands provided
nearly all of the water for hydraulic fracturing in the study area (U.S. EPA. 2015e). Due to this high
reuse rate, the EPA did not identify any locations in the study area where hydraulic fracturing
contributed to locally high water use. In the Susquehanna River Basin, multiple studies and state
reports have identified the potential for hydraulic fracturing water withdrawals in the Marcellus
Shale to impact surface water resources. Evidence suggests, however, that current water
management strategies, including passby flows and reuse of hydraulic fracturing wastewater, help
protect streams from depletion by hydraulic fracturing water withdrawals. A passby flow is a
prescribed, low-streamflow threshold below which water withdrawals are not allowed.
The above examples highlight factors that can affect the frequency or severity of impacts on
drinking water resources from hydraulic fracturing water withdrawals. In particular, areas of the
United States that rely on declining groundwater resources are vulnerable to more frequent and
more severe impacts from all water withdrawals, including withdrawals for hydraulic fracturing.
Extensive groundwater withdrawals can limit the availability of belowground drinking water
resources and can also change the quality of the water remaining in the resource. Because
groundwater recharge rates can be low, impacts can last for many years. Seasonal or long-term
drought can also make impacts more frequent and more severe for groundwater and surface water
resources. Hot, dry weather reduces or prevents groundwater recharge and depletes surface water
bodies, while water demand often increases simultaneously (e.g., for irrigation). This combination
of factors—high hydraulic fracturing water use and relatively low water availability due to
declining groundwater resources and/or frequent drought—was found to be present in southern
and western Texas.
Water management strategies can also affect the frequency and severity of impacts on drinking
water resources from hydraulic fracturing water withdrawals. These strategies include using
hydraulic fracturing wastewater or brackish groundwater for hydraulic fracturing, transitioning
from limited groundwater resources to more abundant surface water resources, and using passby
ES-17
-------
Executive Summary
flows to control water withdrawals from surface water resources. Examples of these water
management strategies can be found throughout the United States. In western and southern Texas,
for example, the use of brackish water is currently reducing impacts on fresh water sources, and
could, if increased, reduce future impacts. Louisiana and North Dakota have encouraged well
operators to withdraw water from surface water resources instead of high-quality groundwater
resources. And, as described above, the Susquehanna River Basin Commission limits surface water
withdrawals during periods of low stream flow.
Water Acquisition Conclusions
With notable exceptions, hydraulic fracturing uses a relatively small percentage of water when
compared to total water use and availability at large geographic scales. Despite this, hydraulic
fracturing water withdrawals can affect the quantity and quality of drinking water resources by
changing the balance between the demand on local water resources and the availability of those
resources. Changes that have the potential to limit the availability of drinking water or alter its
quality are more likely to occur in areas with relatively high hydraulic fracturing water withdrawals
and low water availability, particularly due to limited or declining groundwater resources. Water
management strategies (e.g., encouragement of alternative water sources or water withdrawal
restrictions) can reduce the frequency or severity of impacts on drinking water resources from
hydraulic fracturing water withdrawals.
Chemical Mixing
Activity: The mixing of a base fluid, proppant, and additives at the well site to create hydraulic
fracturing fluids.
Relationship to Drinking Water Resources: Spills of additives and hydraulic fracturing fluids can
reach groundwater and surface water resources.
Hydraulic fracturing fluids are engineered to create and grow fractures in the targeted rock
formation and to carry proppant through the oil and gas production well into the newly-created
fractures. Hydraulic fracturing fluids are typically made up of base fluids, proppant, and additives.
Base fluids make up the largest proportion of hydraulic fracturing fluids by volume. As illustrated in
Text Box ES-6, base fluids can be a single substance (e.g., water in the slickwater example) or can be
a mixture of substances (e.g., water and nitrogen in the energized fluid example). The EPA's analysis
of hydraulic fracturing fluid data reported to FracFocus 1.0 suggests that water was the most
commonly used base fluid between January 2011 and February 2013 (U.S. EPA. 2015a). Non-water
substances, such as gases and hydrocarbon liquids, were reported to be used alone or blended with
water to form a base fluid in fewer than 3% of wells in FracFocus 1.0.
Proppant makes up the second largest proportion of hydraulic fracturing fluids (Text Box ES-6).
Sand (i.e., quartz) was the most commonly reported proppant between January 2011 and February
2013, with 98% of wells in FracFocus 1.0 reporting sand as the proppant (U.S. EPA. 2015a). Other
ES-18
-------
Executive Summaiy
Text Box ES-6. Examples of Hydraulic Fracturing Fluids.
Hydraulic fracturing fluids are engineered to create and extend fractures in the targeted rock formation and to
carry proppant through the production well into the newly-created fractures. While there is no universal hydraulic
fracturing fluid, there are general types of hydraulic fracturing fluids. Two types of hydraulic fracturing fluids are
described below,
Slickwater
Slickwater hydraulic fracturing fluids are water-based fluids that generally contain a friction reducer. The friction
reducer makes it easier for the fluid to be pumped down the oil and gas production well at high rates. Slickwater is
commonly used to hydraulically fracture shale formations.
Total water volume = 4,763,000 gallons
Energized Fluid
Energized fluids are mixtures of liquids and gases. They can be used for hydraulic fracturing in under-pressured gas
formations.
Bradford County, Pennsylv;
Well depth = 7,255 feet
0.01% Friction Reducer (1)
16%* Reused
Wastewater
0.006% Biocide (3)
0.002% Scale Inhibitor (2)
0.0009% Iron
K Control (1)
0.0006% Corrosion
Inhibitor (5)
71% Fresh Water
0.05% Additives (13 Chemicals)
28%* Nitrogen (gas)
1.2% Clay
Control (1)
58% Water
0.004%
1.5% Additives (28 Chemicals)
0.08% Surfactant (3)
0.05% Foamer (2)
, 0.03% Corrosion
Inhibitor (11)
0.03% Biocide (4)
0.01% Friction
Reducer(1)
0.008% Breaker (1)
0.006% Scale
Inhibitor (4)
Rio Arriba County, New Mexico
Well depth = 7,640 feet
Total water volume = 105,000 gallons
*Maximum percent by mass of the total hydraulic fracturing fluid. Data obtained from FracFocus.org
0.004% Iron Control (1)
Additive Dictionary
Acid
Biocide
Breaker
Clay control
Corrosion inhibitor
Foamer
Friction reducer
Iron control
Scale inhibitor
Surfactant
Dissolves minerals and creates pre-fractures in the rock
Controls or eliminates bacteria in the hydraulic fracturing fluid
Reduces the thickness of the hydraulic fracturing fluid
Prevents swelling and migration of formation clays
Protects iron and steel equipment from rusting
Creates a foam hydraulic fracturing fluid
Reduces friction between the hydraulic fracturing fluid and pipes during pumping
Prevents the precipitation of iron-containing chemicals
Prevents the formation of scale buildup within the well
Reduces the surface tension of the hydraulic fracturing fluid
ES-19
-------
Executive Summary
proppants can include man-made or specially engineered particles, such as high-strength ceramic
materials or sintered bauxite.1
Additives generally make up the smallest proportion of the overall composition of hydraulic
fracturing fluids (Text Box ES-6), yet have the greatest potential to impact the quality of drinking
water resources compared to proppant and base fluids. Additives, which can be a single chemical or
a mixture of chemicals, are added to the base fluid to change its properties (e.g., adjust pH, increase
fluid thickness, or limit bacterial growth). The choice of which additives to use depends on the
characteristics of the targeted rock formation (e.g., rock type, temperature, and pressure), the
economics and availability of desired additives, and well operator or service company preferences
and experience.
The variability of additives, both in their purpose and chemical composition, suggests that a large
number of different chemicals may be used in hydraulic fracturing fluids across the United States.
The EPA identified 1,084 chemicals that were reported to have been used in hydraulic fracturing
fluids between 2005 and 2013.2 3 The EPA's analysis of FracFocus 1.0 data indicates that between 4
and 28 chemicals were used per well between January 2011 and February 2013 and that no single
chemical was used in all wells (U.S. EPA. 2015a). Three chemicals—methanol, hydrotreated light
petroleum distillates, and hydrochloric acid—were reported in 65% or more of the wells in
FracFocus 1.0; 35 chemicals were reported in at least 10% of the wells (Table ES-2).
Table ES-2. Chemicals reported in 10% or more of disclosures in FracFocus 1.0.
Disclosures provided information on chemicals used at individual well sites between January 1, 2011, and February
28, 2013.
Chemical Name (CASRN)8
Percent of FracFocus 1.0 disclosures'3
Methanol (67-56-1)
72
Hydrotreated light petroleum distillates (64742-47-8)
65
Hydrochloric acid (7647-01-0)
65
Water (7732-18-5)°
48
Isopropanol (67-63-0)
47
Ethylene glycol (107-21-1)
46
Peroxydisulfuric acid, diammonium salt (7727-54-0)
44
Sodium hydroxide (1310-73-2)
39
Guar gum (9000-30-0)
37
1 Sintered bauxite is crushed and powdered bauxite that is fused into spherical beads at high temperatures.
2 This list includes 1,084 unique Chemical Abstracts Service Registration Numbers (CASRNs], which can be assigned to a
single chemical (e.g., hydrochloric acid] or a mixture of chemicals (e.g., hydrotreated light petroleum distillates].
Throughout this report, we refer to the substances identified by unique CASRNs as "chemicals."
3 Davalu and Konschnik (2016] identified 995 unique CASRNs from data submitted to FracFocus between March 9,2011,
and April 13,2015. Two hundred sixty-three of these CASRNs are not on the list of unique CASRNs identified by the EPA
(Appendix H). Only one of the 263 chemicals was reported at greater than 1% of wells, which suggests that these
chemicals were used at only a few sites.
ES-20
-------
Executive Summary
Chemical Name (CASRN)"
Percent of FracFocus 1.0 disclosures'3
Quartz (14808-60-7)°
36
Glutaraldehyde (111-30-8)
34
Propargyl alcohol (107-19-7)
33
Potassium hydroxide (1310-58-3)
29
Ethanol (64-17-5)
29
Acetic acid (64-19-7)
24
Citric acid (77-92-9)
24
2-Butoxyethanol (111-76-2)
21
Sodium chloride (7647-14-5)
21
Solvent naphtha, petroleum, heavy aromatic (64742-94-5)
21
Naphthalene (91-20-3)
19
2,2-Dibromo-3-nitrilopropionamide (10222-01-2)
16
Phenolic resin (9003-35-4)
14
Choline chloride (67-48-1)
14
Methenamine (100-97-0)
14
Carbonic acid, dipotassium salt (584-08-7)
13
1,2,4-Trimethylbenzene (95-63-6)
13
Quaternary ammonium compounds, benzyl-C12-16-alkyldimethyl,
chlorides (68424-85-1)
12
Poly(oxy-l,2-ethanediyl)-nonylphenyl- hydroxy (mixture)
(127087-87-0)
12
Formic acid (64-18-6)
12
Sodium chlorite (7758-19-2)
11
Nonyl phenol ethoxylate (9016-45-9)
11
Tetrakis(hydroxymethyl)phosphonium sulfate (55566-30-8)
11
Polyethylene glycol (25322-68-3)
11
Ammonium chloride (12125-02-9)
10
Sodium persulfate (7775-27-1)
10
a "Chemical" refers to chemical substances with a single CASRN; these may be pure chemicals (e.g., methanol) or chemical
mixtures (e.g., hydrotreated light petroleum distillates).
b Analysis considered 34,675 disclosures that met selected quality assurance criteria. See Table 5-2 in Chapter 5.
c Quartz and water were reported as ingredients in additives, in addition to proppants and base fluids.
ES-21
-------
Executive Summary
Concentrated additives are delivered to the well site and stored until they are mixed with the base
fluid and proppant and pumped down the oil and gas production well (Text Box ES-7). While the
overall concentration of additives in hydraulic fracturing fluids is generally small (typically 2% or
less of the total volume of the injected fluid), the total volume of additives delivered to the well site
can be large. Because over 1 million gallons (3.8 million liters) of hydraulic fracturing fluid are
generally injected per well, thousands of gallons of additives can be stored on site and used during
hydraulic fracturing.
As illustrated in Text Box ES-7, additives are often stored in multiple, closed containers [typically
200 gallons (760 liters) to 375 gallons (1,420 liters) per container] and moved around the site in
hoses and tubing. This equipment is designed to contain additives and blended hydraulic fracturing
fluid, but spills can occur. Changes in drinking water quality can occur if spilled fluids reach
groundwater or surface water resources.
Several studies have documented spills of hydraulic fracturing fluids or additives. Nearly all of these
studies identified spills from state-managed spill databases. Data gathered for these studies suggest
that spills of hydraulic fracturing fluids or additives were primarily caused by equipment failure or
human error. For example, an EPA analysis of spill reports from nine state agencies, nine oil and gas
well operators, and nine hydraulic fracturing service companies characterized 151 spills of
hydraulic fracturing fluids or additives on or near well sites in 11 states between January 2006 and
April 2012 fU.S. EPA. 2015ml. These spills were primarily caused by equipment failure (34% of the
spills) or human error (25%), and more than 30% of the spills were from fluid storage units (e.g.,
tanks, totes, and trailers). Similarly, a study of spills reported to the Colorado Oil and Gas
Conservation Commission identified 125 spills during well stimulation (i.e., a part of the life of an
oil and gas well that often, but not always, includes hydraulic fracturing) between January 2010 and
August 2013 fCOGCC. 20141. Of these spills, 51% were caused by human error and 46% were due
to equipment failure.
Studies of spills of hydraulic fracturing fluids or additives provide insights on spill volumes, but
little information on chemical-specific spill composition. Among the 151 spills characterized by the
EPA, the median volume of fluid spilled was 420 gallons (1,600 liters), although the volumes spilled
ranged from 5 gallons (19 liters) to 19,320 gallons (73,130 liters). Spilled fluids were often
described as acids, biocides, friction reducers, crosslinkers, gels, and blended hydraulic fracturing
fluid, but few specific chemicals were mentioned.1 Considine etal. f20121 identified spills related to
oil and gas development in the Marcellus Shale that occurred between January 2008 and August
2011 from Notices of Violations issued by the Pennsylvania Department of Environmental
Protection. The authors identified spills greater than 400 gallons (1,500 liters) and spills less than
400 gallons (1,500 liters).
1A crosslinker is an additive that increases the thickness of gelled fluids by connecting polymer molecules in the gelled
fluid.
ES-22
-------
Executive Summaiy
Chemical Mixing Equipment Dictionary
Blender
Chemical additive unit
Flowback tanks
Frac head
High pressure pumps
Hydration unit
Manifold
Proppant
Water tanks
Blends water, proppant, and additives
Transports additives to the site and stores additives onsite
Stores liquid that returns to the surface after hydraulic fracturing
Connects hydraulic fracturing equipment to the production well
Pressurize mixed fluids before injection into the production well
Creates and stores gels used in some hydraulic fracturing fluids
Transfers fluids from the blender to the frac head
Stores proppant (often sand)
Stores water
Text Box ES-7. Chemical Mixing Equipment.
Water Tanks
Manifold
Chemical
Additive Units
Frac Head
Source: Schlumberger
Water Tanks
High Pressure Pumps
Flowback
Tanks
/\/ /V./ r
High Pressure Pumps
Frac
Head
Typical Layout of Chemical Mixing Equipment
This illustration shows how the different pieces of
equipment fit together to contain, mix, and inject
hydraulic fracturing fluid into a production well.
Water, proppant, and additives are blended together
and pumped to the manifold, where high pressure
pumps transfer the fluid to the frac head.
Additives and proppant can be blended with
water at different times and in different amounts
during hydraulic fracturing. Thus, the composition
of hydraulic fracturing fluids can vary during the
hydraulic fracturing job.
Blender
High Pressure
Pump
low pressure lines ~ high pressure lines
Source: Adopted from Olson (2011) and BJ Services Company (2009)
Well Pad During
Hydraulic Fracturing
Equipment set up for
hydraulic fracturing.
Spills of hydraulic fracturing fluids or additives have reached, and therefore impacted, surface
water resources. Thirteen of the 151 spills characterized by the EPA were reported to have reached
a surface water body (often creeks or streams). Among the 13 spills, reported spill volumes ranged
from 28 gallons (105 liters] to 7,350 gallons (27,800 liters). Additionally, Brantley etal. f20141 and
Considine et al. (2012) identified fewer than 10 total instances of spills of additives and/or
hydraulic fracturing fluids greater than 400 gallons (1,500 liters) that reached surface waters in
ES-23
-------
Executive Summary
Pennsylvania between January 2008 and June 2013. Reported spill volumes for these spills ranged
from 3,400 gallons (13,000 liters) to 227,000 gallons (859,000 liters).
Although impacts on surface water resources have been documented, site-specific studies that
could be used to describe factors that affect the frequency or severity of impacts were not available.
In the absence of such studies, we relied on fundamental scientific principles to identify factors that
affect how hydraulic fracturing fluids and chemicals can move through the environment to drinking
water resources. Because these factors influence whether spilled fluids reach groundwater and
surface water resources, they affect the frequency and severity of impacts on drinking water
resources from spills during the chemical mixing stage of the hydraulic fracturing water cycle.
The potential for spilled fluids to impact groundwater or surface water resources depends on the
characteristics of the spill, the environmental fate and transport of the spilled fluid, and spill
response activities (Figure ES-5). Site-specific characteristics affect how spilled liquids move
through soil into the subsurface or over the land surface. Generally, highly permeable soils or
fractured rock can allow spilled liquids to move quickly into and through the subsurface, limiting
the opportunity for spilled liquids to move over land to surface water resources. In low
permeability soils, spilled liquids are less able to move into the subsurface and are more likely to
move over the land surface. In either case, the volume spilled and the distance between the location
of the spill and nearby water resources affects whether spilled liquids reach drinking water
resources. Large-volume spills are generally more likely to reach drinking water resources because
they are more likely to be able to travel the distance between the location of the spill and nearby
water resources.
In general, chemical and physical properties, which depend on the identity and structure of a
chemical, control whether spilled chemicals evaporate, stick to soil particles, or move with water.
The EPA identified measured or estimated chemical and physical properties for 455 of the 1,084
chemicals used in hydraulic fracturing fluids between 2005 and 2013.1 The properties of these
chemicals varied widely, from chemicals that are more likely to move quickly through the
environment with a spilled liquid to chemicals that are more likely to move slowly through the
environment because they stick to soil particles.2 Chemicals that move slowly through the
environment may act as longer-term sources of contamination if spilled.
1 Chemical and physical properties were identified using EPI Suite™. EPI Suite™ is a collection of chemical and physical
property and environmental fate estimation programs developed by the EPA and Syracuse Research Corporation. It can
be used to estimate chemical and physical properties of individual organic compounds. Of the 1,084 hydraulic fracturing
fluid chemicals identified by the EPA, 629 were not individual organic compounds, and thus EPI Suite™ could not be used
to estimate their chemical and physical properties.
2 These results describe how some hydraulic fracturing chemicals behave in infinitely dilute aqueous solutions, which is a
simplified approximation of the real-world mixtures found in hydraulic fracturing fluids. The presence of other chemicals
in a mixture can affect the fate and transport of a chemical.
ES-24
-------
Executive Summaiy
Spilled Hydraulic Fracturing
Fluid or Additive
Groundwater
Surface Water
Spill Characteristics
What chemicals were spilled?
How much was spilled?
Spill Response Activities
What actions were taken to remove the
spilled fluid from the environment?
Environmental Fate and Transport
How would the spilled fluid move
through the surface and
underground environment?
Figure ES-5. Generalized depiction of factors that influence whether spilled hydraulic
fracturing fluids or additives reach drinking water resources, including spill characteristics,
environmental fate and transport, and spill response activities.
Spill prevention practices and spill response activities are designed to prevent spilled fluids from
reaching groundwater or surface water resources and minimize impacts from spilled fluids. Spill
prevention and response activities are influenced by federal, state, and local regulations and
company practices. Spill prevention practices include secondary containment systems (e.g., liners
and berms), which are designed to contain spilled fluids and prevent them from reaching soil,
groundwater, or surface water. Spill response activities include activities taken to stop the spill,
contain spilled fluids (e.g., the deployment of emergency containment systems), and clean up
spilled fluids (e.g., removal of contaminated soil). It was beyond the scope of this report to evaluate
the implementation and efficacy of spill prevention practices and spill response activities.
The severity of impacts on water quality from spills of hydraulic fracturing fluids or additives
depends on the identity and amount of chemicals that reach groundwater or surface water
resources, the toxicity of the chemicals, and the characteristics of the receiving water resource.1
Characteristics of the receiving groundwater or surface water resource (e.g., water resource size
and flow rate) can affect the magnitude and duration of impacts by reducing the concentration of
spilled chemicals in a drinking water resource. Impacts on groundwater resources have the
1 Human health hazards associated with hydraulic fracturing fluid chemicals are discussed in Chapter 9 and summarized
in the "Chemicals in the Hydraulic Fracturing Water Cycle" section below.
ES-25
-------
Executive Summary
potential to be more severe than impacts on surface water resources because it takes longer to
naturally reduce the concentration of chemicals in groundwater and because it is generally difficult
to remove chemicals from groundwater resources. Due to a lack of data, particularly in terms of
groundwater monitoring after spill events, little is publicly known about the severity of drinking
water impacts from spills of hydraulic fracturing fluids or additives.
Chemical Mixing Conclusions
Spills of hydraulic fracturing fluids and additives during the chemical mixing stage of the hydraulic
fracturing water cycle have reached surface water resources in some cases and have the potential
to reach groundwater resources. Although the available data indicate that spills of various volumes
can reach surface water resources, large volume spills are more likely to travel longer distances to
nearby groundwater or surface water resources. Consequently, large volume spills likely increase
the frequency of impacts on drinking water resources. Large volume spills, particularly of
concentrated additives, are also likely to result in more severe impacts on drinking water resources
than small volume spills because they can deliver a large quantity of potentially hazardous
chemicals to groundwater or surface water resources. Impacts on groundwater resources are likely
to be more severe than impacts on surface water resources because of the inherent characteristics
of groundwater. Spill prevention and response activities are designed to prevent spilled fluids from
reaching groundwater or surface water resources and minimize impacts from spilled fluids.
Well Injection
Activity: The injection and movement of hydraulic fracturing fluids through the oil and gas
production well and in the targeted rock formation.
Relationship to Drinking Water Resources: Belowground pathways, including the production
well itself and newly-created fractures, can allow hydraulic fracturing fluids or other fluids to reach
underground drinking water resources.
Hydraulic fracturing fluids primarily move along two pathways during the well injection stage: the
oil and gas production well and the newly-created fracture network. Oil and gas production wells
are designed and constructed to move fluids to and from the targeted rock formation without
leaking and to prevent fluid movement along the outside of the well. This is generally accomplished
by installing multiple layers of casing and cement within the drilled hole (Text Box ES-2),
particularly where the well intersects oil-, gas-, and/or water-bearing rock formations. Casing and
cement, in addition to other well components (e.g., packers), can control hydraulic fracturing fluid
movement by creating a preferred flow pathway (i.e., inside the casing) and preventing
unintentional fluid movement (e.g., from the inside of the casing to the surrounding environment or
vertically along the well from the targeted rock formation to shallower formations).1 An EPA survey
of oil and gas production wells hydraulically fractured between approximately September 2009 and
September 2010 suggests that hydraulically fractured wells are often, but not always, constructed
1 Packers are mechanical devices installed with casing. Once the casing is set in the drilled hole, packers swell to fill the
space between the outside of the casing and the surrounding rock or casing.
ES-26
-------
Executive Summary
with multiple casings that have varying amounts of cement surrounding each casing (U.S. EPA.
2015nl Among the wells surveyed, the most common number of casings per well was two: surface
casing and production casing (Text Box ES-2). The presence of multiple cemented casings that
extend from the ground surface to below the designated drinking water resource is one of the
primary well construction features that protects underground drinking water resources.
During hydraulic fracturing, a well is subjected to greater pressure and temperature changes than
during any other activity in the life of the well. As hydraulic fracturing fluid is injected into the well,
the pressure applied to the well increases until the targeted rock formation fractures; then pressure
decreases. Maximum pressures applied to wells during hydraulic fracturing have been reported to
range from less than 2,000 pounds per square inch (psi) [14 megapascals (MPa)] to approximately
12,000 psi (83 MPa).1 A well can also experience temperature changes as cooler hydraulic
fracturing fluid enters the warmer well. In some cases, casing temperatures have been observed to
drop from 212°F (100°C) to 64°F (18°C). A well can experience multiple pressure and temperature
cycles if hydraulic fracturing is done in multiple stages or if a well is re-fractured.2 Casing, cement,
and other well components need to be able to withstand these changes in pressure and
temperature, so that hydraulic fracturing fluids can flow to the targeted rock formation without
leaking.
The fracture network created during hydraulic fracturing is the other primary pathway along which
hydraulic fracturing fluids move. Fracture growth during hydraulic fracturing is complex and
depends on the characteristics of the targeted rock formation and the characteristics of the
hydraulic fracturing operation. In general, rock characteristics, particularly the natural stresses
placed on the targeted rock formation due to the weight of the rock above, affect how the rock
fractures, including whether newly-created fractures grow vertically (i.e., perpendicular to the
ground surface) or horizontally (i.e., parallel to the ground surface) (Text Box ES-8). Because
hydraulic fracturing fluids are used to create and grow fractures, fracture growth during hydraulic
fracturing can be controlled by limiting the rate and volume of hydraulic fracturing fluid injected
into the well.
Publicly available data on fracture growth are currently limited to microseismic and tiltmeter data
collected during hydraulic fracturing operations in five shale plays in the United States. Analyses of
these data by Fisher and Warpinski (2012) and Davies etal. (2012) indicate that the direction of
fracture growth generally varied with depth and that upward vertical fracture growth was often on
the order of tens to hundreds of feet in the shale formations studied (Text Box ES-8). One percent of
the fractures had a fracture height greater than 1,148 feet (350 meters), and the maximum fracture
height among all of the data reported was 1,929 feet (588 meters). These reported fracture heights
suggest that some fractures can grow out of the targeted rock formation and into an overlying
formation. It is unknown whether these observations apply to other hydraulically fractured rock
formations because similar data from hydraulic fracturing operations in other rock formations are
not currently available to the public.
1 For comparison, average atmospheric pressure is approximately 15 psi.
2 In a multi-stage hydraulic fracturing operation, specific parts of the well are isolated and hydraulically fractured until
the total desired length of the well has been hydraulically fractured.
ES-2 7
-------
Executive Summary
Text Box ES-8. Fracture Growth.
Fracture growth during hydraulic fracturing is complex and depends on the characteristics of the targeted rock
formation and the characteristics of the hydraulic fracturing operation.
Primary Direction of Fracture Growth
In general, the weight of the rock above the point of hydraulic fracturing affects the primary direction of fracture growth.
Therefore, the depth at which hydraulic fracturing occurs affects whether fractures grow vertically or horizontally.
Ground Surface
Production Well
When hydraulic fracturing occurs at depths less than
approximately 2,000 feet, the primary direction of fracture
growth is horizontal, or parallel to the ground surface.
When hydraulic fracturing occurs at depths
greater than approximately 2,000 feet, the
primary direction of fracture growth is vertical,
or perpendicularto the ground surface.
u
Fracture Height
Fisher and Warpinski (2012) and Davies et al. (2012) analyzed microseismic and tiltmeter data collected during thousands of
hydraulic fracturing operations in the Barnett, Eagle Ford, Marcellus, Niobrara, and Woodford shale plays. Their data provide
information on fracture heights in shale. Top fracture heights varied between shale plays and within individual shale plays.
The top fracture height is the vertical distance
upward from the well, between the fracture tip
and the well.
Eagle Ford
Woodford
Barnett
Marcellus
Shale Play
Approximate Median Top
Fracture Height [feet (meters)]
Eagle Ford
130 (40)
Woodford
160 (50)
Barnett
200 (60)
Marcellus
400 (120)
Niobrara
160 (50)
100 200 300 400
Top Fracture Height (m)
Source: Davies et (2012)
The potential for hydraulic fracturing fluids to reach, and therefore impact, underground drinking
water resources is related to the pathways along which hydraulic fracturing fluids primarily move
during hydraulic fracturing: the oil and gas production well itself and the fracture network created
during hydraulic fracturing. Because the well can be a pathway for fluid movement, the mechanical
integrity of the well is an important factor that affects the frequency and severity of impacts from
ES-28
-------
Executive Summaiy
the well injection stage of the hydraulic fracturing water cycle.1 A well with insufficient mechanical
integrity can allow unintended fluid movement, either from the inside to the outside of the well
(pathway 1 in Figure ES-6) or vertically along the outside of the well (pathways 2-5). The existence
of one or more of these pathways can result in impacts on drinking water resources if hydraulic
fracturing fluids reach groundwater resources. Impacts on drinking water resources can also occur
if gases or liquids released from the targeted rock formation or other formations during hydraulic
fracturing travel along these pathways to groundwater resources.
Figure ES-6. Potential pathways for fluid movement in a cemented well.
These pathways (represented by the white arrows) include: (1) a casing and tubing leak into the surrounding rock,
(2) an uncemented annulus (i.e., the space behind the casing), (3) microannuli between the casing and cement,
(4) gaps in cement due to poor cement quality, and (5) microannuli between the cement and the surrounding rock.
This figure is intended to provide a conceptual illustration of pathways that can be present in a well and is not to
scale.
1 Mechanical integrity is the absence of significant leakage within or outside of the well components.
ES-29
-------
Executive Summary
The pathways shown in Figure ES-6 can exist because of inadequate well design or construction
(e.g., incomplete cement around the casing where the well intersects with water-, oil-, or gas-
bearing formations) or can develop over the well's lifetime, including during hydraulic fracturing.
In particular, casing and cement can degrade over the life of the well because of exposure to
corrosive chemicals, formation stresses, and operational stresses (e.g., pressure and temperature
changes during hydraulic fracturing). As a result, some hydraulically fractured oil and gas
production wells may develop one or more of the pathways shown in Figure ES-6. Changes in
mechanical integrity over time have implications for older wells that are hydraulically fractured
because these wells may not be able to withstand the stresses applied during hydraulic fracturing.
Older wells may also be hydraulically fractured at shallower depths, where cement around the
casing may be inadequate or missing.
Examples of mechanical integrity problems have been documented in hydraulically fractured oil
and gas production wells. In one case, hydraulic fracturing of an inadequately cemented gas well in
Bainbridge Township, Ohio, contributed to the movement of methane into local drinking water
resources.1 In another case, an inner string of casing burst during hydraulic fracturing of an oil well
near Kill deer, North Dakota, resulting in a release of hydraulic fracturing fluids and formation fluids
that impacted a groundwater resource.
The potential for hydraulic fracturing fluids or other fluids to reach underground drinking water
resources is also related to the fracture network created during hydraulic fracturing. Because fluids
travel through the newly-created fractures, the location of these fractures relative to underground
drinking water resources is an important factor affecting the frequency and severity of potential
impacts on drinking water resources. Data on the relative location of induced fractures to
underground drinking water resources are generally not available, because fracture networks are
infrequently mapped and because there can be uncertainty in the depth of the bottom of the
underground drinking water resource at a specific location.
Without these data, we were often unable to determine with certainty whether fractures created
during hydraulic fracturing have reached underground drinking water resources. Instead, we
considered the vertical separation distance between hydraulically fractured rock formations and
the bottom of underground drinking water resources. Based on computer modeling studies,
Birdsell etal. (2015a) concluded that it is less likely that hydraulic fracturing fluids would reach an
overlying drinking water resource if (1) the vertical separation distance between the targeted rock
formation and the drinking water resource is large and (2) there are no open pathways (e.g.,
natural faults or fractures, or leaky wells). As the vertical separation distance between the targeted
rock formation and the underground drinking water resource decreases, the likelihood of upward
migration of hydraulic fracturing fluids to the drinking water resource increases (Birdsell etal..
2015a).
Figure ES-7 illustrates how the vertical separation distance between the targeted rock formation
and underground drinking water resources can vary across the United States. The two example
1 Although ingestion of methane is not considered to be toxic, methane can pose a physical hazard. Methane can
accumulate to explosive levels when allowed to exsolve (degas] from groundwater in closed environments.
ES-30
-------
Executive Summaiy
environments depicted in panels a and b represent the range of separation distances shown in
panel c. In Figure ES-7a, there are thousands of feet between the bottom of the underground
drinking water resource and the hydraulically fractured rock formation. These conditions are
generally reflective of deep shale formations (e.g., Haynesville Shale], where oil and gas production
wells are first drilled vertically and then horizontally along the targeted rock formation.
Microseismic data and modeling studies suggest that, under these conditions, fractures created
during hydraulic fracturing are unlikely to grow through thousands of feet of rock into
underground drinking water resources.
(a)
(b)
Drinking Water Resource
A
Drinking Water Resource
No Vertica
Separation
* D ictn rt ra
Distance.
Drinking Water Resource
and Targeted Rock Formation
Targeted Rock Formation
15,000
10,000
5,000
(C)
. ii ii ii n
III--
.<$* jp <# , <*> . <*>- ' J>' Jo' JV Jb' .54' •£'
C?> r© rQ rO r© (O ¦y
^ .<5° J5° .$> , 0° , cP
v "v> "y V' <)'
Separation Distance in Measured Depth (feet)
Figure ES-7. Examples of different subsurface environments in which hydraulic fracturing
takes place.
In panel a, there are thousands of feet between the base of the underground drinking water resource and the part
of the well that is hydraulically fractured. Panel b illustrates the co-location of groundwater and oil and gas
resources. In these types of situations, there is no separation between the shallowest point of hydraulic fracturing
within the well and the bottom of the underground drinking water resource. Panel c shows the estimated
distribution of separation distances for approximately 23,000 oil and gas production wells hydraulically fractured
by nine service companies between 2009 and 2019 (U.S. EPA. 2015n). The separation distance is the distance along
the well between the point of shallowest hydraulic fracturing in the well and the base of the protected
groundwater resource (illustrated in panel a). The error bars in panel c display 95% confidence intervals.
When drinking water resources are co-located with oil and gas resources and there is no vertical
separation between the hydraulically fractured rock formation and the bottom of the underground
drinking water resource (Figure ES-7b), the injection of hydraulic fracturing fluids impacts the
quality of the drinking water resource. According to the information examined in this report, the
overall occurrence of hydraulic fracturing within a drinking water resource appears to be low, with
Targeted Rock Formation
Separation
Distance in
Measured Depth
Vertical
Separation
Distance
ES-31
-------
Executive Summary
the activity generally concentrated in some areas in the western United States (e.g., the Wind River
Basin near Pavillion, Wyoming, and the Powder River Basin of Montana and Wyoming).1 Hydraulic
fracturing within drinking water resources introduces hydraulic fracturing fluid into formations
that may currently serve, or in the future could serve, as a drinking water source for public or
private use. This is of concern in the short-term if people are currently using these formations as a
drinking water supply. It is also of concern in the long-term, because drought or other conditions
may necessitate the future use of these formations for drinking water.
Regardless of the vertical separation between the targeted rock formation and the underground
drinking water resource, the presence of other wells near hydraulic fracturing operations can
increase the potential for hydraulic fracturing fluids or other subsurface fluids to move to drinking
water resources. There have been cases in which hydraulic fracturing at one well has affected a
nearby oil and gas well or its fracture network, resulting in unexpected pressure increases at the
nearby well, damage to the nearby well, or spills at the surface of the nearby well. These well
communication events, or "frac hits," have been reported in New Mexico, Oklahoma, and other
locations. Based on the available information, frac hits most commonly occur when multiple wells
are drilled from the same surface location and when wells are spaced less than 1,100 feet (335
meters) apart Frac hits have also been observed at wells up to 8,422 feet (2,567 meters) away from
a well undergoing hydraulic fracturing.
Abandoned wells near a well undergoing hydraulic fracturing can provide a pathway for vertical
fluid movement to drinking water resources if those wells were not properly plugged or if the plugs
and cement have degraded over time. For example, an abandoned well in Pennsylvania produced a
30-foot (9-meter) geyser of brine and gas for more than a week after hydraulic fracturing of a
nearby gas well. The potential for fluid movement along abandoned wells may be a significant issue
in areas with historic oil and gas exploration and production. Various studies estimate the number
of abandoned wells in the United States to be significant. For instance, the Interstate Oil and Gas
Compact Commission estimates that over 1 million wells were drilled in the United States prior to
the enactment of state oil and gas regulations (IOGCC. 20081. The location and condition of many of
these wells are unknown, and some states have programs to find and plug abandoned wells.
Well Injection Conclusions
Impacts on drinking water resources associated with the well injection stage of the hydraulic
fracturing water cycle have occurred in some instances. In particular, mechanical integrity failures
have allowed gases or liquids to move to underground drinking water resources. Additionally,
hydraulic fracturing has occurred within underground drinking water resources in parts of the
United States. This practice introduces hydraulic fracturing fluids into underground drinking water
resources. Consequently, the mechanical integrity of the well and the vertical separation distance
between the targeted rock formation and underground drinking water resources are important
factors that affect the frequency and severity of impacts on drinking water resources. The presence
of multiple layers of cemented casing and thousands of feet of rock between hydraulically fractured
1 Section 6.3.2 in Chapter 6.
ES-32
-------
Executive Summary
rock formations and underground drinking water resources can reduce the frequency of impacts on
drinking water resources during the well injection stage of the hydraulic fracturing water cycle.
Produced Water Handling
Activity: The on-site collection and handling of water that returns to the surface after hydraulic
fracturing and the transportation of that water for disposal or reuse.
Relationship to Drinking Water Resources: Spills of produced water can reach groundwater and
surface water resources.
After hydraulic fracturing, the injection pressure applied to the oil or gas production well is
released, and the direction of fluid flow reverses, causing fluid to flow out of the well. The fluid that
initially returns to the surface after hydraulic fracturing is mostly hydraulic fracturing fluid and is
sometimes called "flowback" (Text Box ES-9). As time goes on, the fluid that returns to the surface
contains water and economic quantities of oil and/or gas that are separated and collected. Water
that returns to the surface during oil and gas production is similar in composition to the fluid
naturally found in the targeted rock formation and is typically called "produced water." The term
"produced water" is also used to refer to any water, including flowback, that returns to the surface
through the production well as a by-product of oil and gas production. This latter definition of
"produced water" is used in this report.
Produced water can contain many constituents, depending on the composition of the injected
hydraulic fracturing fluid and the type of rock hydraulically fractured. Knowledge of the chemical
composition of produced water comes from the collection and analysis of produced water samples,
which often requires advanced laboratory equipment and techniques that can detect and quantify
chemicals in produced water. In general, produced water has been found to contain:
• Salts, including those composed from chloride, bromide, sulfate, sodium, magnesium, and
calcium;
• Metals, including barium, manganese, iron, and strontium;
• Naturally-occurring organic compounds, including benzene, toluene, ethylbenzene,
xylenes (BTEX), and oil and grease;
• Radioactive materials, including radium; and
• Hydraulic fracturing chemicals and their chemical transformation products.
The amount of these constituents in produced water varies across the United States, both within
and among different rock formations. Produced water from shale and tight gas formations is
typically very salty compared to produced water from coalbed methane formations. For example,
the salinity of produced water from the Marcellus Shale has been reported to range from less than
1,500 milligrams per liter (mg/L) of total dissolved solids to over 300,000 mg/L, while produced
water from coalbed methane formations has been reported to range from 170 mg/L of total
ES-33
-------
Executive Summary
Text Box ES-9. Produced Water from Hydraulically Fractured Oil and Gas Production Wells.
Water of varying quality is a byproduct of oil and gas production. The composition and volume of produced water
varies by well, rock formation, and time after hydraulic fracturing. Produced water can contain hydraulic fracturing
fluid, formation water, and chemical transformation products.
Produced
Water
Hydraulic Fracturing Fluid
Base fluid, proppant, and additives in hydraulic fracturing fluids.
Chemical Transformation Products
New chemicals that are formed when chemicals in
hydraulic fracturing fluids undergo
chemical reactions, degrade, or transform.
Formation Water
Water naturally found in the pore spaces of the targeted rock formation. Formation water is often salty and can have
different amounts and types of metals, radioactive materials, hydrocarbons (e.g., oil and gas), and other chemicals.
Water Produced Immediately After Hydraulic Fracturing
Generally, the fluid that initially returns to the surface is
mostly a mixture of the injected hydraulic fracturing fluid
and its reaction and degradation products.
Water Produced During Oil or Gas Production
The fluid that returns to the surface when oil and/or gas is
produced generally resembles the formation water.
Produced Water
(Also called "flowback")
r
< i
Produced
Water
J
The volume of water produced per day immediately after hydraulic
fracturing is generally greater than the volume of water produced
per day when the well is also producing oil and/or gas.
dissolved solids to nearly 43,000 mg/L.1 Shale and sandstone formations also commonly contain
radioactive materials, including uranium, thorium, and radium. As a result, radioactive materials
have been detected in produced water from these formations.
Produced water volumes can vary by well, rock formation, and time after hydraulic fracturing.
Volumes are often described in terms of the volume of hydraulic fracturing fluid used to fracture
the well. For example, Figure ES-4 shows that wells in the Marcellus Shale typically produce 10-
30% of the volume injected in the first 10 years after hydraulic fracturing. In comparison, some
wells in the Barnett Shale have produced 100% of the volume injected in the first three years.
1 For comparison, the average salinity of seawater is approximately 35,000 mg/L of total dissolved solids.
ES-34
-------
Executive Summary
Because of the large volumes used for hydraulic fracturing [about 4 million gallons (15 million
liters) per well in the Marcellus Shale and the Barnett Shale], hundreds of thousands to millions of
gallons of produced water need to be collected and handled at the well site. The volume of water
produced per day generally decreases with time, so the volumes handled on site immediately after
hydraulic fracturing can be much larger than the volumes handled when the well is producing oil
and/or gas (Text Box ES-9).
Produced water flows from the well to on-site tanks or pits through a series of pipes or flowlines
(Text Box ES-10) before being transported offsite via trucks or pipelines for disposal or reuse.
While produced water collection, storage, and transportation systems are designed to contain
produced water, spills can occur. Changes in drinking water quality can occur if produced water
spills reach groundwater or surface water resources.
Produced water spills have been reported across the United States. Median spill volumes among the
datasets reviewed for this report ranged from approximately 340 gallons (1,300 liters) to 1,000
gallons (3,800 liters) per spill.1 There were, however, a small number of large volume spills. In
North Dakota, for example, there were 12 spills greater than 21,000 gallons (79,500 liters), five
spills greater than 42,000 gallons (160,000 liters), and one spill of 2.9 million gallons (11 million
liters) in 2015. Common causes of produced water spills included human error and equipment
leaks or failures. Common sources of produced water spills included hoses or lines and storage
equipment
Spills of produced water have reached groundwater and surface water resources. In U.S. EPA
(2015m). 30 of the 225 (13%) produced water spills characterized were reported to have reached
surface water (e.g., creeks, ponds, or wetlands), and one was reported to have reached
groundwater. Of the spills that were reported to have reached surface water, reported spill volumes
ranged from less than 170 gallons (640 liters) to almost 74,000 gallons (280,000 liters). A separate
assessment of produced water spills reported to the California Office of Emergency Services
between January 2009 and December 2014 reported that 18% of the spills impacted waterways
CCCST. 2015a1.
Documented cases of water resource impacts from produced water spills provide insights into the
types of impacts that can occur. In most of the cases reviewed for this report, documented impacts
included elevated levels of salinity in groundwater and/or surface water resources.2 For example,
the largest produced water spill reported in this report occurred in North Dakota in 2015, when
approximately 2.9 million gallons (11 million liters) of produced water spilled from a broken
pipeline. The spilled fluid flowed into Blacktail Creek and increased the concentration of chloride
and the electrical conductivity of the creek; these observations are consistent with an increase in
water salinity. Elevated levels of electrical conductivity and chloride were also found downstream
in the Little Muddy River and the Missouri River. In another example, pits holding flowback fluids
overflowed in Kentucky in 2007. The spilled fluid reached the Acorn Fork Creek, decreasing the pH
of the creek and increasing the electrical conductivity.
1 See Section 7.4 in Chapter 7.
2 Groundwater impacts from produced water management practices are described in Chapter 8 and summarized in the
"Wastewater Disposal and Reuse" section below.
ES-35
-------
Executive Summary
Text Box ES-10. On-Site Storage of Produced Water.
Water that returns to the surface after hydraulic fracturing is collected and stored on site in pits or tanks.
Above: Produced water storage pit. (Source: US EPA)
Left: Produced water storage tanks. (Source: US EPA)
Above: Flowback pit. (Source: US DOE/NETL)
Right: Flowback tanks. (Source: US EPA)
Water Tanks
High Pressure Pumps
Manifold
High Pressure Pumps
Flowback
Tanks
Frac
bo.Ki
Produced Water Storage Immediately after
Hydraulic Fracturing
After hydraulic fracturing, water is returned
to the surface. Water initially produced
from the well after hydraulic fracturing is
sometimes called "flowback." This water can
be stored onsite in tanks or pits before being
taken offsite for injection in Class II wells,
reuse in other hydraulic fracturing operations,
or aboveground disposal.
Source: Adapted from Olson (2011) and BJ
Services Company (2009)
Produced Water Storage During Oil or Gas Production
Water is generally produced throughout the life of an oil and gas production well. During oil and gas production, the
equipment on the well pad often includes the wellhead and storage tanks or pits for gas, oil, and produced water.
ES-36
-------
Executive Summary
Site-specific studies of historical produced water releases highlight the role of local geology in the
movement of produced water through the environment. Whittemore f20071 described a site in
Kansas where low permeability soils and rock caused produced water to primarily flow over the
land surface to nearby surface water resources, reducing the amount of produced water that
infiltrated soil. In contrast, Ottonetal. (2007) explored the release of produced water and oil from
two pits in Oklahoma. In this case, produced water from the pits flowed through thin soil and into
the underlying, permeable rock. Produced water was also identified in deeper, less permeable rock.
The authors suggest that produced water moved into the deeper, less permeable rock through
natural fractures. Together, these studies highlight the role of preferential flow paths (i.e., paths of
least resistance) in the movement of produced water through the environment.
Spill response activities likely reduce the severity of impacts on groundwater and surface water
resources from produced water spills. For example, in the North Dakota example noted above,
absorbent booms were placed in the affected creek and contaminated soil and oil-coated ice were
removed from the site. In another example, a pipeline leak in Pennsylvania spilled approximately
11,000 gallons (42,000 liters) of produced water, which flowed into a nearby stream. In response,
the pipeline was shut off, a dam was constructed to contain the spilled produced water, water was
removed from the stream, and the stream was flushed with fresh water. In both examples, it was
not possible to quantify how spill response activities reduced the severity of impacts on
groundwater or surface water resources. However, actions taken after the spills were designed to
stop produced water from entering the environment (e.g., shutting off a pipeline), remove produced
water from the environment (e.g., using absorbent booms), and reduce the concentration of
produced water constituents introduced into water resources (e.g., flushing a stream with fresh
water).
The severity of impacts on water quality from spills of produced water depends on the identity and
amount of produced water constituents that reach groundwater or surface water resources, the
toxicity of those constituents, and the characteristics of the receiving water resource.1 In particular,
spills of produced water can have high levels of total dissolved solids, which affects how the spilled
fluid moves through the environment When a spilled fluid has greater levels of total dissolved
solids than groundwater, the higher-density fluid can move downward through groundwater
resources. Depending on the flow rate and other properties of the groundwater resource, impacts
from produced water spills can last for years.
Produced Water Handling Conclusions
Spills of produced water during the produced water handling stage of the hydraulic fracturing
water cycle have reached groundwater and surface water resources in some cases. Several cases of
water resource impacts from produced water spills suggest that impacts are characterized by
increases in the salinity of the affected groundwater or surface water resource. In the absence of
direct pathways to groundwater resources (e.g., fractured rock), large volume spills are more likely
to travel further from the site of the spill, potentially to groundwater or surface water resources.
1 Human health hazards associated with chemicals detected in produced water are discussed in Chapter 9 and
summarized in the "Chemicals in the Hydraulic Fracturing Water Cycle" section below.
ES-37
-------
Executive Summary
Additionally, saline produced water can migrate downward through soil and into groundwater
resources, leading to longer-term groundwater contamination. Spill prevention and response
activities can prevent spilled fluids from reaching groundwater or surface water resources and
minimize impacts from spilled fluids.
Wastewater Disposal and Reuse
Activity: The disposal and reuse of hydraulic fracturing wastewater.
Relationship to Drinking Water Resources: Disposal practices can release inadequately treated
or untreated hydraulic fracturing wastewater to groundwater and surface water resources.
In general, produced water from hydraulically fractured oil and gas production wells is managed
through injection in Class II wells, reuse in other hydraulic fracturing operations, or various
aboveground disposal practices (Text Box ES-11). In this report, produced water from hydraulically
fractured oil and gas wells that is being managed through one of the above management strategies
is referred to as "hydraulic fracturing wastewater." Wastewater management choices are affected
by cost and other factors, including: the local availability of disposal methods; the quality of
produced water; the volume, duration, and flow rate of produced water; federal, state, and local
regulations; and well operator preferences.
Available information suggests that hydraulic fracturing wastewater is mostly managed through
injection in Class II wells. Veil (2015) estimated that 93% of produced water from the oil and gas
industry was injected in Class II wells in 2012. Although this estimate included produced water
from oil and gas wells in general, it is likely indicative of nationwide management practices for
hydraulic fracturing wastewater. Disposal of hydraulic fracturing wastewater in Class II wells is
often cost-effective, especially when a Class II disposal well is located within a reasonable distance
from a hydraulically fractured oil or gas production well. In particular, large numbers of active Class
II disposal wells are found in Texas (7,876), Kansas (5,516), Oklahoma (3,837), Louisiana (2,448),
and Illinois (1,054) (U.S. EPA. 2016d). Disposal of hydraulic fracturing wastewater in Class II wells
has been associated with earthquakes in several states, which may reduce the availability of
injection in Class II wells as a wastewater disposal option in these states.
Nationwide, aboveground disposal and reuse of hydraulic fracturing wastewater are currently
practiced to a much lesser extent compared to injection in Class II wells, and these management
strategies appear to be concentrated in certain parts of the United States. For example,
approximately 90% of hydraulic fracturing wastewater from Marcellus Shale gas wells in
Pennsylvania was reused in other hydraulic fracturing operations in 2013 (Figure ES-4a). Reuse in
hydraulic fracturing operations is practiced in some other areas of the United States as well, but at
lower rates (approximately 5-20%). Evaporation ponds and percolation pits have historically been
used in the western United States to manage produced water from the oil and gas industry and have
likely been used to manage hydraulic fracturing wastewater. Percolation pits, in particular, were
commonly reported to have been used to manage produced water from stimulated wells in Kern
ES-38
-------
Executive Summaiy
Class II wells are used to inject wastewater associated with oil and
gas production underground. Fluids can be injected for disposal
or to enhance oil or gas production from nearby oil and gas
production wells.
Reuse in other hydraulic fracturing operations depends on
the quality and quantity of the available wastewater, the cost
associated with treatment and transportation of the wastewater,
and local water demand for hydraulic fracturing.
Aboveground Disposal Practices
Abovegrourid disposal of treated and untreated hydraulic fracturing wastewater can take many forms, including release to
surface water resources and land application.
Federal and state regulations affect aboveground disposal management options. For example, existing federal
regulations generally prevent the direct release of wastewater pollutants to waters of the United States from
onshore oil and gas extraction facilities east of the 98th meridian. However, in the arid western portion of the
continental United States (west of the 98th meridian), direct discharges of wastewater from onshore oil and gas
extraction facilities to waters of the United States may be permitted if the produced water has a use in agriculture
or wildlife propagation and meets established water quality criteria when discharged.
Text Box ES-11. Hydraulic Fracturing Wastewater Management.
Produced water from hydrauiically fractured oil and gas production wells is often, but not always, considered a
waste product to be managed. Hydraulic fracturing wastewater (i.e., produced water from hydrauiically fractured
wells) is generally managed through injection in Class II wells, reuse in other hydraulic fracturing operations, and
various aboveground disposal practices.
Injection in Class II Wells
Most oil and gas wastewater—including hydraulic fracturing
wastewater—is injected in Class II wells, which are regulated
under the Underground Injection Control Program of the
Safe Drinking Water Act.
Reuse in Other Hydraulic Fracturing Operations
Hydraulic fracturing wastewater can be used, in combination
with fresh water, to make up hydraulic fracturing fluids at
nearby hydraulic fracturing operations.
Evaporation ponds and
percolation pits can be used
for hydraulic fracturing
wastewater disposal.
Evaporation ponds allow
liquid waste to naturally
evaporate. Percolation pits
allow wastewater to move
into the ground, although
this practice has been
discontinued in most states.
Reused Hydraulic
Fracturing
Wastewater
Some wastewater treatment
facilities treat hydraulic
fracturing wastewater
and release the treated
wastewater to surface
water. Solid or liquid
by-products of the
treatment process can be
sent to landfills or injected
underground.
ES-39
-------
Executive Summary
County, California, between 2011 and 2014.1 Beneficial uses (e.g., livestock watering and irrigation)
are also practiced in the western United States if the water quality is considered acceptable,
although available data on the use of these practices are incomplete.
Aboveground disposal practices generally release treated or, under certain conditions, untreated
wastewater directly to surface water or the land surface (e.g., wastewater treatment facilities,
evaporation pits, or irrigation). If released to the land surface, treated or untreated wastewater can
move through soil to groundwater resources. Because the ultimate fate of the wastewater can be
groundwater or surface water resources, the aboveground disposal of hydraulic fracturing
wastewater, in particular, can impact drinking water resources.
Impacts on drinking water resources from the aboveground disposal of hydraulic fracturing
wastewater have been documented. For example, early wastewater management practices in the
Marcellus Shale region in Pennsylvania included the use of wastewater treatment facilities that
released (i.e., discharged) treated wastewater to surface waters (Figure ES-8). The wastewater
treatment facilities were unable to adequately remove the high levels of total dissolved solids found
in produced water from Marcellus Shale gas wells, and the discharges contributed to elevated levels
of total dissolved solids (particularly bromide) in the Monongahela River Basin. In the Allegheny
River Basin, elevated bromide levels were linked to increases in the concentration of hazardous
disinfection byproducts in at least one downstream drinking water facility and a shift to more toxic
brominated disinfection byproducts.2 In response, the Pennsylvania Department of Environmental
Protection revised existing regulations to prevent these discharges and also requested that oil and
gas operators voluntarily stop bringing certain kinds of hydraulic fracturing wastewater to facilities
that discharge inadequately treated wastewater to surface waters.3
The scientific literature and recent data from the Pennsylvania Department of Environmental
Protection suggest that other produced water constituents (e.g., barium, strontium, and radium)
may have been introduced to surface waters through the release of inadequately treated hydraulic
fracturing wastewater. In particular, radium has been detected in stream sediments at or near
wastewater treatment facilities that discharged inadequately treated hydraulic fracturing
wastewater. Such sediments can migrate if they are disturbed during dredging or flood events.
Additionally, residuals from the treatment of hydraulic fracturing wastewater (i.e., the solids or
liquids that remain after treatment) are concentrated in the constituents removed during
treatment, and these residuals can impact groundwater or surface water resources if they are not
managed properly.
1 Hydraulic fracturing was the predominant stimulation practice. Other stimulation practices included acid fracturing and
matrix acidizing. California updated its regulations in 2015 to prohibit the use of percolation pits forthe disposal of fluids
produced from stimulated wells.
2 Disinfection byproducts form through chemical reactions between organic material and disinfectants, which are used in
drinking water treatment. Human health hazards associated with disinfection byproducts are described in Section 9.5.6 in
Chapter 9.
3 See Text Box 8-1 in Chapter 8.
ES-40
-------
Executive Summaiy
¦ Other
Includes road spreading, landfill, and
other disposal practices
¦ Reuse in Oil and Gas Activities
Includes non-hydraulic fracturing oil
and gas activities
Centralized Waste Treatment
Wastewater is treated and either
discharged to surface waters or
reused in other hydraulic fracturing
operations
¦ Publicly-Owned Treatment Works
Wastewater is treated and
discharged to surface waters
Underground Injection
Wastewater is injected into Class II
wells
¦ On-site Reuse in Hydraulic Fracturing
Figure ES-8. Changes in wastewater management practices over time in the Marcellus Shale
area of Pennsylvania.
Data from PA PEP (2015a).
Impacts on groundwater and surface water resources from current and historic uses of lined and
unlined pits, including percolation pits, in the oil and gas industry have been documented. For
example, Kell (20111 reported 63 incidents of non-public water supply contamination from unlined
or inadequately constructed pits in Ohio between 1983 and 2007, and 57 incidents of groundwater
contamination from unlined produced water disposal pits in Texas prior to 1984. Other cases of
impacts have been identified in several states, including New Mexico, Oklahoma, Pennsylvania, and
Wyoming.1 Impacts among these cases included the detection of volatile organic compounds in
groundwater resources, wastewater reaching surface water resources from pit overflows, and
wastewater reaching groundwater resources through liner failures. Based on documented impacts
on groundwater resources from u nlined pits, many states have implemented regulations that
prohibit percolation pits or unlined storage pits for either hydraulic fracturing wastewater or oil
and gas wastewater in general.
The severity of impacts on drinking water resources from the aboveground disposal of hydraulic
fracturing wastewater depends on the volume and quality of the discharged wastewater and the
characteristics of the receiving water resource. In general, large surface water resources with high
flow rates can reduce the severity of impacts through dilution, although impacts may not be
eliminated. In contrast, groundwater is generally slow moving, which can lead to an accumulation
of hydraulic fracturing wastewater contaminants in groundwater from continuous or repeated
discharges to the land surface; the resulting contamination can be long-lasting. The severity of
1 See Section 8.4.5 in Chapter 8.
-
— 1 fl m
O 60%
e
3 50%
i= I I I I I I
¦ . I II I
2009 2010 2011 2012 2013 2014
ES-41
-------
Executive Summary
impacts on groundwater resources will also be influenced by soil and sediment properties and
other factors that control the movement or degradation of wastewater constituents.
Wastewater Disposal and Reuse Conclusions
The aboveground disposal of hydraulic fracturing wastewater has impacted the quality of
groundwater and surface water resources in some instances. In particular, discharges of
inadequately treated hydraulic fracturing wastewater to surface water resources have contributed
to elevated levels of hazardous disinfection byproducts in at least one downstream drinking water
system. Additionally, the use of lined and unlined pits for the storage or disposal of oil and gas
wastewater has impacted surface and groundwater resources. Unlined pits, in particular, provide a
direct pathway for contaminants to reach groundwater. Wastewater management is dynamic, and
recent changes in state regulations and practices have been made to limit impacts on groundwater
and surface water resources from the aboveground disposal of hydraulic fracturing wastewater.
Chemicals in the Hydraulic Fracturing Water Cycle
Chemicals are present in the hydraulic fracturing water cycle. During the chemical mixing stage of
the hydraulic fracturing water cycle, chemicals are intentionally added to water to alter its
properties for hydraulic fracturing (Text Box ES-6). Produced water, which is collected, handled,
and managed in the last two stages of the hydraulic fracturing water cycle, contains chemicals
added to hydraulic fracturing fluids, naturally occurring chemicals found in hydraulically fractured
rock formations, and any chemical transformation products (Text Box ES-9). By evaluating
available data sources, we compiled a list of 1,606 chemicals that are associated with the hydraulic
fracturing water cycle, including 1,084 chemicals reported to have been used in hydraulic
fracturing fluids and 599 chemicals detected in produced water. This list represents a national
analysis; an individual well would likely have a fraction of the chemicals on this list and may have
other chemicals that were not included on this list.
In many stages of the hydraulic fracturing water cycle, the severity of impacts on drinking water
resources depends, in part, on the identity and amount of chemicals that enter the environment.
The properties of a chemical influence how it moves and transforms in the environment and how it
interacts with the human body. Therefore, some chemicals in the hydraulic fracturing water cycle
are of more concern than others because they are more likely to move with water (e.g., spilled
hydraulic fracturing fluid) to drinking water resources, persist in the environment (e.g., chemicals
that do not degrade), and/or affect human health.
Evaluating potential hazards from chemicals in the hydraulic fracturing water cycle is most useful
at local and/or regional scales because chemical use for hydraulic fracturing can vary from well to
well and because the characteristics of produced water are influenced by the geochemistry of
hydraulically fractured rock formations. Additionally, site-specific characteristics (e.g., the local
landscape, and soil and subsurface permeability) can affect whether and how chemicals enter
drinking water resources, which influences how long people may be exposed to specific chemicals
and at what concentrations. As a first step for informing site-specific risk assessments, the EPA
ES-42
-------
Executive Summary
compiled toxicity values for chemicals in the hydraulic fracturing water cycle from federal, state,
and international sources that met the EPA's criteria for inclusion in this report12
The EPA was able to identify chronic oral toxicity values from the selected data sources for 98 of the
1,084 chemicals that were reported to have been used in hydraulic fracturing fluids between 2005
and 2013. Potential human health hazards associated with chronic oral exposure to these chemicals
include cancer, immune system effects, changes in body weight, changes in blood chemistry,
cardiotoxicity, neurotoxicity, liver and kidney toxicity, and reproductive and developmental
toxicity. Of the chemicals most frequently reported to FracFocus 1.0, nine had toxicity values from
the selected data sources (Table ES-3). Critical effects for these chemicals include kidney/renal
toxicity, hepatotoxicity, developmental toxicity (extra cervical ribs), reproductive toxicity, and
decreased terminal body weight
Table ES-3. Available chronic oral reference values for hydraulic fracturing chemicals reported
in 10% or more of disclosures in FracFocus 1.0.
Chemical name (CASRN)8
Chronic oral
reference value
(mg/kg/day)
Critical effect
Percent of
FracFocus 1.0
disclosuresb
Propargyl alcohol (107-19-7)
0.002°
Renal and hepatotoxicity
33
1,2,4-Trimethylbenzene (95-63-6)
0.01°
Decreased pain sensitivity
13
Naphthalene (91-20-3)
0.02°
Decreased terminal body
weight
19
Sodium chlorite (7758-19-2)
0.03°
Neuro-developmental
effects
11
2-Butoxyethanol (111-76-2)
0.1°
Hemosiderin deposition in
the liver
23
Quaternary ammonium compounds, benzyl-
C12-16-alkyldimethyl, chlorides (68424-85-1)
0.44d
Decreased body weight
and weight gain
12
Formic acid (64-18-6)
0.9e
Reproductive toxicity
11
Ethylene glycol (107-21-1)
2C
Kidney toxicity
47
Methanol (67-56-1)
2C
Extra cervical ribs
73
a "Chemical" refers to chemical substances with a single CASRN; these may be pure chemicals (e.g., methanol) or chemical
mixtures (e.g., hydrotreated light petroleum distillates).
b Analysis considered 35,957 disclosures that met selected quality assurance criteria. See Table 9-2 in Chapter 9.
c From the EPA Integrated Risk Information System database.
d From the EPA Human Health Benchmarks for Pesticides database.
0 From the EPA Provisional Peer-Reviewed Toxicity Value database.
1 Specifically, the EPA compiled noncancer oral reference values and cancer oral slope factors (Chapter 9]. A reference
value describes the dose of a chemical that is likely to be without an appreciable risk of adverse health effects. In the
context of this report, the term "reference value" generally refers to reference values for noncancer effects occurring via
the oral route of exposure and for chronic durations. An oral slope factor is an upper-bound estimate on the increased
cancer risk from a lifetime oral exposure to an agent.
2 The EPA's criteria for inclusion in this report are described in Section 9.4.1 in Chapter 9. Sources of information that met
these criteria are listed in Table 9-1 of Chapter 9.
ES-43
-------
Executive Summary
Chronic oral toxicity values from the selected data sources were identified for 120 of the 599
chemicals detected in produced water. Potential human health hazards associated with chronic oral
exposure to these chemicals include liver toxicity, kidney toxicity, neurotoxicity, reproductive and
developmental toxicity, and carcinogenesis. Chemical-specific toxicity values are included in
Chapter 9.
Chemicals in the Hydraulic Fracturing Water Cycle Conclusions
Some of the chemicals in the hydraulic fracturing water cycle are known to be hazardous to human
health. Of the 1,606 chemicals identified by the EPA, 173 had chronic oral toxicity values from
federal, state, and international sources that met the EPA's criteria for inclusion in this report.
These data alone, however, are insufficient to determine which chemicals have the greatest
potential to impact drinking water resources and human health. To understand whether specific
chemicals can affect human health through their presence in drinking water, data on chemical
concentrations in drinking water would be needed. In the absence of these data, relative hazard
potential assessments could be conducted at local and/or regional scales using the multi-criteria
decision analysis approach outlined in Chapter 9. This approach combines available chemical
occurrence data with selected chemical, physical, and toxicological properties to place the severity
of potential impacts (i.e., the toxicity of specific chemicals) into the context of factors that affect the
likelihood of impacts (i.e., frequency of use, and chemical and physical properties relevant to
environmental fate and transport).
Data Gaps and Uncertainties
The information reviewed for this report included cases of impacts on drinking water resources
from activities in the hydraulic fracturing water cycle. Using these cases and other data,
information, and analyses, we were able to identify factors that likely result in more frequent or
more severe impacts on drinking water resources. However, there were instances in which we
were unable to form conclusions about the potential for activities in the hydraulic fracturing water
cycle to impact drinking water resources and/or the factors that influence the frequency or severity
of impacts. Below, we provide perspective on the data gaps and uncertainties that prevented us
from drawing additional conclusions about the potential for impacts on drinking water resources
and/or the factors that affect the frequency and severity of impacts.
In general, comprehensive information on the location of activities in the hydraulic fracturing water
cycle is lacking, either because it is not collected, not publicly available, or prohibitively difficult to
aggregate. This includes information on the:
• Above- and belowground locations of water withdrawals for hydraulic fracturing;
• Surface locations of hydraulically fractured oil and gas production wells, where the
chemical mixing, well injection, and produced water handling stages of the hydraulic
fracturing water cycle take place;
• Belowground locations of hydraulic fracturing, including data on fracture growth; and
• Locations of hydraulic fracturing wastewater management practices, including the
disposal of treatment residuals.
ES-44
-------
Executive Summary
There can also be uncertainty in the location of drinking water resources. In particular, depths of
groundwater resources that are, or in the future could be, used for drinking water are not always
known. If comprehensive data about the locations of both drinking water resources and activities in
the hydraulic fracturing water cycle were available, it would have been possible to more completely
identify areas in the United States in which hydraulic fracturing-related activities either directly
interact with drinking water resources or have the potential to interact with drinking water
resources.
In places where we know activities in the hydraulic fracturing water cycle have occurred or are
occurring, data that could be used to characterize the presence, migration, or transformation of
hydraulic fracturing-related chemicals in the environment before, during, and after hydraulic
fracturing were scarce. Specifically, local water quality data needed to compare pre- and post-
hydraulic fracturing conditions are not usually collected or readily available. The limited amount of
data collected before, during, and after activities in the hydraulic fracturing water cycle reduces the
ability to determine whether these activities affected drinking water resources.
Site-specific cases of alleged impacts on underground drinking water resources during the well
injection stage of the hydraulic fracturing water cycle are particularly challenging to understand
(e.g., methane migration in Dimock, Pennsylvania; the Raton Basin of Colorado; and Parker County,
Texas1). This is because the subsurface environment is complex and below ground fluid movement
is not directly observable. In cases of alleged impacts, activities in the hydraulic fracturing water
cycle may be one of several causes of impacts, including other oil and gas activities, other industries,
and natural processes. Thorough scientific investigations are often necessary to narrow down the
list of potential causes to a single source at site-specific cases of alleged impacts.
Additionally, information on chemicals in the hydraulic fracturing water cycle (e.g., chemical
identity; frequency of use or occurrence; and physical, chemical, and toxicological properties) is not
complete. Well operators claimed at least one chemical as confidential at more than 70% of wells
reported to FracFocus 1.0 (U.S. EPA, 2015a).2 The identity and concentration of these chemicals,
their transformation products, and chemicals in produced water would be needed to characterize
how chemicals associated with hydraulic fracturing activities move through the environment and
interact with the human body. Identifying chemicals in the hydraulic fracturing water cycle also
informs decisions about which chemicals would be appropriate to test for when establishing pre-
hydraulic fracturing baseline conditions and in the event of a suspected drinking water impact.
Of the 1,606 chemicals identified by the EPA in hydraulic fracturing fluid and/or produced water,
173 had toxicity values from sources that met the EPA's criteria for inclusion in this report Toxicity
values from these selected data sources were not available for 1,433 (89%) of the chemicals,
although many of these chemicals have toxicity data available from other data sources.3 Given the
1 See Text Boxes 6-2 (Dimock, Pennsylvania], 6-3 (Raton Basin], and 6-4 (Parker County, Texas] in Chapter 6.
2 Chemical withholding rates in FracFocus have increased over time. Konschnikand Dayalu (2016] reported that 92% of
wells reported in FracFocus 2.0 between approximately March 2011 and April 2015 used at least one chemical that was
claimed as confidential.
3 Chapter 9 describes the availability of data in other data sources. The quality of these data sources was not evaluated as
part of this report.
ES-45
-------
Executive Summary
large number of chemicals identified in the hydraulic fracturing water cycle, this missing
information represents a significant data gap that makes it difficult to fully understand the severity
of potential impacts on drinking water resources.
Because of the significant data gaps and uncertainties in the available data, it was not possible to
fully characterize the severity of impacts, nor was it possible to calculate or estimate the national
frequency of impacts on drinking water resources from activities in the hydraulic fracturing water
cycle. We were, however, able to estimate impact frequencies in some, limited cases (i.e., spills of
hydraulic fracturing fluids or produced water and mechanical integrity failures).1 The data used to
develop these estimates were often limited in geographic scope or otherwise incomplete.
Consequently, national estimates of impact frequencies for any stage of the hydraulic fracturing
water cycle have a high degree of uncertainty. Our inability to quantitatively determine a national
impact frequency or to characterize the severity of impacts, however, did not prevent us from
qualitatively describing factors that affect the frequency or severity of impacts at the local level.
Report Conclusions
This report describes how activities in the hydraulic fracturing water cycle can impact—and have
impacted—drinking water resources and the factors that influence the frequency and severity of
those impacts. It also describes data gaps and uncertainties that limited our ability to draw
additional conclusions about impacts on drinking water resources from activities in the hydraulic
fracturing water cycle. Both types of information—what we know and what we do not know—
provide stakeholders with scientific information to support future efforts.
The uncertainties and data gaps identified throughout this report can be used to identify future
efforts to further our understanding of the potential for activities in the hydraulic fracturing water
cycle to impact drinking water resources and the factors that affect the frequency and severity of
those impacts. Future efforts could include, for example, groundwater and surface water
monitoring in areas with hydraulically fractured oil and gas production wells or targeted research
programs to better characterize the environmental fate and transport and human health hazards
associated with chemicals in the hydraulic fracturing water cycle. Future efforts could identify
additional vulnerabilities or other factors that affect the frequency and/or severity of impacts.
In the near term, decision-makers could focus their attention on the combinations of hydraulic
fracturing water cycle activities and local- or regional-scale factors that are more likely than others
to result in more frequent or more severe impacts. These include:
• Water withdrawals for hydraulic fracturing in times or areas of low water availability,
particularly in areas with limited or declining groundwater resources;
• Spills during the management of hydraulic fracturing fluids and chemicals or produced
water that result in large volumes or high concentrations of chemicals reaching
groundwater resources;
1 See Chapter 10.
ES-46
-------
Executive Summary
• Injection of hydraulic fracturing fluids into wells with inadequate mechanical integrity,
allowing gases or liquids to move to groundwater resources;
• Injection of hydraulic fracturing fluids directly into groundwater resources;
• Discharge of inadequately treated hydraulic fracturing wastewater to surface water
resources; and
• Disposal or storage of hydraulic fracturing wastewater in unlined pits, resulting in
contamination of groundwater resources.
The above combinations of activities and factors highlight, in particular, the vulnerability of
groundwater resources to activities in the hydraulic fracturing water cycle. By focusing attention on
the situations described above, impacts on drinking water resources from activities in the hydraulic
fracturing water cycle could be prevented or reduced.
Overall, hydraulic fracturing for oil and gas is a practice that continues to evolve. Evaluating the
potential for activities in the hydraulic fracturing water cycle to impact drinking water resources
will need to keep pace with emerging technologies and new scientific studies. This report provides
a foundation for these efforts, while helping to reduce current vulnerabilities to drinking water
resources.
ES-47
-------
Executive Summary
This page is intentionally left blank.
ES-48
-------
Chapter 1 - Introduction
Chapter 1. Introduction
1-1
-------
Chapter 1 - Introduction
This page is intentionally left blank.
1-2
-------
Chapter 1 - Introduction
1. Introduction
1.1 Background
People rely on clean and plentiful water resources to meet their basic needs. In the early 2000s,
members of the public began to raise concerns about the use of hydraulic fracturing for oil and gas
production and its potential impacts on drinking water resources. Hydraulic fracturing involves the
injection of fluids into a well under pressures great enough to fracture oil- and gas-bearing
formations. While hydraulic fracturing has been used to enhance oil and gas production from
conventional rock formations, the combination of hydraulic fracturing and directional drilling has
made it economical to produce oil and gas from previously unused unconventional rock
formations.1 This has led to increases in oil and gas production and expanded activity throughout
the United States.
Concerns about the impacts of hydraulic fracturing activities on both the quality and quantity of
drinking water resources have been raised by the public. Some residents living close to oil and gas
production wells report changes in the quality of groundwater resources used for drinking water
and assert that hydraulic fracturing is responsible for these changes. Other concerns include
impacts on water availability due to water use in hydraulic fracturing, especially in areas of the
country experiencing drought, and impacts on water quality from the disposal of wastewater
generated after hydraulic fracturing.
In response to public concerns, the U.S. Congress urged the U.S. Environmental Protection Agency
(EPA) to study the relationship between hydraulic fracturing and drinking water fH.R. Rep. 111-
316, 2009). In 2011, the EPA published its Plan to Study the Potential Impacts of Hydraulic
Fracturing on Drinking Water Resources ("U.S. EPA. 201 Id: hereafter Study Plan), which described
the research the Agency would be conducting on activities involving water that support hydraulic
fracturing (referred to as the "hydraulic fracturing water cycle"). The research described in the
Study Plan began the same year. In 2012, the EPA issued Study of the Potential Impacts of Hydraulic
Fracturing on Drinking Water Resources: Progress Report (U.S. EPA. 2 012h: hereafter Progress
Report) to update the public on the status of EPA's research. Since its initiation, the EPA's hydraulic
fracturing study has directly resulted in the publication of 27 separate government reports and
scientific journal articles. This assessment integrates results from those reports and scientific
journal articles with publicly available data and information. It represents the culmination of the
EPA's hydraulic fracturing study focused on characterizing the relationship between hydraulic
fracturing and drinking water.
1 Conventional oil- and gas-bearing rock formations are often described as "permeable" and tend to have many large, well-
connected pore spaces that allow fluids to move within the rock formation. Unconventional oil- and gas-bearing rock
formations do not exhibit these characteristics. See Chapter 3 for more information on uses of the terms conventional and
unconventional.
1-3
-------
Chapter 1 - Introduction
1.2 Goals
The goals of this assessment are to assess the potential for activities in the hydraulic fracturing
water cycle to impact the quality or quantity of drinking water resources and to identify factors that
affect the frequency or severity of those impacts.
1.3 Scope
The hydraulic fracturing water cycle defines the activities that are within the scope of this
assessment. This cycle encompasses activities involving water that support hydraulic fracturing
and consists of five stages:
1. Water Acquisition: the withdrawal of groundwater or surface water to make hydraulic
fracturing fluids;
2. Chemical Mixing: the mixing of a base fluid (typically water), proppant, and additives at
the well site to create hydraulic fracturing fluids;1
3. Well Injection: the injection and movement of hydraulic fracturing fluids through the oil
and gas production well and in the targeted rock formation;
4. Produced Water Handling: the on-site collection and handling of water that returns to
the surface after hydraulic fracturing and the transportation of that water for disposal or
reuse; and2
5. Wastewater Disposal and Reuse: the disposal and reuse of hydraulic fracturing
wastewater.3
The hydraulic fracturing water cycle, and thus the scope of this assessment, was developed with
input from stakeholders (i.e., federal, state, and tribal partners; industry and non-governmental
organizations; and the general public) and the EPA's Science Advisory Board (SAB) fU.S. EPA.
2011d). The hydraulic fracturing water cycle and our assessment scope reflect interest from
stakeholders in understanding impacts from the act of hydraulic fracturing itself as well as the
activities involving water that support it, without examining impacts from oil and gas production
development broadly.
1A base fluid is the fluid into which proppants and additives are mixed to make a hydraulic fracturing fluid; water is an
example of a base fluid. Additives are chemicals or mixtures of chemicals that are added to the base fluid to change its
properties.
2 "Produced water" is defined in this report as water that flows from and through oil and gas wells to the surface as a by-
product of oil and gas production.
3 "Hydraulic fracturing wastewater" is defined in this report as produced water from hydraulically fractured oil and gas
wells that is being managed using practices that include, but are not limited to, injection in Class II wells, reuse in other
hydraulic fracturing operations, and various aboveground disposal practices. The term "wastewater" is being used as a
general description of certain waters and is not intended to constitute a term of art for legal or regulatory purposes. Class
II wells are used to inject wastewater associated with oil and gas production underground and are regulated under the
Underground Injection Control Program of the Safe Drinking Water Act.
1-4
-------
Chapter 1 - Introduction
Well Injection
Produced Water Handling
Water Acquisition
Chemical Mixing
Wastewater Disposal and Reuse
Figure 1-1. Conceptualized view of the stages of the hydraulic fracturing water cycle.
Shown here is a generalized landscape depicting simplified activities of the hydraulic fracturing water cycle, their relationship to each other, and their
relationship to drinking water resources. Activities may take place in the same watershed or different watersheds and close to or far from drinking water
resources. Drinking water resources are any groundwater or surface water that now serves, or in the future could serve, as a source of drinking water for public
or private use. Arrows depict the movement of water and chemicals. Specific activities in the "Wastewater Disposal and Reuse" inset are (a) disposal via
injection well, (b) wastewater treatment with reuse or discharge, and (c) evaporation or percolation pit disposal. Note: Figure not to scale.
1-5
-------
Chapter 1 - Introduction
This assessment focuses on hydraulic fracturing in onshore oil and gas wells in the contiguous
United States; limited available information on hydraulic fracturing in Alaska is included. To the
extent possible, this assessment addresses hydraulic fracturing in all types of oil- and gas-bearing
formations in which it is conducted, including shale, so-called 'tight' formations (e.g., certain
sandstones, siltstones, and carbonates), coalbeds, and conventional rock formations. The
assessment tends to focus on hydraulic fracturing in shale, reflecting the abundance and availability
of literature and data on hydraulic fracturing in this type of rock formation.
In this assessment, we consider how activities in the hydraulic fracturing water cycle interact with
drinking water resources. Consistent with the Study Plan (U.S. EPA. 2011dl. drinking water
resources are defined within this assessment as any groundwater or surface water that now serves,
or in the future could serve, as a source of drinking water for public or private use. This definition is
broader than most regulatory definitions of "drinking water" to include both fresh and non-fresh
bodies of water that are and could be used now or could be used in the future as sources of drinking
water (Chapter 2). We note that drinking water resources provide not only water that individuals
actually drink but also water used for many additional purposes such as cooking and bathing.
As part of the assessment, we evaluated immediate, near-term, and delayed effects on drinking
water resources from normal operations and accidents. For example, we considered how surface
spills of hydraulic fracturing fluids may have immediate or near-term impacts on neighboring
surface water and shallow groundwater quality (Chapters 5 and 7). We also considered how the
potential release of hydraulic fracturing fluids in the subsurface may take years to impact
groundwater resources, because liquids and gas often move slowly in the subsurface (Chapter 6).
Additionally, impacts may be transient or long-term, often depending on the characteristics of the
affected drinking water resource. Finally, impacts may be detected near the hydraulic fracturing
water cycle activity or some distance away. For instance, we considered that, depending on the
constituents of treated hydraulic fracturing wastewater discharged to a stream and the flow in that
stream, drinking water resource quality could be affected a significant distance downstream
(Chapter 8).
This assessment focuses predominantly on activities supporting a single well or multiple wells at
one site, accompanied by a more limited discussion of cumulative activities and the impacts that
could result from having many wells on a landscape. Studies of cumulative effects are generally
lacking, but we use the scientific literature to address this topic where possible.1
We examine impacts of hydraulic fracturing for oil and gas on drinking water resources and address
factors that affect the frequency or severity of impacts. Specific definitions used in this assessment
are provided below:
• An impact is any change in the quality or quantity of drinking water resources, regardless
of severity, that results from an activity in the hydraulic fracturing water cycle.
1 Cumulative effects refer to combined changes in the environment that can take place as a result of multiple activities
over time and/or space.
1-6
-------
Chapter 1 - Introduction
• A factor is a feature of hydraulic fracturing operations or an environmental condition that
affects the frequency or severity of impacts.
• Frequency is the number of impacts per a given unit (e.g., per geographic area, per unit
time, per number of hydraulically fractured wells, per number of water bodies). Reflecting
the scientific literature, the most common representation of frequency in this assessment
is number of impacts per hydraulically fractured well.
• Severity is the magnitude of change in the quality or quantity of a drinking water resource
as measured by a given metric (e.g., duration, spatial extent, contaminant concentration).
We identify and discuss factors affecting the frequency or severity of impacts to avoid a simple
inventory of all specific situations in which hydraulic fracturing might alter drinking water quality
or quantity. This allows knowledge about the conditions under which impacts are likely or unlikely
to occur to be applied to new circumstances (e.g., a new area of oil or gas development where
hydraulic fracturing is expected to be used) and could inform the development of strategies to
prevent impacts. Although no attempt has been made in this assessment to identify or evaluate
comprehensive best practices for states, tribes, or the industry, we describe ways to avoid or
reduce the frequency or severity of impacts from hydraulic fracturing activities as they have been
reported in the scientific literature. Laws, regulations, and policies also exist to protect drinking
water resources (Text Box 1-1), but a comprehensive summary and evaluation of current or
proposed regulations and policies is beyond the scope of this assessment.
Text Box 1-1. Regulatory Protection for Drinking Water Resources.
The quality and quantity of drinking water resources are protected in the United States by a collection of
federal, state, tribal, and local laws, regulations, and polices. They differ with respect to how water resources
are defined (Chapter 2) and thus which resources qualify for protection. Some policies protect water
resources from oil and gas industry activities as part of a larger set of regulated industries, or from oil and gas
industry activities only, or from hydraulic fracturing-related activities, specifically. Multiple federal and state
agencies, departments, or divisions are responsible for implementing these laws, regulations, and policies. An
exhaustive summary of current and emerging laws, regulations, and policies, those responsible for
implementing them, and enforcement or effectiveness is not in the scope of this assessment. The following
information is designed to give the reader a general understanding of how the U.S. government and states
protect drinking water resources from the potential impacts of activities in the hydraulic fracturing water
cycle.
On the federal level, the U.S. government regulates some activities in the hydraulic fracturing water cycle to
protect drinking water resources. For example, under the Clean Water Act, the National Pollution Discharge
Elimination System (NPDES) program regulates surface discharge of wastewater from the oil and gas sector
(in addition to many other industries). Issuance and enforcement of NPDES discharge permits is primarily the
responsibility of the states that have received NPDES program authorization from the EPA. In addition, the
Safe Drinking Water Act's (SDWA) Underground Injection Control program regulates the underground
disposal of hydraulic fracturing wastewater (and wastewater generated in other industries) and, like the
NPDES program, allows states to seek program authorization from the EPA. The federal government does not
have the authority to regulate hydraulic fracturing as an injection activity under the SDWA except when it
(Text Box 1-1 is continued on the following page.)
1-7
-------
Chapter 1 - Introduction
Text Box 1-1 (continued). Regulatory Protection for Drinking Water Resources.
(1) involves diesel fuel, a result of legislation passed in 2005, or (2) causes an imminent and substantial
endangerment to the health of persons. Additionally, produced water is exempted from regulation as a
hazardous waste under the Resource Conservation and Recovery Act Subtitle C. In 2015, the U.S. Department
of the Interior published a set of regulations for conducting hydraulic fracturing operations on federal public
and tribal lands. It includes requirements to help protect groundwater by updating standards for well
mechanical integrity, wastewater disposal, and public disclosure of chemicals. As of late 2016, a federal
district court judge has set aside these regulations as outside the scope of the U.S. Department of the
Interior's authority, and this decision is being appealed.
States generally have the primary responsibility for protecting drinking water resources from the impacts of
hydraulic fracturing activities fGuralnick. 2016: Zirogannis et al„ 20161 Some states have put in place broad
restrictions or moratoria on hydraulic fracturing activities due in part to concerns about potential risks to
drinking water resources. Many other states allow hydraulic fracturing activities, and several sources of
information track and/or summarize their laws, regulations, and policies. An online database of statutes and
regulations applicable to the oil and gas industry and related to water quality, water quantity, and air quality
in 17 states is maintained by LawAtlas (www. lawatlas.org/oilandgasl
State approaches vary widely, from comprehensive laws addressing all aspects of hydraulic fracturing
activities to regulations addressing specific activities fGuralnick. 20161 In 2009 and 2014, the Ground Water
Protection Council (GWPC) summarized regulations that are designed to protect water resources and
applicable to the oil and gas industry in 27 states; they did not investigate compliance (GWPC. 2014. 20091
The summaries revealed that regulations are carried out by either oil and gas agencies, environmental
agencies, or both, depending on the state. They also identified general categories of existing regulations that
could control impacts on drinking water resources from activities in the hydraulic fracturing water cycle,
including permitting, well design and integrity, injection activities, and surface management of fluids.
Categories were comprised of regulatory "elements." Certain elements had been adopted across 90% or more
of states included in the summaries that allowed hydraulic fracturing as of July 2013: surface casing generally
must be set below the deepest protected groundwater zone; protected groundwater depth is determined on a
well-specific basis or by rule; and surface casing must be cemented from bottom to top. All other elements
were adopted at lower and widely varying rates. For example, as of July 2013, a requirement for water well
testing and monitoring adjacent to hydraulic fracturing operations existed in five states. Other states,
including California, have added this requirement since then.
State laws, regulations, and policies are continually changing. Changes may be initiated by state legislatures
or regulatory agencies (sometimes in response to legal decisions) and generally apply to new wells or future
hydraulic fracturing operations and not existing wells or wells that have been hydraulically fractured in the
past. Third-party groups, like the State Review of Oil and Natural Gas Environmental Regulations
(STRONGER) organization, offer multi-stakeholder reviews of state oil and gas regulatory programs and
recommendations to improve those programs according to guidelines developed by their workgroups.
Interstate organizations of state agency representatives also have initiatives to develop oil and gas resources
while protecting water and other environmental resources, initiatives like the GWPC and Interstate Oil and
Gas Compact Commission's States First. In combination with changing policies, new technologies (such as
those that make it possible to reuse hydraulic fracturing wastewater in subsequent hydraulic fracturing
operations) have the potential to further reduce impacts on drinking water resources.
1-8
-------
Chapter 1 - Introduction
We identify and evaluate potential human health hazards of hydraulic fracturing-related chemicals
(Chapter 9), but this assessment is not a human health risk assessment It does not identify
populations that are exposed to chemicals or other stressors in the environment, estimate the
extent of exposure, or estimate the incidence of human health impacts. Relatedly, we did not
conduct site-specific predictive modeling to quantitatively estimate contaminant concentrations in
drinking water resources, although modeling studies conducted by others are described.
This assessment focuses on the potential for impacts from activities in the hydraulic fracturing
water cycle on drinking water resources. It does not address all concerns that have been raised
about hydraulic fracturing nor about oil and gas exploration and production more generally.
Activities that are not considered in this assessment include acquisition and transport of
constituents of hydraulic fracturing fluids besides water (e.g., sand mining and chemical
production); site selection and development; other infrastructure development (e.g., roads,
pipelines, compressor stations); site reclamation; and well closure. We consider these activities to
be outside the scope of the hydraulic fracturing water cycle and, therefore, their impacts are not
addressed in this assessment Disposal of hydraulic fracturing wastewater in underground injection
control wells is described and characterized, but consistent with the Study Plan, potential for
impacts of this practice on drinking water resources is not included. Additionally, this report does
not discuss the potential impacts of hydraulic fracturing on other water uses (e.g., agriculture or
industry), other aspects of the environment (e.g., air quality, induced seismicity, or ecosystems),
worker health and safety, or communities. Finally, this assessment focuses on the available science
and does not review, consider, or recommend policy options.
1.4 Approach
This assessment relies on scientific literature and data that address topics within the scope of the
hydraulic fracturing water cycle. Scientific journal articles and peer-reviewed EPA reports
containing results from the EPA's hydraulic fracturing study comprise one set of applicable
literature. Other literature evaluated includes articles published in science and engineering
journals, federal and state government reports, non-governmental organization (NGO) reports, and
oil and gas industry publications. Data sources examined include federal- and state-collected data
sets, databases curated by federal and state government agencies, other publicly available data and
information, and data submitted by industry to the EPA.1 In total, we cite approximately 1,200
sources of scientific data and information in this assessment
1.4.1 EPA Hydraulic Fracturing Study Publications
The research topic areas and projects described in the Study Plan were developed with substantial
expert and public input and were designed to meet the data and information needs of this
assessment. As such, peer-reviewed results of research that the EPA conducted under the Study
Plan, published separately as EPA reports or as journal articles, are incorporated and cited
1 Confidential and non-confidential business information was provided to the EPA by nine hydraulic fracturing service
companies in response to a September 2010 information request and by nine oil and gas well operators in response to an
August 2011 information request.
1-9
-------
Chapter 1 - Introduction
frequently throughout this assessment. As is customary in assessments that synthesize a large body
of literature and data, the results of EPA research are contextualized and interpreted in
combination with the other literature and data described in Section 1.4.2. The journal articles and
EPA reports that give complete and detailed project results can be found on the EPA's hydraulic
fracturing study website (www.epa.gov /h fstudv). For ease of reference, a description of the
individual projects, the type of research activity they represent (i.e., analysis of existing data,
scenario evaluation, laboratory study, or case study), and the corresponding citations of published
journal articles and EPA reports that are referenced in this assessment can be found in Appendix A.
1.4.2 Literature and Data Search Strategy
We used a broad search strategy to identify approximately 4,100 sources of scientific information
applicable to this assessment This strategy included requesting input from scientists, stakeholders,
and the public about relevant data and information, and thorough searches of published
information and applicable data.1
Over 1,600 articles, reports, data, and other sources of information were obtained through outreach
to the public, stakeholders, and scientific experts. The EPA requested material through many
venues, as follows. We received recommended literature from the SAB, the EPA's independent
federal scientific advisory committee, from its review of the EPA's draft Study Plan; from its
consultation on the EPA's Progress Report; during an SAB briefing on new and emerging
information related to hydraulic fracturing in fall 2013; and from its peer review of the external
review draft of this assessment. Subject matter experts and stakeholders also recommended
literature through a series of technical workshops and roundtables organized by the EPA between
2011 and 2013. In addition, the public submitted literature recommendations to the SAB during the
SAB review of the draft Study Plan, consultation on the Progress Report, briefing on emerging
information, and review of the external review draft of this assessment, as well as in response to a
formal request for data and information posted in the Federal Register (EPA-HQ-ORD-2010-0674)
in November 2012. The submission deadline was extended from April to November 2013 to
provide the public with additional opportunity to provide information to the EPA.
Approximately 2,500 additional sources were identified by conducting searches via online scientific
databases and federal, state, and stakeholder websites. We searched these databases and websites
in particular for (1) materials addressing topics not covered by the documents submitted by
experts, stakeholders, and the public as noted above, and (2) newly emerging scientific studies.
Multiple targeted and iterative searches on topics determined to be within the scope of the
assessment were conducted until June 1, 2016. After that time, we included newer literature as it
was recommended to us during our internal technical reviews or as it came to our attention and
was determined to be important for filling a gap in information.
1 This study did not review information contained in state and federal enforcement actions concerning alleged
contamination of drinking water resources.
1-10
-------
Chapter 1 - Introduction
1.4.3 Literature and Data Evaluation Strategy
We evaluated the literature and data identified in the search strategy using the five assessment
factors outlined by the EPA Science Policy Council in A Summary of General Assessment Factors for
Evaluating the Quality of Scientific and Technical Information (U.S. EPA. 2003c). The factors are (1)
applicability and utility, (2) evaluation and review, (3) soundness, (4) clarity and completeness, and
(5) uncertainty and variability. Table 1-1 lists these factors along with the specific criteria
developed for this assessment. We first evaluated all materials for applicability. If we determined
that the material was "applicable" under the criteria, the reference was evaluated on the basis of the
other four factors.
Our objective was to consider and then cite literature in the assessment that fully conforms to all
criteria defining each assessment factor. However, in some cases, literature on a topic did not fully
conform to an aspect of the outlined criteria. For instance, the preponderance of literature in some
technical areas is published as white papers and reports for which independent peer review is not
standard practice or is not well documented. To address these areas in which peer-reviewed
literature was limited, we cited literature that may not have been peer-reviewed. These references
often provided useful background information or corroborated conclusions in the peer-reviewed
literature.
Table 1-1. The five factors and accompanying criteria used to evaluate literature and data
cited in this assessment.
Criteria are consistent with those outlined by the EPA's Science Policy Council (U.S. EPA, 2003c). Criteria are
incorporated into the Quality Assurance Project Plans for this assessment (U.S. EPA, 2014d, 2013d).
Factor
Criteria
Applicability
Document provides information useful for assessing the potential pathways for
hydraulic fracturing activities to change the quality or quantity of drinking water
resources, identifies factors that affect the frequency and severity of impacts, or
suggests ways that potential impacts may be avoided or reduced.
Review
Document has been peer-reviewed.
Soundness
Document relies on sound scientific theory and approaches, and conclusions are
consistent with data presented.
Clarity/completeness
Document provides underlying data, assumptions, procedures, and model parameters,
as applicable, as well as information about sponsorship and author affiliations.
Uncertainty/variability
Document identifies uncertainties, variability, sources of error, and/or bias and
properly reflects them in any conclusions drawn.
1.4.4 Quality Assurance and Peer Review
The use of quality assurance (QA) and peer review helps ensure that the EPA conducts high-quality
science that can be used to inform policymakers, industry, and the public. Quality assurance
activities performed by the EPA ensure that the agency's environmental data are of sufficient
quantity and quality to support the data's intended use. The EPA prepared a programmatic Quality
1-11
-------
Chapter 1 - Introduction
Management Plan (U.S. EPA. 2014e) for all of the research conducted under the EPA's Study Plan,
including the review and synthesis of the scientific literature in this assessment The hydraulic
fracturing Quality Management Plan describes the QA program's organizational structure; defines
and assigns QA and quality control (QC) responsibilities; and describes the processes and
procedures used to plan, implement, and assess the effectiveness of the quality system. The broad
plan is then supported by more detailed QA Project Plans (QAPPs). The QAPPs developed for this
assessment provide the technical approach and associated QA/QC procedures for our data and
literature search and evaluation strategies introduced in Section 1.4.2 and 1.4.3 fU.S. EPA. 2014d.
2013d). A QA audit was conducted by the QA Manager during the preparation of this assessment to
verify that the appropriate QA procedures, criteria, reviews, and data verification were adequately
performed and documented. Identifying uncertainties is another aspect of QA; uncertainty,
including data gaps and data limitations, is discussed throughout this assessment.
This report is classified as a Highly Influential Scientific Assessment (HISA), which is defined by the
Office of Management and Budget (OMB) as a scientific assessment that (1) could have a potential
impact of more than $500 million in any year or (2) is novel, controversial, or precedent-setting or
has significant interagency interest fOMB. 20041. The OMB describes specific peer review
requirements for HISAs. To meet these requirements, the EPA often engages the SAB as an
independent federal advisory committee to conduct peer reviews of high-profile scientific matters
relevant to the agency. Members of an ad hoc panel, the same panel that was convened under the
auspices of the SAB to provide comment on the Progress Report, also provided comment on an
external review draft of this assessment.1 Panel members were nominated by the public and chosen
to create a balanced review panel based on factors such as technical expertise, knowledge,
experience, and absence of any real or perceived conflicts of interest. Both peer review comments
provided by the SAB panel (SAB. 2016) and public comments submitted to the panel during their
deliberations about the external review draft of this assessment were carefully considered in the
development of this final document
1.5 Organization
This assessment begins with an Executive Summary that summarizes our overall content and
conclusions. The Executive Summary is written to be accessible to all members of the public.2
This introductory chapter establishes the goals, scope, and approach for the rest of the assessment
Following is a characterization of drinking water resources in the contiguous United States
(Chapter 2). Next, we present a general description of hydraulic fracturing activities and the role of
hydraulic fracturing in the oil and gas industry in the United States (Chapter 3). Chapter 1 is written
1 Information about this process is available online at http: / /vosemite.epa.gov/sab /sabproduct.nsf/
02ad90bl36fc21ef85256eba00436459/b436304ba804e3f885257a5b00521b3b!C)penDocument.
2 The terminology used in the data and literature cited in this assessment can be very technical in nature and sometimes
inconsistent. An attempt has been made throughout this document to provide definitions of technical terms and to use
terminology in a consistent way that enhances understanding of the topics presented for the audiences targeted by each
part of the assessment.
1-12
-------
Chapter 1 - Introduction
to be accessible to all members of the public. Chapters 2 and 3 are written to be accessible to an
audience with general science knowledge.
Chapters 4 through 8 are organized around the stages of the hydraulic fracturing water cycle
(Figure 1-1) and address the potential for activities conducted during those stages to change the
quality or quantity of drinking water resources. Each stage is covered by a separate chapter. There
is also a chapter devoted to an examination of the properties of the chemicals and constituents in
hydraulic fracturing-related fluids (Chapter 9). These chapters are written to be accessible to an
audience with a moderate amount of technical training and expertise in the respective topic areas.
The final chapter provides a synthesis of the information in the assessment (Chapter 10). This
chapter is written to be accessible to an audience with general science knowledge.
The appendices supply information that support the chapters of the assessment This includes an
appendix with a table of all individual products published under the EPA's hydraulic fracturing
study and cited in this assessment, as well as answers to the research questions posed in the Study
Plan (Appendix A). These answers were informed by the products of the study and the data and
literature reviewed in this assessment.
1.6 Intended Use
This state-of-the-science assessment will contribute to the understanding of the potential impacts
of activities in the hydraulic fracturing water cycle on drinking water resources and the factors that
influence those impacts. The data and findings can be used by federal, tribal, state, and local
officials; industry; and the public to better understand and address vulnerabilities of drinking water
resources to hydraulic fracturing activities.
We expect this report will be used to help facilitate and inform dialogue among interested
stakeholders, including Congress, other federal agencies, states, tribal governments, the
international community, industry, NGOs, academia, and the general public. Additionally, the
identification of knowledge gaps will promote greater attention to these areas by researchers.
This report may support future assessment efforts. We anticipate that it could contribute context to
site-specific exposure or risk assessments of hydraulic fracturing, to regional public health
assessments, or to assessments of cumulative impacts of hydraulic fracturing on drinking water
resources over time or over defined geographic areas of interest
Finally, and most importantly, this assessment presents the science to inform decisions by federal,
state, tribal, and local officials; industry; and the public on how best to protect drinking water
resources now and in the future.
1-13
-------
Chapter 1 - Introduction
This page is intentionally left blank.
1-14
-------
Chapter 2 - Drinking Water Resources in the United States
Chapter 2. Drinking Water
Resources in the United States
Abstract
In this assessment, drinking water resources are defined as any body of groundwater or surface water
that now serves, or in the future could serve, as a source of drinking water for public or private use. An
estimated 86% of the United States population derives its household drinking water from public water
systems (PWSs), which mostly use surface water sources, while nearly all of the remaining 14% of
people self-supply their drinking water from groundwater.
Future access to high-quality drinking water in the United States will likely be affected by changes in
climate and water use. The existing distribution and abundance of the drinking water resources may not
be sufficient in some locations to meet future demand. Since 2000, about 30% of the total area of the
contiguous United States has experienced moderate drought conditions and about 20% has experienced
severe drought conditions, which often correlates with diminishment of drinking water supplies. As a
result, non-fresh water resources, such as wastewater from sewage treatment plants, brackish surface
water and groundwater, and seawater are increasingly treated and used to meet the demand for
drinking water.
Hydraulically fractured oil and gas production wells can be located near drinking water sources.
Between 2000 and 2013, approximately 3,900 PWSs had between one and 144 wells hydraulically
fractured within 1 mile of their water source; these PWSs served more than 8.6 million people year-
round in 2013. An additional 740,000 people self-supply their drinking water in counties where at least
30% of the population relies on groundwater and where there were at least 400 hydraulically fractured
wells. Belowground, hydraulic fracturing can occur in close vertical proximity to drinking water
resources. Available data show that depths to hydraulically fractured rock formations containing oil and
gas resources can range from less than 1,000 feet (300 meters) to more than 10,000 feet (3,000 meters),
while drinking water resources may be found between a few tens of feet to as much as 8,000 feet (2,000
meters) below the surface. The EPA found that, along individual wellbores, where data were available,
the distance between these two resources ranged from no separation to more than 10,000 feet (3,000
meters). There is considerable uncertainty in this range of values, however. In many cases, the lack of
accessible information about the depth to the base of formations containing groundwater resources in
need of current and future protection prevents calculation of a vertical separation distance.
The locations of drinking water resources relative to hydraulically fractured oil and gas production wells
influence the potential for activities in the hydraulic fracturing water cycle to impact drinking water
resources. With increased proximity, activities in the hydraulic fracturing water cycle have more
potential to affect aboveground and belowground drinking water resources.
2-1
-------
Chapter 2 - Drinking Water Resources in the United States
This page is intentionally left blank.
2-2
-------
Chapter 2 - Drinking Water Resources in the United States
2. Drinking Water Resources in the United States
2.1 Introduction
Drinking water resources provide the water humans consume, cook with, bathe in, and need for
other purposes. In this assessment, drinking water resources are considered to be any groundwater
or surface water that now serves, or in the future could serve, as a source of drinking water for
public or private use.1 This chapter provides information about drinking water resources in the
United States, including current sources and indications of future trends for drinking water
resources. Assessment of whether and where activities in the hydraulic fracturing water cycle may
impact drinking water resources requires consideration, in part, of the locations of water and oil
and gas resources and what physically separates them. More information about oil and gas
resources and the areas of the United States where hydraulic fracturing occurs is described in
Chapter 3, however this chapter focuses on the lateral (horizontal) and vertical distances between
hydraulic fracturing operations and drinking water resources.
2.2 Ground and Surface Water Resources
All drinking water derives from the finite amount of water found on or below the earth's surface.
Fresh water serves as the source for most drinking water.2 To get an idea of the fresh water fraction
of all water, this section presents an estimate of the earth's water abundance. Shiklomanov (1993)
estimates the amounts of all water on earth, and here these amounts are expressed as the percent
of the earth's total water volume:
• Oceans account for about 96.5%.
• Saline groundwater, saline lakes, and water in the form of ice or vapor account for 2.7%.
• Fresh groundwater, swamps, lakes, and rivers account for the remaining 0.8%, of which
about 99% is groundwater.
Hydrologic Cycle. The process describing the movement of the earth's water through the
atmosphere, land, and oceans is referred to as the hydrologic cycle. Text Box 2-1 describes the
hydrologic cycle, including the manner in which the finite amount of water on the earth moves
through different locations during the stages of the cycle. On land, surface water and groundwater
interact, shown in the text box as surface water infiltrating into the ground, and separately as an
example of groundwater flowing into the river. Water consumption (for example when used for
agriculture, incorporated into a product, or for drinking purposes), temporarily removes water
1 In this chapter, a "drinking water source" means the body of water is now supplying, or is known to be capable of
supplying drinking water.
2 Published estimates of worldwide water supplies, such as by Shiklomanov, do not use a salinity threshold value to define
"fresh" or "saline" water. "Fresh" water is characterized in these published estimates as serving as a source for domestic,
agricultural, and industrial uses. As described further in Section 2.2.1.1, the term "fresh" in this chapter refers to water
having total dissolved solids content up to 3,000 milligrams per liter.
2-3
-------
Chapter 2 - Drinking Water Resources in the United States
from one local place in the hydrologic cycle, but it may be returned to a different point in the
hydrologic cycle. See Chapter 4 for additional discussion of water consumption.
Text Box 2-1. The Hydrologic Cycle.
The finite amount of water and its movement on earth is often called the hydrologic cycle, depicted below.
The three basic, and repeating, stages of this cycle include:
1. Rainfall transfers water from the atmosphere into oceans or onto land,
2. Water on land moves among surface water bodies and groundwater, and
3. Evaporation from land and the oceans returns water to the atmosphere.
Rainwater and melted snow collect into rivers, lakes or other water bodies to become surface water, or
infiltrates into the ground to become groundwater. Humans drink fresh surface and groundwater, and in
some locations, ocean water treated by desalination. Water resides on land or in the ground for varying
amounts of time before moving into another of stage of the hydrologic cycle. Residence times for water found
in different land locations can range from days to millions of years, depending on the path water takes.
Residence time affects water quality on land or in the ground because water dissolves natural earth salts
when in contact with those materials. When water on or under land reaches the ocean, its salt content
ultimately stays in the ocean because evaporation leaves behind dissolved salt creating freshwater vapor.
Evaporation from land and the ocean contribute fresh water to the atmosphere where it can precipitate once
again, thus completing a hydrologic cycle. As drawn in this depiction, evaporation includes the release of
water vapor from plant leaves that originally entered plant root systems in a process known as transpiration.
Snow
Snow melt
/, &jun off
Evaporation
' * S* , * '
Evaporation *•
Ocean
Reservoir/lake*
water-filled pores
within sedimentary
rock
water-filled
fractures within
bedrock
Rain
1 L *Ci
2-4
-------
Chapter 2 - Drinking Water Resources in the United States
2.2.1 Groundwater Resources
Groundwater can be found in the subsurface nearly everywhere, but it varies in quality and
quantity. Groundwater exists in that part of the hydrologic cycle where surface water infiltrates
through soil into subsurface cracks and voids in rock, creating and sustaining aquifers, a natural
process known as groundwater recharge.1 The opposite natural process from recharge is discharge,
where groundwater flows to the surface at springs or through the bottoms of lakes and rivers.
Groundwater also includes water trapped in the pores of sedimentary rocks as they were
deposited.
The scale of groundwater flow systems can be local, regional, or something in between. Local
groundwater flows may be small enough to be measured in the tens of feet while regional
groundwater flows may be large enough to be measured in hundreds of miles f Alley etal.. 19991.
Groundwater movement is related to the rate of groundwater recharge, gravity's effect on the
groundwater, and the permeability of the rock through which groundwater flows. Localized
groundwater flow tends to occur along shallower flow paths with shorter overall residence times,
whereas regional groundwater flow tends to occur along deeper flow paths with longer residence
times fWinter etal.. 19981. Text Box 2-1 depicts differences between local and regional flow
regimes.
The U.S. Geological Survey (USGS) has mapped and described more than 60 principal aquifers in the
United States, although these aquifers are not the only occurrences of groundwater fUSGS. 20091.2
Although the depth to the water table can vary from ground surface to a few tens of feet below
ground surface, the depth to the base of groundwater can be tens of thousands of feet below
ground.3 The depth to the base of individual principal aquifers can be a relatively uniform or may
vary by thousands of feet across the aquifer's areal extent due to sloping geologic formations
and/or changes in topography.
2.2.1.1 Groundwater Quality
The quality of groundwater often correlates with its age, which ranges from days to millions of
years f Alley etal.. 1999: Freeze and Cherry. 1979a: Chebotarev. 195 51.4 As explained in T ext Box
2-1, groundwater salinity tends to increase with increasing residence time due to gradual
dissolution of contacted earth materials. Some groundwater can become very saline. These waters
can result from exposure to soluble sedimentary rocks and/or concentration of salt content due to
evaporation of liquid water in the subsurface fZolfaghari etal.. 2016: Levorsen. 19651. It is also
possible that sea water was trapped in sediments during deposition in ancient oceans, which were
subsequently buried over geologic time. There are instances where groundwater is found at great
1 An aquifer is a water-bearing geologic formation, group of formations, or part of a formation. Groundwater is the water
in an aquifer.
2 Principal aquifers are defined as a regionally extensive aquifer or aquifer system that has the potential to be used to
supply potable water. Principal aquifers in Puerto Rico and the U.S. Virgin Islands are included.
3 The water table refers to the top, or uppermost surface, of groundwater. Below the water table, the ground is saturated
with water.
4 Groundwater age used here refers to how long the water has been in the ground.
2-5
-------
Chapter 2 - Drinking Water Resources in the United States
depths but is relatively fresh. This can be caused by groundwater moving from the surface to deep
locations relatively quickly with little time to pick up dissolved solids and become saline. This
phenomenon is more pronounced in mountains where rainwater or melted snow in upland areas
supply groundwater that moves downward through steeply dipping, permeable sedimentary rock
layers to reach great depths. Chemicals occurring naturally in groundwater include both inorganic
(e.g., salts, metals) and organic (carbon-based) types.
Salinity variation. Salinity is often the principal characteristic used to describe the overall quality of
groundwater. The term "fresh" groundwater often means groundwater containing no more than
1,000 milligrams per liter of total dissolved solids (mg/L TDS) but it is sometimes used to refer to
groundwater containing no more than 3,000 mg/L TDS fMaupin et al.. 2014: U.S. EPA. 2012e:
Freeze and Cherry. 1979al. When characterizing groundwater quality, scientists generally consider
the relative abundance of sodium, calcium, potassium, magnesium, chloride, bicarbonate, and
sulfate to account for the bulk of dissolved constituents (Freeze and Cherry. 1979a). Natural salinity
ranges from less than 100 mg/L to over 300,000 mg/L TDS (Lauer etal.. 2016: Clark and Veil.
20091. Higher salinity groundwater can contribute to palatability problems, and in the very high
salinity ranges, causes water to be unhealthful for human consumption fEllis. 19971. People have a
range of reactions to drinking water salinity. Some people object to the taste of drinking water
having comparatively lower salinity levels while other people reach this objection threshold at
higher salinity levels (Burlingame and Waer. 2002). Desalinating water containing salinity values of
10,000 mg/L TDS to render it potable is technically and economically feasible fEsser etal.. 20151.1
As a result, groundwater with salinity values up to 10,000 mg/L TDS is often defined as a protected
groundwater resource under several laws, including the regulations implementing the federal Safe
Drinking Water Act (SDWA) and the U.S. Bureau of Land Management (BLM) Onshore Order #2.
The complete basis and standards for defining a protected groundwater in all locations within the
United States is beyond the scope of this report. Additional information about protections given to
groundwater is described in Chapter 1 in Text Box 1-1.
Groundwater suitable for drinking is found within a large range of depths around the United States.
The groundwater quality profile with depth varies around the United States. Feth (1965) described
patterns in the relationship of depth to groundwater containing salinity ranging from 1,000 to
3,000 mg/L TDS.2 The patterns include: (1) large portions of the Southeast and middle Midwest
have at least 1,000 ft (300 m) of separation between the land surface and groundwater containing
1,000-3,000 mg/L TDS, and (2) significant portions of the Northeast, northern Midwest, and parts
of the West have less than 500 ft (200 m) separating the land surface from groundwater containing
1,000-3,000 mg/L TDS. The report does not contain information about the base or thickness of
groundwater having certain quality. As a result, these depths represent minimum distances
between the land surface and bottom depth of groundwater having this salinity range.
1 For instance, desalination of sea water (approximately 35,000 mg/L TDS] now occurs in Florida, California, and Texas.
2 Salinity and total dissolved solids are frequently interchangeable terms. The vast majority of dissolved constituents in
natural water are inorganic salts, although a minor fraction of dissolved constituents can be organic matter. Feth f19651
maps groundwater found at ranges of depth with spans of salinity. Singular depth and salinity values are not present on
the map.
2-6
-------
Chapter 2 - Drinking Water Resources in the United States
Methane in groundwater. Methane can be found naturally at detectable levels in groundwater
fKappel and Nvstrom. 2012: Eltschlager etal.. 2001: Coleman et al.. 19881. There are different
origins of methane in groundwater. Biogenic methane is produced at comparatively low
temperature and pressure from biologic decay of carbon-bearing matter, while thermogenic
methane is formed over geologic time when carbon-bearing matter is exposed to elevated pressure
and temperature conditions typically associated with deep burial (Baldassare etal.. 20141. Given
the buoyancy of natural gas, if a pathway exists or enough time is available, it can move upward and
accumulate at shallower depths. Natural gas found in small, uneconomic quantities in shallow zones
may have originated in place or may have migrated upward, and is often referred to as stray gas.
For more discussion about the issue of stray gas, see Text Box 6-3 in Chapter 6. When consumed in
drinking water, methane does not generally have human health effects,1 however, it is an explosive
gas if it comprises between 5% and 15% of a volume of air fAstle and Weast. 19841. If methane
from well water enters the atmosphere within a confined space under conditions that allow it to
concentrate, it can pose an explosive threat if it reaches this threshold.
2.2.1.2 Groundwater Quantity
Groundwater quantity can be characterized as the total subsurface water available, although a
practical limiting property is the rate at which groundwater can be withdrawn from the subsurface,
a property known as yield (Freeze and Cherry. 1979al. If rock formations in the subsurface contain
water within exceedingly small or poorly connected pore spaces, then the low yield may preclude
its practical use as a source of drinking water.
When recharge and discharge are in balance, the volume of groundwater existing in the subsurface
remains the same. Recharge and discharge also occur in connection with human-caused activity.
Groundwater recharge increases due to irrigation, underground injection wells, surface
impoundments, and dammed reservoirs, while groundwater discharge increases through well
withdrawals for irrigation, household use, etc. (Winter et al.. 1998). These activities can locally
affect the natural balance between groundwater recharge and discharge. Climatic variation that
changes precipitation rates also affects groundwater recharge rates, which in turn leads to changes
in subsurface groundwater volume fWinter et al.. 19981.
When an aquifer consistently yields water at rates suitable for human use, and the water is of good
enough quality to drink or be treated for drinking, it can serve as a source of drinking water.
2.2.2 Surface Water Resources
Surface water is that part of the hydrologic cycle that occurs on land surface and includes water in
the ocean as well as rainwater or meltwater. Surface water collects into depressions or along
channels in sufficient volume to create standing or running water all or much of the time. Non-
ocean surface water has often had little time to become saline, because much of it is not in direct
contact with anything other than more water in the surrounding surface water body. Non-ocean
surface water can quickly move into the next phase of the hydrological cycle, either evaporating
1 There is no enforceable drinking water standard established for dissolved methane in drinking water.
2-7
-------
Chapter 2 - Drinking Water Resources in the United States
into the atmosphere or infiltrating the subsurface. Because surface water is open to the atmosphere
and is generally located at the lowest points on a landscape, it is susceptible to contamination.
Contamination sources include atmospheric deposition, and run-off from urban land areas or lands
used for agricultural or industrial activities fWinter et al.. 19981. Many non-ocean surface water
bodies in the United States have a set of water quality standards based on their designated use,
which can include recreation, drinking water, supportive of aquatic life, fishery, industrial supply,
and other uses. In turn, National Discharge Pollution Elimination (NPDES) permits governing point
source discharge into the surface water bodies are issued under the Clean Water Act and contain
limits on pollutants designed to achieve these water quality standards.1 When taken together, these
permits are meant to ensure that the surface water achieves a water quality consistent with the
designated use.
2.2.2.1 Surface Water Quality
Studies conducted in connection with the National Water Quality Assessment Program show the
presence of human-made chemicals at low concentrations in the streams surveyed (Kingsbury et
al.. 20081.2 Based on dissolved solids alone, sampled streams range from less than 100 mg/L TDS to
more than 500 mg/L TDS fArming and Flvnn. 20141. Large lakes can range in salinity from less than
500 mg/L TDS to more than 200,000 mg/L. By comparison, ocean water has a salinity of about
35,000 mg/L TDS. Considering the vast array of possible chemical, biological, and radiological
content in surface water, it is beyond the scope of this report to describe in detail the surface water
qualities that exist in the United States.
2.2.2.2 Surface Water Quantity
About 7% of the surface area of the United States is covered by surface water, but it is not uniformly
distributed. The portion of the United States located east of the Mississippi River comprises about
25% of the total area, yet it contains about 42% of the total land area covered by surface water
(USGS. 2016: U.S. Census Bureau. 2012). The Great Lakes alone, located in the eastern half of the
United States, contain about one-fifth of the world's surface fresh water (Government of Canada
and U.S. EPA. 19951.3 In contrast, the western part of the United States has a lower proportion of
land covered by surface water with streams that tend to be more intermittent in nature.4 For
instance, 81 percent of the streams in Arizona, New Mexico, Nevada, Utah, Colorado, and California
are not permanent streams (Levick etal.. 2008). Certain parts of the western U.S. are presently
experiencing less surface water availability as indicated by declining water reservoir levels with
some reservoirs in the southwest currently below 50% of their capacity.5 For example, according to
1 Title 40, United States Code of Federal Regulations, Part 131, as of May 25,2016.
2 See USGS ("20121 for more information about this program.
3 Including the portion of the Great Lakes lying within Canada.
4 Not all western states follow this trend. Hawaii and Alaska, for instance, have a significantly higher percentage of land
mass covered by surface water (41% and 14%, respectively] than the national average.
5 See for instance U.S. DPI (~2016bl. California Department ofWater Resources ("20161. and SRP ("20161.
2-8
-------
Chapter 2 - Drinking Water Resources in the United States
the U.S. Department of the Interior (DOI), the largest capacity reservoir in the United States, Lake
Mead, holds about 37% of its volume capacity as of the fall of 2016 fU.S. DOI. 2016al
2.3 Current Drinking Water Sources
Drinking water is supplied to households and businesses by either public water systems (PWSs) or
non-public systems (non-PWSs).1 In 2010, approximately 270 million people (86% of the
population) in the United States relied on PWSs to supply their homes with drinking water fMaupin
etal.. 2014: U.S. EPA. 2013bl. These PWSs provided households with nearly 24 billion gal (91
billion L) of water per day (Maupin etal.. 2014).2 In areas without service by PWSs, approximately
45 million people (14% of the population) obtain drinking water from non-PWSs, using mostly
water wells. Non-PWSs account for about 3.6 billion gal (14 billion L) of daily water withdrawals
fMaupin et al.. 20141.3
Both groundwater and surface water serve as drinking water sources in the United States. Surface
water accounts for about 58% of all drinking water withdrawals and groundwater supplies the
remaining 42%. Table 2-1 portrays the relative abundance of surface water and groundwater as
sources for both publicly and non-publicly supplied drinking water.
Of the population receiving water supplied by PWSs, the relative importance of surface and
groundwater sources for supplying drinking water varies geographically (Figure 2-1). Most larger
PWSs rely on surface water and are located in urban areas fU.S. EPA. 2011cl. whereas most smaller
PWSs rely on groundwater and are located in rural areas (U.S. EPA. 2014h. 2013b). More than 95%
of households in rural areas obtain their drinking water from groundwater (U.S. EPA. 2011c).
PWSs are subject to routine monitoring and testing requirements required under the National
Primary Drinking Water Standards regulations, whereas no such monitoring or testing is required
for non-PWSs.4 The required monitoring and testing at PWSs ensures that the public has
information regarding the extent to which delivered water meets drinking water standards,
whereas users of non-PWSs (e.g., private water wells) make individual, voluntary decisions about
how often they monitor and test their water. Lack of monitoring may make non-PWS users more
vulnerable to contamination, if present, than PWS users.
1 PWSs provide water for human consumption from surface water or groundwater through pipes or other infrastructure
to at least 15 service connections or serve an average of at least 25 people for at least 60 days a year fU.S. EPA. 2012gl.
The EPA categorizes PWSs as either community water systems, which supply water to the same population year-round, or
non-community water systems, which supply water to at least 25 of the same people at least six months per year, but not
year-round. Non-public water systems (non-PWSs] have fewer than 15 service connections and serve fewer than 25
individuals fU.S. EPA. 19911. Non-PWSs are often private water wells supplying drinking water to a singular residence.
2 The USGS compiles data in cooperation with local, state, and federal environmental agencies to produce water-use
information aggregated at the county, state, and national levels. Every five years, data at the county level are compiled
into a national water use census and state-level data are published. The most recent USGS water use report was released
in 2014, and contains water use estimates from 2010. Water withdrawals are distinguished from and are greater than
water deliveries due to water loss during the process of delivering finished water fMaupin etal.. 2014: USGS. 2014b).
3 A withdrawal means the volume of water taken from its source regardless of how much of that volume is either returned
to the local hydrologic cycle or is consumed without being returned to the local hydrologic cycle.
4 See Title 40 of the Code of Federal Regulations, Part 141, promulgated pursuant to the SDWA.
2-9
-------
Chapter 2 - Drinking Water Resources in the United States
Table 2-1. Summary of drinking water sources in the United States in 2010.
The volume and percentages of daily domestic water withdrawals in the United States are shown by public and
non-public water systems, total withdrawal, and whether the source is surface water or groundwater. Volume is in
billions of gallons per day (Bgal/day) and percentages are of either water supply type or total volume withdrawn,
as indicated in italics. Some figures shown are rounded values. Source of data: Maupin et al. (2014).
Drinking water source
Public water
supply
Non-public
water supply
Total volume
withdrawn
Surface Water
Daily volume withdrawn (billion gallons)
26.3
0.1
26.4
Percent of water supply type
63
2
58
Groundwater
Daily volume withdrawn (billion gallons)
15.7
3.5
19.2
Percent of water supply type
37
98
42
Total
Daily volume withdrawn (billion gallons)
42.0
3.6
45.6
Percent of water supply type
92
8
100
Drinking water sources used by public water
systems, as a percentage of all sources in a state
h > 75% surface water sources
50 - 75% surface water sources
50 - 75% ground water sources
> 75% ground water sources
Projection: North American Albers Equal Area Conic
Figure 2-1. Geographic variability in drinking water sources for public water systems.
The relative importance of surface and groundwater as sources for public water systems varies by state. The public
water system sources used in this analysis include infiltration galleries, intakes, reservoirs, springs, and wells.
Sources: ESRI (2010). U.S. Census Bureau (2013). and U.S. EPA (2013b).
2-10
-------
Chapter 2 - Drinking Water Resources in the United States
2.3.1 Factors Affecting How Water Becomes a Drinking Water Source
The most common source of drinking water in the world, including in the United States, is fresh
water (see Section 2.2.1.1). There can be exceptions to the use of fresh water as a drinking water
source. For instance, projects in California, Florida, Arizona and Texas desalinate sea water or
brackish groundwater to produce drinking water.1 The principle of supply and demand that affects
availability of commercial products in the marketplace is also applicable to drinking water
resources. Water not considered a practical drinking water source under one demand condition
may become desirable as a drinking water source under different demand conditions. Text Box 2-2
presents El Paso, Texas as such an example.
Text Box 2-2. El Paso's Use of Higher Salinity Water for Drinking Water.
The El Paso Water Utility (EPWU] provides drinking water to over 600,000 people in the City of El Paso,
Texas and surrounding communities. Historically, the EPWU has withdrawn surface water from the Rio
Grande River and groundwater to meet water needs. Salinity from the freshwater aquifers typically ranges
between 300 and 1,000 mg/L TDS. With increases in population and periodic drought conditions stressing
the water supply, the EPWU instituted a number of different measures to diversify its water supply portfolio.
Components of the EPWU water supply portfolio include water conservation, surface water, groundwater
and, more recently, desalinating saline groundwater. Continued long-term pumping of fresh groundwater
allowed higher salinity groundwater to enter into one of EPWU's well fields from more saline parts of the
aquifer. This well field is now used as the source for the Kay Bailey Desalination Plant, which began operation
in 2007 and desalinates groundwater with salinity ranging from 1,000 and 5,000 mg/L TDS (El Paso Water
Utilities. 20161 The plant uses reverse osmosis technology to remove the high salt content thereby creating
additional fresh water supplies. Use of this higher salinity water supply has added approximately 25% more
water availability, decreasing the stress on the original fresh water supplies available to the EPWU and
highlights the potential value of groundwater that had not formerly been considered a drinking water source.
2.3.1.1 General Considerations Applicable to All Water as Source of Drinking Water
Factors to consider when assessing a possible source of drinking water include availability,
contaminants in the water, and the cost to obtain and treat water. Surface water in streams, lakes,
or reservoirs is almost always considered to be a source for drinking water, because they contain
fresh, readily accessible water. Groundwater is a critically important drinking water source in many
parts of the United States, especially where surface water is less abundant Challenges for use as
drinking water exist for both surface and groundwater. Surface water may not suffice as a drinking
water source when it exists only temporarily or cannot supply the volume demand. Both surface
water and groundwater may have contaminant levels that require expensive treatment technology.
For instance, in an extensive report, the USGS describes how human activities cause unnaturally
fast and deep groundwater movement, which degrades water quality over long periods in the
1 Brackish water is often a general term used for water having a salinity content intermediate between fresh water and
sea water, although it may also have a more specific definition, such as the 1,000 - 10,000 mg/L TDS value used in some
USGS publications.
2-11
-------
Chapter 2 - Drinking Water Resources in the United States
nation's principal aquifers (DeSimone etal.. 20141. Despite these challenges, changes in the demand
for water affect the consideration of sources of water for drinking purposes.
2.3.1.2 Considerations Applicable to Groundwater as a Drinking Water Source
Determining what groundwater is eligible for use as a drinking water source can include additional
challenges. Groundwater may be located at significant depth or within low-yield aquifers, requiring
additional engineering solutions to make them practical and/or cost effective as a drinking water
source. Aquifers, or parts of aquifers, not in use today for drinking water purposes may nonetheless
eventually be considered a drinking water source. The future viability of currently unused aquifers
depends on the definition of what constitutes a drinking water resource and knowledge of the
physical and chemical characteristics of the aquifers. The extent of knowledge about what exists in
the subsurface depends on extrapolation from limited subsurface data (e.g., water samples
collected from wells in, or passing through, aquifers). Although salinity is a common criterion for
designating an aquifer as a drinking water resource (see Section 2.2.1.4), there is not a uniform
threshold value for making that determination. The Groundwater Protection Council (GWPC) notes:
There is a great deal of variation between states with respect to defining protected
groundwater. 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 groundwater, and
the actual use of groundwater fGWPC. 2009I1
In addition to variation in applicable water quality criteria, the availability of information regarding
groundwater that meets an applicable criterion (if one exists) is also variable. For instance, the
bottom depth of aquifers or parts of aquifers that may be defined as a drinking water resource are
not always readily publicly available. In some locations, such as the State of Texas, estimates of the
bottom depth of groundwater meeting certain regulatory threshold criteria are made public on a
website.2 In other parts of the United States the depth of identified protected subsurface drinking
water resources may not be publicly available. No centralized compilation of groundwater depth
and quality exists for all locations in the United States, nor does such a reference exist for depths to
protected groundwater resources. The depths to protected groundwater resources can vary. In one
example, the EPA described the reported bottom depths of protected groundwater resources as
ranging from just below ground surface to 8,000 ft (2,000 m) (U.S. EPA. 2015n).3
Even in regions where the bottom depth of protected groundwater resources are generally known,
there can remain uncertainty regarding precise depths at specific locations. Examples include the
states of Indiana and Michigan according to the EPA Region 5 Underground Injection Control (UIC)
1 An underground source of drinking water (USDW] is defined in the federal regulations that implement the UIC program.
A USDW is generally considered to be any aquifer, or its portion, that currently serves as a source for a public water
system; or which contains enough groundwater to supply a public drinking water system, and either now supplies water
for human consumption, or contains fewer than 10,000 mg/L TDS. See Title 40 of the Code of Federal Regulation, Section
144.3.
2 See http://www.beg.utexas.edu/sce/index.html.
3 This reference provided 1,000-foot (305 meters] depth resolution for the reported base of protected groundwater.
2-12
-------
Chapter 2 - Drinking Water Resources in the United States
program, the State of Utah according to the Utah Geological Survey, and the State of California
according to the California State Water Resources Board fEsser etal.. 2015: Anderson etal.. 2012:
U.S. EPA. 2012el. In these examples, the depth to groundwater meeting the salinity threshold
necessary for decision-making is stated not to be known with precision, and collection of additional
groundwater quality information is advised.1
2.4 Future Drinking Water Sources
The future availability of fresh drinking water sources in the United States (Section 2.2.1.1) will
likely be affected by changes in climate and water use (Georgakakos etal.. 2014). Since 2000, about
30% of the total area of the contiguous United States has experienced moderate drought conditions
and about 20% has experienced severe drought conditions (National Drought Mitigation Center.
2015: U.S. EPA. 2015p). Declines in surface water resources have already led to increased
withdrawals and cumulative net depletions of groundwater in some areas fCastle etal.. 2014:
Georgakakos etal.. 2014: Konikow. 2013: Famiglietti etal.. 2011). Loss of approximately 240 mi3
(1,000 km3) of groundwater between 1900 and 2008 has been documented by the USGS. USGS
reports that about 20% of that loss occurred in the final eight years of that timeframe and that
depletion is greater in the arid and semi-arid western states than in the more humid eastern states
f Konikow. 20131. Other sources of water that might not be considered fresh, such as wastewater
from sewage treatment plants, brackish and saline surface and groundwater, as well as sea water,
are also increasingly being used to meet water demand. Through treatment or desalination, these
water sources can reduce the use of high-quality, potable fresh water for industrial processes,
irrigation, recreation, and toilet flushing (i.e., non-potable uses). In addition, in 2010, approximately
355 million gal per day (1.3 billion L per day) of treated wastewater was reclaimed through potable
reuse projects (NRC. 2012). Such projects use reclaimed wastewater to augment surface drinking
water sources or to recharge aquifers that supply drinking water to PWSs (NRC. 2012: Sheng.
20051. In 2007, among approximately 13,000 desalination plants worldwide, there existed the
capacity to produce about 14.7 billion gal (55.6 billion L) of fresh water each day. In 2005, the
United States had approximately 11 % of that volume capacity (Gleick. 2008: Coolev etal.. 2006).
An increasing number of states are developing new water supplies to augment existing drinking
water sources through reuse of reclaimed water, recycling of storm water, and desalination fU.S.
GAP. 20141. Most desalination programs currently use brackish water as a source, although plans
are underway to expand the use of sea water. States with the highest installed capacity for
desalination include Florida, California, Arizona, and Texas (Coolev et al.. 2006). It is likely that
various water treatment technologies will continue to expand drinking water sources beyond those
that are currently being considered. In addition to treatment technologies, there are efforts by
public water systems to alleviate demand on drinking water supplies such as encouraging more
modest consumer water usage and repairing leaks in water infrastructure.
1 Decisions dependent on knowledge of threshold salinity values in groundwater can include permitting injection wells
and oil and gas production well construction design approvals.
2-13
-------
Chapter 2 - Drinking Water Resources in the United States
2.5 Proximity of Drinking Water Resources to Hydraulic Fracturing Operations
Hydraulic fracturing in oil and gas production wells necessarily takes place where oil and gas
resources are located. The relative locations of drinking water resources influences the degree to
which they may be affected by activities in the hydraulic fracturing water cycle. With increased
proximity, hydraulic fracturing activities have a greater potential to affect surface and subsurface
sources of current and future drinking water (Vengosh etal.. 2014: Entrekin etal.. 20111. To
estimate potentially vulnerability populations that use drinking water resources, the EPA
performed an analysis of the number of hydraulically fractured production wells that are located
within 1 mi (1.6 km) of a PWS source. The EPA also presents subsurface separation distances
between the depths of drinking water resources and hydraulic fracturing in production wells.
2.5.1 Lateral Distance between Public Water System Sources and Hydraulic Fracturing
The EPA analyzed the locations of the approximately 275,000 oil and gas wells that were assumed
to be hydraulically fractured in 25 states between 2000 and 2013 (Chapter 3) to determine the
number of fractured wells within a 1-mile radius of facilities that withdraw water for a PWS.1-2-3
Based on 2000-2013 Drillinglnfo data, the lateral distance from the nearest facility that withdraws
water for PWS to a hydraulically fractured well ranged from 0.01 to 41 mi (0.02 to 66 km), with an
average distance of 6.2 mi (10.0 km) and a median distance of 4.8 mi (7.7 km) (Drillinglnfo. 2014a:
U.S. EPA. 2014h). Of the approximately 275,000 wells that were estimated to have been
hydraulically fractured in 25 states between 2000 and 2013, an estimated 21,900 (8%) were within
1 mile of at least one PWS groundwater well or surface water intake. Most of these approximately
6,800 individual facilities that withdraw water for a PWS were located in Colorado, Louisiana,
Michigan, North Dakota, Ohio, Oklahoma, Pennsylvania, Texas, and Wyoming (Figure 2-2). These
facilities that withdraw water for a PWS had an average of seven hydraulically fractured production
wells and a maximum of 144 such production wells within a 1-mile radius. These water sources
supplied water to 3,924 PWSs—1,609 of which are community water systems—that served more
than 8.6 million people year-round in 2013 fU.S. EPA. 2014h: U.S. Census Bureau. 2013: U.S. EPA.
2013b").4
!The EPA estimated the number of oil and gas production wells hydraulically fractured between 2000 and 2013. To do
this, EPA assumed that all horizontal wells were hydraulically fractured in the year they started producing and assumed
that all wells within a shale, coalbed, or low-permeability formation, regardless of well orientation, were hydraulically
fractured in the year they started producing. More details are provided in U.S. EPA f2013cl Not all coalbed methane wells
are hydraulically fractured, but coalbed methane wells represent production wells that sometimes uses hydraulic
fracturing. Given that there were 15% of coalbed methane wells relative to all hydraulically fractured wells and the lack of
data that distinguishes whether or not coalbed wells are hydraulically fractured, EPA included coalbed wells into all
counts of wells that are hydraulically fractured.
2 The selected 1-mile distance used in this analysis provides a consistent approach. Local topographic conditions could
support the use of a different analysis at any specific site.
3 A facility that withdraws water for a PWS includes water intakes, water wells, springs, infiltration galleries, and
reservoirs. It is common for a PWS to operate multiple individual facilities to withdraw the cumulative water supplied by
the PWS.
4 All PWS types were included in the locational analyses performed. However, only community water systems were used
to calculate the number of customers obtaining water from a PWS with at least one source within 1 mile of a hydraulically
2-14
-------
Chapter 2 - Drinking Water Resources in the United States
Number of hydraulically
fractured wells within 1
mile of a public water
system (PWS) source
1 -8
9-24
• 25-48
• 49-144
0 125 250 500 A
Miles
Source data credits: Drillinglnfo;
U S Environmental Protection Agency
Basemap credits: ESRI
Figure 2-2. The location of public water system sources having hydraulically fractured wells
within 1 mile.
Points indicate the location of public water system (PWS) sources; point color indicates the number of hydraulically
fractured wells within 1 mile of each PWS source. The estimates of wells hydraulically fractured from 2000 to 2013
developed from the Drillinglnfo data were based on assumptions described in Chapter 3. Sources: Drillinglnfo
(2014). U.S. EPA (2013b), and ESRI (2010).
The EPA also analyzed the location of hydraulically fractured wells relative to populations where a
high proportion (>30%, or at least twice the national average] obtain drinking water from non-
PWSs (mostly private groundwater wells].1 Based on Drillinglnfo well location data and USGS
drinking water data, between 2000 and 2013, approximately 3.6 million people live in counties
fractured well. If non-community water systems are included, the estimated number of customers increases by 533,000
people (U.S. EPA. 2012gl A community water system is a PWS which serves at least 15 service connections used by year-
round residents or regularly serves at least 25 year-round residents.
'There is no national data set of non-PWSs. In Maupin etal. f 20141 the USGS estimates the proportion of the population
reliant on non-PWSs, referred to as the "self-supplied population," by county, based on estimates of the population
without connections to a public water system. The USGS estimates were used for this analysis.
2-15
-------
Chapter 2 - Drinking Water Resources in the United States
with at least one hydraulically fractured well and where at least 30% of the population relies on
non-PWSs for drinking water fDrillinglnfo. 2014a: USGS. 2014b]. The population changes to
approximately 740,000 people living in counties with more than 400 hydraulically fractured wells
and at least 30% of the population relies on non-PWSs for drinking water fDrillinglnfo. 2014a:
USGS. 2014b).1The counties having more than 400 hydraulically fractured wells and at least 30% of
the population relying on non-PWSs for drinking water were located in Colorado, Kentucky,
Michigan, Montana, New Mexico, New York, Oklahoma, Pennsylvania, Texas, and Wyoming.
As described in Chapter 1, this assessment defines five stages in the hydraulic fracturing water
cycle. The lateral distance analysis presented here relates to the wellhead locations of hydraulically
fractured production wells, and therefore addresses three stages that take place near production
wellheads, evaluated in Chapters 5, 6, and 7, respectively (chemical mixing, well injection, and
produced water handling).2 A lateral distance analysis was not possible for the other two stages
(water acquisition, wastewater disposal and reuse) because there is a lack information about where
water is acquired for hydraulic fracturing and where the wastewater from any given hydraulically
fractured well is disposed or treated.
2.5.2 Vertical Distance between Drinking Water Resources and Hydraulic Fracturing
The depth at which hydraulic fracturing takes place varies depending on the depth to the targeted
production zone. For instance, in a study of wells representing approximately 23,000 production
wells hydraulically fractured by nine service companies during 2009 and 2010, the EPA found that,
when measured vertically from the surface to total depth, well depths ranged from less than 2,000
ft (600 m) to more than 11,000 ft (3,000 m) (U.S. EPA. 2015nl Similarly, based on more than
38,000 hydraulic fracturing disclosures to the FracFocus registry website, the middle 90% of these
well disclosures had vertical depths between 2,900 and 13,000 ft (880 and 4,000 m) with a median
value of about 8,100 ft (2,500 m) fU.S. EPA. 2015al. Hydraulic fracturing can occur at or near the
bottom of a production well or it may take place at different intermediate depths depending on the
location of economically producible oil and gas, and thus the total vertical depth of a production
well does not necessarily correlate to the depth at which hydraulic fracturing occurs (Chapter 6).
Hydraulic fracturing has been conducted at depths ranging from less than 1,000 ft (300 m) to
greater than 10,000 ft (3,000 m) depth fU.S. EPA. 2015n: NETL. 20131. The distance from the base
of the drinking water resource to the shallowest hydraulic fracturing initiation point in a
production well serves as a separation distance.3 The EPA reports separation distances in depth
measured along the well ranging from no separation distance (where hydraulic fracturing took
1 Approximately 14% of the U.S. population is self-supplied by non-PWSs fMaupin et al.. 20141 This analysis considers
only counties in which more than double the national average—that is, at least 30% of the county's population—was
supplied by non-PWSs.
2 Chapter 7 (Produced Water Handling] examines potential effects on drinking water resources at hydraulically fractured
wellhead locations, as well as away from wellhead locations.
3 If measured vertically from the shallowest hydraulic fracturing initiation point to the bottom of the drinking water
resource, this is referred to as a vertical separation distance. If measured along a borehole from the shallowest hydraulic
fracturing initiation point to the bottom of the drinking water resource, this is referred to as a separation distance in
measured depth.
2-16
-------
Chapter 2 - Drinking Water Resources in the United States
place at depths shallower than the reported base of the drinking water resource) to more than
10,000 ft (3,000 m) fll.S. RPA. 2015n"l.
In a given setting it is the geologic and hydrologic history that determines the depths to potential
oil and gas and/or subsurface drinking water resources. In some settings, rock formations bearing
economic quantities of oil or gas also contain groundwater that, based on salinity value alone,
qualifies it as a drinking water resource. Large distances vertically separate these two resources in
other settings. Figure 2-3 depicts two different types of these settings.
Drinking Water Resource
Separation
Distance in
Measured Depth
Vertical
Separation
Distance
(b) O,
1
X.
^ Drinking Water Resource
, No Vertical
. f Separation
Drinking Water Resource
and Targeted Rock Formation
Targeted Rock Formation
15,000
10,000
5,000
(C)
T II 1
ii i I
I I = .
t# cf? cf? <$> <# dp <*? $ cS?
* ^ «jP # # # «§? rf? # *
V V *V V "V >
Separation Distance in Measured Depth (feet)
Figure 2-3. Separation distance between drinking water resources and hydraulically fractured
intervals in wells
Schematic examples showing a relatively large separation distance (panel a) and the absence of any separation
distance (panel b) between the shallowest fracture initiation depth in a well to the base of the protected drinking
water resource. Distances may be presented as vertical or as a measured distance along a non-vertical well. Panel c
shows result from wells studied representing approximately 23,000 production wells hydraulically fractured
between 2009 and 2010 (U.S. EPA, 2015n). Error bars in panel c display 95% confidence intervals.
In Figure 2-3, panel (a), the hydraulically fractured oil- and gas-bearing zone is much deeper than
drinking water resources, therefore separation distance is large. In panel (b), the hydraulically
fractured oil- and gas-bearing zone is at the same depth as drinking water resources and there is no
separation. The lack of separation distance can be due to the oil- and gas-bearing zone being
shallow and/or the drinking water resource being deep. Panel (c] illustrates the distribution of
separation distances in measured depth for study wells representing approximately 23,000 oil and
gas production wells hydraulically fractured by nine service companies between 2009 and 2010, as
2-17
-------
Chapter 2 - Drinking Water Resources in the United States
reported in U.S. EPA (2015n). The calculation of 95% confidence intervals shown in panel (c) is
described in the EPA report and was affected by the number of companies in the study and the well
file selection methods.
2.6 Conclusions
Drinking water resources provide the water humans consume, cook with, bathe in, and need for
other purposes. An estimated 86% of the United States population derives its household drinking
water from PWSs that serve at least 25 people. The remaining 14% self-supply their homes with
drinking water from non-PWSs, which are largely private water wells. Publicly supplied drinking
water is subject to monitoring and testing to determine compliance with drinking water standards
while no such monitoring and testing is required at non-PWSs. Surface water is the source for an
estimated 58% of the volume needed to supply drinking water and groundwater is the source for
the remaining 42%.
The existing distribution and abundance of the drinking water resources in the United States may
not be sufficient in some locations to meet future demand. The future availability of sources of
drinking water that are considered fresh will likely be affected by changes in climate and water use.
Since at least 2000, many areas of the United States have experienced significant drought, which
often correlate with diminishment of ground and surface water supplies in these areas. Locally,
measures are now being implemented to prolong use of current drinking water sources such as
encouraging more modest drinking water use and using treated wastewater or other non-potable
water sources to help meet demand.
Between 2000 and 2013, the EPA estimates there were approximately 275,000 oil and gas
production wells hydraulically fractured in 25 states. To produce a consistent measure of proximity
between these hydraulically fractured oil and gas production wells and drinking water resources
during this time frame, the EPA counted the number hydraulically fractured oil and gas production
wells located within 1 mile of public drinking water sources, and performed a count of the counties
with a relatively high reliance on self-supplied drinking water that also contain one or more of
these hydraulically fractured production wells. Between 2000 and 2013, approximately 3,900
public water systems had between one and 144 wells hydraulically fractured within 1 mile of their
water source; these public water systems served more than 8.6 million people year-round in 2013.
An additional 740,000 people between 2000 and 2013 self-supplied their drinking water in
counties where at least 30% of the population relies on groundwater and having at least 400
hydraulically fractured wells.
Depending on the nature of the geologic setting, hydraulically fractured oil and gas production
wells can be located near where people get their drinking water. Depths to hydraulically fractured
oil and gas resources can range from less than 1,000 ft (300 f) to more than 10,000 ft (3,000 m)
while drinking water resources may be found between a few tens of feet to as much as 8,000 ft
(2,000 m) below the surface. There is limited publicly available information to determine the
vertical distance separating the shallowest hydraulic fracturing initiation point in a production well
from the deepest drinking water resource. The EPA found, among 323 wells studied statistically
representing more than 23,000 production wells hydraulically fractured by nine service companies
between 2009 and 2010, the distance along the wells between these two resources ranged from
none to more than 10,000 ft (3,000 m).
2-18
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Chapter 3. Hydraulic Fracturing for
Oil and Gas in the United States
Abstract
This chapter provides a general description of the practice of hydraulic fracturing, where it is conducted,
how prevalent it is, and how hydraulic fracturing-based oil and gas production fits into the context of
energy production in the United States. Some of the information in this chapter also serves as an
introduction to the more in-depth technical chapters in the assessment.
Hydraulic fracturing is a technique used to increase oil and gas production from underground oil-
and/or gas-bearing rock formations (reservoirs). The technique involves the injection of hydraulic
fracturing fluids through the production well and into the reservoir under pressures great enough to
fracture the reservoir rock. Hydraulic fracturing fluids typically consist mainly of water, a "proppant"
(typically sand) that props open the created fractures, and additives (usually chemicals) that modify the
properties of the fluid for fracturing. Fractures created during hydraulic fracturing enable better flow of
oil and gas from the reservoir into the production well. Water that naturally occurs in the oil and gas
reservoirs also typically flows into and through the production well to the surface as a byproduct of the
oil and gas production process.
Since the mid-2000s, the combination of modern hydraulic fracturing and directional drilling has
become widespread and significantly contributed to a surge in oil and gas production in the United
States. Slightly more than 50% of oil production and nearly 70% of gas production in 2015 is estimated
to have occurred using hydraulic fracturing. Hydraulic fracturing is widely used in unconventional (low
permeability) oil and gas reservoirs that include shales, so-called tight oil and tight gas formations, and
coalbeds, but it is also used in conventional reservoirs.
There is no comprehensive national database of wells that are hydraulically fractured in the United
States. Using data from several commercial and public sources, the EPA estimates that 25,000 to 30,000
new wells were drilled and hydraulically fractured in the United States annually between 2011 and
2014. These hydraulic fracturing wells are geographically concentrated; in 2011 and 2012 almost half of
hydraulic fracturing wells were located in Texas, and a little more than a quarter were located in the
four states of Colorado, Pennsylvania, North Dakota, and Oklahoma.
New drilling activity for hydraulic fracturing wells is generally linked with oil and gas prices, and those
peaked in the United States between 2005 and 2008 for gas and between 2011 and 2014 for oil.
Following price declines, the number of new hydraulically fractured wells in 2015 decreased to about
20,000. Despite recent declines in prices and new drilling, U.S. gas and oil production continues at levels
above those in recent decades, and production for both is predicted to continue growing in the long
term, led by hydraulic fracture-based production from unconventional reservoirs.
3-1
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
This page is intentionally left blank.
3-2
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
3. Hydraulic Fracturing for Oil and Gas in the
United States
3.1 Introduction
This chapter provides general background information on hydraulic fracturing and will help the
reader understand the in-depth technical chapters that follow. We describe the purpose and
process of hydraulic fracturing and the situations and settings in which it is used (Section 3.1). Then
we provide a general description of activities at a hydraulic fracturing well site including assessing
and preparing the well site, well drilling and construction, the hydraulic fracturing event, the oil
and gas production phase, and eventual site closure (Section 3.3). A characterization of the
prevalence of hydraulic fracturing in the United States is then presented (Section 3.4), followed by a
review of its current and future importance in the oil and gas industry and its role in the U.S. energy
sector (Section 3.5), and a brief conclusion (Section 3.6).
3.2 What is Hydraulic Fracturing?
Hydraulic fracturing is a technique used to increase oil and gas production from underground oil-
or gas-bearing rock formations (reservoirs).1 The technique involves the injection of hydraulic
fracturing fluids through the production well and into the reservoir under pressures great enough
to fracture the reservoir rock. The injected hydraulic fracturing fluid carries "proppant" (typically
sand) into the fractures so that they remain propped open after the pressurized injection is
stopped. In addition to water, which typically makes up most of the injected fracturing fluid, the
fluid also contains chemical additives (additives) that serve a variety of purposes. These additives,
for example, can increase the fluid viscosity (how "thick" the fluid is) so that it carries the proppant
into the fractures more effectively, can help control well corrosion, can help minimize microbial
growth in the well, and so on (King and Durham. 2015: Gupta and Valko. 2007). The resulting
fractures enable better flow of oil and gas from the reservoir into the production well. Water that
naturally occurs in the reservoirs also typically flows into and through the production well to the
surface as a byproduct of the production process.
Although hydraulic fracturing is not new, how and where it is employed has changed (Text Box
3-1). For about a half-century after its introduction in the late 1940s, it was used to increase
production from vertical wells in conventional oil and gas reservoirs. Conventional reservoirs
develop over geologic time (many millions of years) when naturally buoyant oil and gas very slowly
migrate upward from the shale rock formations in which they formed until they are trapped by
geologic formations or structures and accumulate under a confining layer (Figure 3-1). As the oil
and gas accumulate, the pressure may increase. If the reservoir is under enough pressure and has
1A version of hydraulic fracturing, sometimes called hydrofracturing or hydrofracking, can be used to increase water
yields from water wells and is typically done by injecting only water under pressure. This application of hydraulic
fracturing is out of the scope of this assessment.
3-3
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Text Box 3-1. Hydraulic Fracturing: Mot New, but Different and Still Changing.
From the mid-1800s to the 1940s, operators of oil and gas wells occasionally tried to increase production by
pumping fluids or sometimes dropping explosives into wells. In the late 1940s, a fracturing technique to
increase production was patented by the Stanolind Oil and Gas Company and licensed to the Halliburton Oil
Well Cementing Company (Montgomery and Smith, 20101. Close to 1 million wells were hydraulically
fractured from the late 1940s to about 2000 flOGCC. 2002], The typical well design and hydraulic fracturing
operations during most of that time, though, were very different from today's modern hydraulic fracturing
operations.
The groundwork for the transformation to modern hydraulic fracturing was laid in the 1970s and early
1980s. Public-private research and development (R&D] partnerships that included industry, the Department
of Energy, and the Gas Research Institute were established because large amounts of natural gas were known
to occur in some shale rock formations yet traditional production well technology was not able to recover
much of the gas (Avila. 19761. These R&D programs played a key role in advancing technologies such as deep
horizontal drilling and fracturing with higher water volumes that ultimately enabled production from shales
and other unconventional sources of gas and oil (DOE. 2015: NRC; Committee on Benefits of DOE R&D on
Energy Efficiency and Fossil Energy. 20011. During this period, the U.S. Congress began offering tax incentives
for producers to use the developing technologies in the field (Wang and Krupnick. 2013: EIA. 2011a: Yergin.
20111. Advances in directional drilling technologies led to the first horizontal wells being drilled in the mid-
1980s in the Austin Chalk oil-bearing rock formation in Texas (Pearson. 2011: Havmond. 19911. Directional
drilling and other technologies matured in the late 1990s. In 2001, the Mitchell Energy company developed a
cost-effective technique to fracture the Barnett Shale in Texas. The company was bought by Devon Energy, a
company with advanced experience in directional and horizontal drilling, that, in 2002, drilled seven wells
and developed in the Barnett Shale using the combination of horizontal drilling and hydraulic fracturing; fifty-
five more wells were completed in 2003 (Yergin. 20111. The techniques were rapidly adopted and further
developed by others fDOE. 2011b: Montgomery and Smith. 20101. By 2005, the techniques were being used
in unconventional (low-permeability) oil and gas reservoirs outside of Texas. Modern hydraulic fracturing
quickly became the industry standard, driving a surge in U.S. production of oil and natural gas.
Hydraulic fracturing techniques and technologies continue to evolve. Wells are being drilled with longer
horizontal sections and are more closely spaced. Multiple, horizontal sections extending from a single vertical
well enable production from larger subsurface areas from a single well pad on the land's surface. These
historic and continuing technological developments enable production from previously unused oil and gas-
bearing geologic formations, altering and expanding the geographic range of oil and gas production activities.
Left: Early hydraulic fracturing site, late 1940s (source: Halliburton, used with permission). Right: Contemporary
hydraulic fracturing operation, late 2000s (source: NYSDEC (2015), used with permission).
3-4
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
adequate natural permeability, the economic extraction of oil and/or gas may only require using a
drilled well to bring the oil or gas to the surface.1
If the natural pressure is not high enough for the oil and gas to readily flow to the surface, various
pumping and "lift" techniques can be used to help the oil and gas move up the well to the surface
fHvne. 20121. In other situations, operators may pump water or a mix of water and carbon dioxide
(or other similar mixtures) into the reservoir through injection wells to help move and enhance the
extraction of oil and gas through nearby production wells. These techniques address pressure and
fluid characteristics in the reservoir, are not designed to fracture the reservoir rock, and therefore
are production-increasing techniques that are distinct from hydraulic fracturing. The discussions in
the remainder of this chapter focus on hydraulic fracturing in unconventional reservoirs.
Hydraulic fracturing is now combined with directional drilling technologies to access oil and gas in
unconventional reservoirs (although hydraulic fracturing is still used in conventional reservoirs,
too).2 Unconventional reservoirs have a very low natural permeability, which prevents oil and gas
from flowing through the rock into wells in economic amounts. Production from unconventional
reservoirs becomes economically feasible when wells, typically horizontal or deviated, are drilled
and hydraulically fractured through long portions of the production zone (the targeted oil- and gas-
bearing zones within a reservoir). See Figure 3-1 for a diagram of horizontal and other well types
and the reservoir types from which they can produce. Text Box 3-2 provides a brief discussion on
the use of the terms conventional and unconventional.
More details about the geologic formations that can be unconventional reservoirs are presented
below:
• Shales. Some organic-rich black shales serve as the source of oil and gas found in
conventional resources when, over geologic time, the lighter and more buoyant oil and gas
migrate upward from these shales and become trapped under impermeable confining
layers (Figure 3-1). Shales have very low permeability and the oil and gas are contained in
poorly connected pore space in the shale rock. With hydraulic fracturing and directional
drilling now enabling oil and gas production from very low permeability formations, some
of these shale source rocks are now unconventional reservoirs in addition to being
sources. Some shales produce predominantly gas and others predominantly oil; often
there will be some co-production of gas from oil wells and co-production of liquid oil from
gas wells fUSGS. 2013a: FTA. 201 lal.
• Tight formations. Some oil- and gas-bearing sandstone, siltstone, and carbonate
formations can be referred to as "tight" formations (for example, "tight sands") because of
their relatively low permeability and the fact that oil and gas are contained in small, poorly
connected pore spaces. Given a range of permeabilities, some tight formations require
1 Permeability in rocks is the ability of fluids, including oil and gas, to flow through well-connected pores or small
openings in the rock.
2 Directional drilling is the practice of controlling the direction and deviation (angle] of a borehole during drilling to
extend the borehole in a predetermined orientation and to a targeted area in the subsurface. Directional drilling is
required for drilling a deviated or horizontal well and is common in unconventional reservoirs. The terms deviated wells
and directional wells are often used interchangeably.
3-5
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
hydraulic fracturing for economic production and some do not. In the literature, "tight gas"
generally refers to gas in tight sands and is distinguished from "shale gas." Oil resources
from shale and other tight formations, in contrast, are frequently referred together under
the label "shale oil" or "tight oil" fSchlumberger. 2014: [JSCS. 2014a).
Coalbeds. Organic-rich coal, found in coalbeds, can be a source of methane (natural gas].
The gas primarily adheres to the coal surface rather than being contained in pore space or
structurally trapped in the formation. A range of techniques can be used to extract
methane from coalbeds and these techniques sometimes, but not always, employ hydraulic
fracturing. A key component of all coalbed methane production is the need to "dewater"
the coalbeds (pumping out naturally occurring or injected water) to reduce the pressure in
the coal allowing the methane to be released and flow from the coal into the production
well (Palmer. 2010: Al-Tubori etal.. 2009: IJSGS. 20001.
land surface
vertical
well
deviated
well
horizontal
well
coalbed
with methane
tight sand
gas
confining
sandstone
tight sand
oil
Figure Not to Scale
migration of oil
and gas over
geologic time
Figure 3-1. Conceptual illustration of the types of oil and gas reservoirs and production wells
used in hydraulic fracturing.
A vertical well is producing from a conventional oil and gas reservoir (right). The impermeable gray confining layer
(sometimes called a cap rock) traps the lighter and more buoyant gas (red) and oil (green) as it migrates up from
the deeper oil- or gas-rich shale source rock. Also shown are wells producing from unconventional reservoirs: a
horizontal well producing from a deep shale (center); a vertical well producing methane (gas) from coalbeds
(second from left); and a deviated well producing from a tight sand reservoir (left). Multiple deviated or horizontal
wells can be constructed and operated from a single well site. Note that the oil- or gas-rich shale serves as both a
source and a reservoir. Modified from Schenk and Poilastro (2002 and Newell (2011).
3-6
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Text Box 3-2. "Conventional" Versus "Unconventional."
The terms "conventional" and "unconventional" are widely used in articles and reports to distinguish types of
oil and gas reservoirs, wells, production techniques, and more. In this report, the terms are mainly used to
distinguish different types of oil and gas reservoirs: "conventional" reservoirs are those that can support the
economically feasible production of oil and gas using long-established technologies, and "unconventional"
reservoirs are those in which production has become economical only with the advances that have occurred
in hydraulic fracturing (often combined with directional drilling) in recent years.
Note that as hydraulic fracturing has increasingly become a standard industry technique, the word
"unconventional" is less apt than it once was to describe these oil and gas reservoirs. In a sense, "the
unconventional has become the new conventional" fNETL. 20131
The following three maps show the locations of major shale gas and oil resources, tight gas
resources, and coalbed methane resources, respectively, in the contiguous United States (Figure
3-2, Figure 3-3, and Figure 3-4). To explain the terminology used in the maps: a group of known or
possible oil and gas accumulations in the same region and with similar geologic characteristics can
be referred to as a play (Schlumberger. 2014). Plays can sometimes be geologically layered atop one
another (or "stacked") and are located in broad depressions filled with sedimentary rock
formations in the earth's continental crust known as basins. A group of similar coalbed methane
(gas) reservoirs can be referred to as coalbed methane fields (rather than plays) and are also found
in basins. The plays and fields in the maps below represent unconventional reservoirs that are
being exploited now or could be exploited in the future using hydraulic fracturing.
There is a wide range of depths at which hydraulic fracturing occurs across the country. For
example, approximate average depths for some of the largest gas-producing reservoirs are as deep
as 6,000 ft (2,000 m) in the Marcellus Shale in Pennsylvania and West Virginia, 7,500 ft (2,300 m) in
the Barnett Shale in Texas, and 12,000 ft (3,700 m) for the Haynesville-Bossier Shale in Louisiana
and Texas fNETL. 2013).1A few other, smaller plays are shallower, with depths less than 2,000 ft
(600 m) in parts of the Antrim (Michigan), Fayetteville (Arkansas), and New Albany (Indiana and
Kentucky) shale plays (NETL. 2013: GWPC and ALL Consulting. 2009). Coal seams that can be
drilled to produce gas (coalbed methane) range in depth from less than 600 ft (200 m) to more than
6,000 ft (2,000 m) with production often occurring at depths between 1,000 and 3,000 ft (300 and
900 m) (U.S. EPA. 2006: ALL Consulting. 2004). Coalbed methane production occurs in the San Juan
Basin in New Mexico, the Powder River Basin in Wyoming and Montana, and the Black Warrior
Basin in Alabama and Mississippi. See Chapter 6 for more information on the general locations and
depths of formations being hydraulically fractured.
1 These are approximate average depths; hydraulic fracturing occurs in shallower and deeper zones in all these plays.
3-7
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Bakken
Wiliiston
—Basin—
lichigan
Powder
- River
Basin
Antrim
obrara
d others
forest
\ City
Uinta-
Pi;ceance
iBasin
Illinois
Basin
San \
Joaquin \
Basin
lonterey
Appalac i
Bask
^ Lewis Pierre
I Raton
j San Juan Basin
/ Basin
tin
i/oodford
Los
Angeles
Basins
Arkoma Basin
Vat\rior
Basin
Fort
Worth
Basin
Bone Spi
Haynesville
Western
Gulf
Basiny
Major Shale Gas & Oil Plays
Current plays
Prospective plays
Basin
500
I Miles
Source data credits: U.S. Energy Information Administration (EIA)
Basemap Credits: U.S. Census Bureau, Esri, DeLorme, GEBCO, NOAA
NGDC, and other contributors
Figure 3-2. Major shale gas arid oil plays in the contiguous United States.
The plays represent geologically similar accumulations of oil and gas that are or could be developed. Adapted from EIA (2015).
3-8
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
yi 11 is ton
Basin
' Mesaverde
j Lance
and others
Niobrara
Mancos and others
Denver
Basin
Tuscarora
and others
Uinta-
Piceance
Basin
)palachian
Basin ,
Anadarko
Basin
Sap Juan
Basin
(Permian
' Basin
Fort-
Worth
Basin
1%LA-MS
Sah Basin
Austin
Chalk
Western
Gulf
Basing
Major Tight Gas Plays
Tight gas plays
Basins
375 500
M iles
Source data credits: U.S. Energy Information Administration (EIA)
Basemap Credits: U.S. Census Bureau, Esri, DeLorme, GEBCO, NOAA
NGDC. and other contributors
Figure 3-3. Major tight gas plays in the contiguous United States.
The plays represent geologically similar accumulations of gas that are or could be developed. Adapted from EIA (2011b).
3-9
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
terGree
;r Basin
I Uinta
' Basin
Cherokei
Platfori
Arkoma4M$
Basiii|^~
Basin
Major Coalbed Methane Fields
Coalbed methane fields
Coal basins and regions
Source data credits: U.S. Energy Information Administration (EIA)
Basemap Credits: U.S. Census Bureau, Esri, DeLorme, GEBCO, NOAA
NGDC, and other contributors
Figure 3-4. Coalbed methane fields and coal basins in the contiguous United States.
The fields represent gas-bearing coal deposits that are or could be developed. Adapted from EIA (2011b).
3-10
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
How a hydraulic fracturing operation is conducted depends on the characteristics of the oil- or gas-
bearing formation (such as the geology, depth, and other factors). Hydraulic fracturing operations
in shales, such as the Marcellus and Haynesville, require that relatively large volumes of water and
proppantto be pumped at high pressures through deep wells with long horizontal sections in the
production zone. In some tight formations, such as in the Permian Basin, hydraulic fracturing can be
conducted with smaller water volumes and using less pressure in shorter vertical or deviated wells
fGallegos and Varela. 20151. Hydraulic fracturing technologies can be applied to coalbed methane
production in various ways, for example, with much smaller water volumes and no proppant, or
with water-based gels or foams and proppant Coalbed methane production sometimes involves no
hydraulic fracturing, with only pumping of the naturally occurring formation water out of the
coalbeds to enable the release and production of the trapped methane.1
3.3 Hydraulic Fracturing and the Life of a Well
A variety of activities take place at a well site over the course of the operational life of a
hydraulically fractured oil and gas production well. Not all of these activities are within the scope of
this assessment (that includes water acquisition, chemical mixing, well injection, produced water
handling, and wastewater disposal and reuse). However, in this chapter we include some
information on a wider range of activities related to the well site to provide context for the reader.
The overview of well operations presented in this section is broad, illustrates common activities,
and describes some specific operational details. The details of well preparation, hydraulic
fracturing and production operations, and closure can vary between companies, reservoirs, and
states, and even from well to well. The activities involved in well development and operations may
be conducted by the well owner and/or operator, their representatives, and/or service companies
working for the well owner.
Figure 3-5 shows the general sequence and duration of activities at a hydraulic fracturing well site,
including the activities that comprise the five stages of the hydraulic fracturing water cycle (noted
above and defined in Chapter 1). The hydraulic fracturing event itself is the period of the most
operational activity during the life of a well and is short in duration compared to the other well site
activities. The hydraulic fracturing activity typically lasts from about a day to several weeks fU.S.
EPA. 2016c: Halliburton. 2013: NYSDEC. 2011). The subsequent phase of oil and gas production,
during which produced water also flows from the well, is the longest phase during the life of the
well and can last decades fKing and Durham. 20151.2
1 Some subsurface geologic formations, including coalbeds and oil and gas reservoirs, can contain naturally occurring
water that is commonly referred to as "formation water," "native water," or (if salty] "native brines."
2 In general, produced water is water that flows from the subsurface through oil and gas wells to the surface as a by-
product of oil and gas production. See Section 3.3.3 and Chapter 7 for more details.
3-11
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Site assessment
and preparation
• Surveys, core
sampling, and
test well logging
• Permitting
• Construction of
roads, well pads,
pipelines, and
other structures
Well drilling Hydraulic
and construction fracturing
• Water acquisition
• Drilling and
installation of
casing and cement
• Setup of water
tanks, pumps,
blenders, and other
equipment on site
• Transport of water,
proppant, and
chemicals to site
• Well preparation,
including perforation,
if necessary
• Chemical mixing
• Well injection of
fracturing fluids
in multiple stages
• Removal of
hydraulic fracturing
equipment
Produced water
handling and
wastewater
disposal/reuse
• Produced water
handling
• Wastewater disposal
and reuse, either
onsite or offsite
• Storage, treatment,
or reuse of water
onsite
• Hauling of produced
water via truck or
pipeline for
treatment, reuse, or
disposal offsite
Oil and gas
production
• Capture of oil
and/or gas
• Transport from
site via pipelines,
trucks, and/or
trains
• Re-fracturing
or recompletion
of the well, if
necessary
• Continued
produced water
recovery and
wastewater
management
Site and well
closure
• Wellhead and
surface structure
removal
• Well plugging
• Site reclamation
Figure 3-5. General timeline and summary of activities that take place during the preparation
and through the operations of an oil or gas well site at which hydraulic fracturing is used.
3.3.1 Site Preparation and Well Construction
Before hydraulic fracturing and production can occur, preliminary steps include assessing and
preparing the site, and drilling and constructing the production well.
3.3.1.1 Site Assessment and Preparation
Selecting a suitable well site requires an assessment of geologic (subsurface) and geographic
(surface) factors. Geophysical surveys of the subsurface can be conducted using data gathering
techniques from the land surface or subsurface, and rock samples may be gathered from outcrops
or from exploratory or test wells. Other information is obtained by well logging in which
geophysical instruments that collect data on subsurface conditions are lowered into or installed in a
well (Kundertand Mullen. 2009).1 Analyzing all of this information together enables operators to
develop an understanding of the potential reservoir characteristics (such as permeability and the
presence of natural fractures and water), the position of such formations in relation to other
1 Well logging is used to obtain information on mechanical integrity, well performance, and reservoir properties that can
affect oil and gas production. Well logging data from other wells in the nearby area also provides information on the
reservoir. More information on well logging is found in Chapter 6 and Appendix D.
3-12
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
formations, including water-bearing zones, and details about the quantity and quality of the oil and
gas resource.
Geographic factors involved in well site assessment include topography and land cover; proximity
to roads, pipelines, water sources, other oil and gas wells, and abandoned oil or gas wells; possible
well setback requirements; potential for site erosion; location relative to environmentally sensitive
areas; and location relative to populated areas (Drohan and Brittingham. 2012: Arthur et al..
2009a!1 Land ownership also plays an important role in well site selection. During site assessment
and before site development and well drilling, the well owner/operator obtains a mineral rights
lease, negotiates with landowners, and applies for necessary permits from the appropriate federal,
state, and local authorities fHvne. 20121. This initial site assessment phase of the process may take
several months (King and Durham. 2015: King. 2012).
The site is typically surveyed to plan and finalize well site location and access. Sometimes an access
road may need to be built to accommodate trucks delivering equipment and supplies to be used at
the site fHvne. 20121. The operator levels and grades the well site to manage drainage, complete
access routes, and prepare the well pad. The well pad is a smaller area within the broader well site
where the production well will be drilled and the hydraulic fracturing activities will be
concentrated. Well pads can range in size from less than an acre to several acres depending on the
scope of the operations (King. 2012: NYSDEC. 20111. Multiple wells can be located on a single well
pad at a well site fKing. 2012: NYSDEC. 20111
To manage the various fluids that are used for or generated during operations, storage pits
(sometimes referred to as impoundments) are excavated, graded and constructed on the well site,
and/or steel tanks are installed. These are used to hold water and materials (such as drilling mud)
related to the well-drilling activities, water used in the hydraulic fracturing process, or the
produced water that is generated post-fracturing (Hvne. 2012). Pit construction is generally
governed by local regulations. In some areas, regulations may prohibit the use of pits or require pits
to be lined to prevent fluid seepage into the shallow subsurface. One alternative to constructing a
pit for drilling fluids is the use of a closed loop drilling system that stores, partly treats, and recycles
the drilling fluid fAstrella and Wiemers. 19961. Often piping is installed along the surface or in the
shallow subsurface of the well site to deliver water for hydraulic fracturing, remove produced
water, or transport the oil and gas once production begins (Arthur et al.. 2009a).
Water may be acquired from local surface water or groundwater resources, or reused from other
well sites. Water is required for the drilling phase as well as for hydraulic fracturing (Chapter 4).
Figure 3-6 depicts the pumping of water for well site operation from a local surface water source.
After site and well pad preparation, drill rigs and associated equipment (including the drill rig
platform, generators, well blowout preventer, fuel storage tanks, cement pumps, drill pipe, and
casing) are brought onto the site.
1 Regarding well setbacks, some states and sometimes local city or county governments can have requirements that define
how close an oil and gas well can be located to drinking water supplies or other water bodies.
3-13
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Figure 3-6. Surface water being pumped for oil and gas development.
Photo credit: Arkansas Water Science Center (USGS).
3.3.1.2 Well Drilling and Construction
Wells are generally drilled and constructed by repeating several basic steps. The operator begins by
using the drill rig (temporarily located on the well pad) to hoist a section of long drill pipe up and
attaching a drill bit to the bottom of the drill pipe. The drill rig is then used to rotate and advance
the drill pipe/drill bit combination (also known as the drill string) downward through the soil and
rock. As the drill string continues moves downward, new sections of pipe are added at the surface,
enabling the drilling to proceed deeper (Hvne. 2012). During drilling a drilling fluid is pumped
down through the center of the drill string to the drill bit to lubricate and cool it, and to help
remove the drill cuttings from the well fKing. 2012).1
Drilling is temporarily halted at certain pre-determined intervals, the drill string is removed from
the wellbore (also called the borehole), and long sections of another type of steel pipe called casing
are lowered into the wellbore and set in place.2 Cement is then pumped into the space between the
outside of the casing and the wellbore. This process is repeated, with the next interval of drilling
1 Drilling fluids, sometimes called drilling mud, consist primarily of water, foam, oil or air, with the most common drilling
mud consisting mainly of water and clay (Williamson. 2013). Drill cuttings are the small pieces of broken and ground-up
rock generated during the drilling process.
2 The wellbore is the drilled hole and can refer to both the open hole or an uncased portion of the well.
3-14
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
using a smaller diameter drill bit that fits inside the existing casing. The result can be multiple
layers of casing and cement with surface casing and cement typically set below the groundwater
resource to be protected. Figure 3-7 illustrates different types of casing as defined by their locations
within the well, shows multiple casing and cement layers, and shows examples of two wells with
differences in the extent of cement1
Ground Surface
Protected
Ground Water
Casing »
Cement ~
^Conductor *
•Surface
Production
Wellbore
Targeted Rock
Formation
Conductor, surface, and production casings
Figure 3-7. Illustration of well construction showing different types of casing and cement.
The well on the left is cemented continuously from the surface to the production zone and the well
on the right has cement in sections, including sections cemented across protected groundwater.
The cement protects the casing from corrosion by formation water, helps physically support the
casing in the borehole, and stabilizes the borehole against collapse or deformation (Renpu. 2011).2
The casing and cement help to isolate geologic zones of high pressure, isolate water-bearing zones,
and maintain the integrity of the production well for transporting oil and gas to the surface. Casing
and cement provide important barriers that keep fluids within the well (oil, gas, hydraulic
fracturing fluids) isolated and separated from fluids outside the well (formation water) fHvne.
2012). Figure 3-8 shows sections of casing ready for installation.
1 In different portions ofthe well, multiple concentric sections of casing of different diameters can be used as shown by
the surface and production casings in Figure 3-7. The largest casing diameter can range between 30 in. (76 cm] to 42 in.
(107 cm] with casing diameters typically larger in the shallower portions of a well and smaller in the deeper portions
fHvne. 20121. See Appendix D for details on well construction and casing diameters.
2 Some naturally occurring formation water can be veiy saline (salty or briny], which can be corrosive to metal.
3-15
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Figure 3-8. Sections of well casing ready for installation at a well site in Colorado.
Photo credit: Gregory Oberley (U.S. EPA).
Some wells are cemented continuously from the surface down to the production zone. Other wells
are partially cemented with, for example, cement from the surface to some distance below the
deepest protected groundwater zone and perhaps cement across high pressure or water- or oil-
bearing zones. Sometimes there can be multiple casing and cement layers (Figure 3-7). There are
advantages, in some situations, to not fully cementing the casing as long as high pressure or water-
and oil-bearing zones are cemented. For example, some sections may not be cemented to allow
monitoring of the pressure in the space between the casing and the borehole or to prevent damage
to weak rock formations due to the weight of the cement1 (King and Durham. 2015: API. 2009).
Although wells are initially drilled vertically (more or less straight down), the sections of the wells
that are hydraulically fractured in the production zone of the reservoir can be vertical, deviated, or
horizontal (Figure 3-1). The operator determines the well orientation that will provide the best
access to the targeted zone(s) within a reservoir and that will align the production section of the
well with natural fractures and other geologic structures in a way that helps improve production.
Deviated wells may be "S" shaped or continuously slanted. So-called "horizontal wells" have one or
more extensions or branches oriented approximately 90 degrees from the vertical portion of the
well; these horizontal sections are often referred to as "laterals." The lengths of laterals can range
from 2,000 to 10,000 ft (600 to 3,000 m) or more (Hvne. 2012: Miskimins. 2008: Bosworth et al..
1998). Multiple laterals can extend in different directions from a single well (and multiple wells can
be located on a single well site). This allows access to more of the production zone with a higher
well density in the subsurface, which can be required for unconventional reservoirs, while having
fewer well sites on the land surface.
1 The use of lighter cement or special cementing techniques can also prevent damage of weaker rock formations. See
Chapter 6 and Appendix D for more details on well construction and cementing.
3-16
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Once well construction is completed, the operator can move the drilling rig and related drilling
equipment, install the wellhead (the top portion of the well], and prepare the well for hydraulic
fracturing and subsequent production of oil and gas. Chapter 6 and Appendix D contain more
details on well construction, casing, and cement.
Figure 3-9 (from northeastern Pennsylvania) and Figure 3-10 (from northwestern North Dakota)
show, in the context of the local landscape, well sites during well drilling and construction prior to
hydraulic fracturing activities.
Figure 3-9. Aerial photograph of two hydraulic fracturing well sites and a service road in
Springville Township, Pennsylvania,
Photo credit: lmage@J Henry Fair / Flights provided by LightHawk.
3-17
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Figure 3-10. Aerial photograph of hydraulic fracturing well sites near Williston, North Dakota.
Photo credit: lmage@J Henry Fair / Flights provided by LightHawk.
3.3.2 Hydraulic Fracturing
The hydraulic fracturing phase is an intense phase of work in the life of the well that involves
complex operational activities at the well site. This phase of work is short in duration, compared to
other work phases in the life of a well, and typically lasts less than two weeks per well. It consists of
multiple activities, is typically a process done in repetitive stages, and requires a variety of
equipment and materials. During this phase of work, the well is prepared for hydraulic fracturing,
specialized equipment is hauled to the well site, the hydraulic fracturing fluid components -the
water, proppant, and additives- are moved to the well site, and the hydraulic fracturing fluid is
mixed and injected under pressure through the well and into the targeted production zone in the
subsurface (Figure 3-11).
3-18
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
I
I
Figure 3-11. Well site with equipment (and pits in the background) in preparation for
hydraulic fracturing in Troy, Pennsylvania.
Image from NYSDEC (2015). Reprinted with permission.
3.3.2.1 Injection Process
The section of well located in the production zone can be prepared for the injection and fracturing
process in several different ways. One approach is used when the production casing and cement
extend all the way into the production zone; this requires the use of focused explosive charges to
perforate (blast holes in) the casing and cement in a segment of the well within the production
zone. In another approach, known as a formation packer completion, only the casing, equipped with
holes that can be opened and closed, is extended into the production zone. The resulting
perforations or holes allow the injected hydraulic fracturing fluids to flow out of the well to fracture
the reservoir rock and allow the oil and gas to flow into the well. Another technique is an open hole
completion in which the casing is set and cemented just to the edge of the production zone, so the
borehole extends open (with no casing or cement) into the production zone. In open hole
completions, oil and gas flow directly into the borehole and eventually into the cased section of the
well leading to the surface fliyne. 2012: Cramer. 2008: Economides and Martin. 20071.
3-19
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
After the subsurface portion of the well is prepared for injection, a wellhead assembly is
temporarily installed on the wellhead to which high pressure fluid lines are connected for injection
of the fluids into the well. Figure 3-12 shows three wellheads with injection piping attached in
preparation for hydraulic fracturing injection. Pressures required for fracturing can vary widely
depending on depth, formation pressure, and rock type and can range from 2,000 psi to 12,000 psi
fU.S. EPA. 2016c: Salehi and Ciezobka. 2013: Abou-Saved etal.. 2011: Thompson. 20101.
Figure 3-12. Three wellheads on a multi-well pad connected to the piping used for hydraulic
fracturing injection.
Photo credit: DOE/NETL
The portion of the well to be fractured can sometimes be done all at once or done in multiple
interval fU.S. EPA. 2016c: GWPC and ALL Consulting. 20091. When done in multiple intervals,
shorter lengths or segments of the well are closed-off (using equipment inserted down into the
well) and fractured independently in "stages" (Lee et al.. 20111. Fluids are first injected to clean the
well (removing any cement or debris). Then, for each stage fractured, a series of hydraulic
fracturing fluid mixtures is injected to initiate fractures and carry the proppant into the fractures
fHvne. 2012: GWPC and ALL Consulting. 20091. The fracturing process can require moving millions
of gallons of fluids around the well site through various hoses and lines, blending and mixing the
fluids with proppant, and injecting the mixture at high pressures down the well. For more details on
hydraulic fracturing chemical mixtures and stages, see Chapter 5.
The hydraulic fracturing produces propped-open fractures that extend into the production zone
and create more flow paths that contact a greater volume of the oil- and gas-bearing rock within the
production zone of the reservoir. This increase in flow paths and in the volume of the production
3-20
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
zone accessed by the production well is how hydraulic fracturing increases production. In this
regarding, hydraulic fracturing can be considered a production or well "stimulation" technique.
The process and the fracturing pressures during injection are closely monitored throughout the
fracturing event. Microseismic monitoring (a geophysical survey technique) can be used to estimate
the horizontal and vertical extent of the fractures created and, used with other monitoring and
operational data, provides important information for designing subsequent fracture jobs (Cipolla et
al.. 20111. Engineers can design fracture systems using modeling software to help optimize the
process. More details of injection, fracturing, and related monitoring are provided in Chapter 6 and
Appendix D.
3.3.2.2 Fracturing Fluids
To conduct the chemical mixing and preparation of the hydraulic fracturing fluids, water- and
chemical-filled tanks and other storage containers are transported and installed on site. The
components that make up the hydraulic fracturing fluid for injection are commonly mixed on a
truck-mounted blender on the well pad. Hoses and pipes are used to transfer the water, proppant,
and chemicals from storage units to the mixing equipment and to the well into which the mixed
hydraulic fracturing fluid will be injected. The injection process happens in stages with specific
chemicals added at different times during each stage. The composition of the hydraulic fracturing
fluid, therefore, can change over time during the process fKnappe and Fireline. 2012: Fink. 20031.
See Chapter 5 for more details on mixing and staged injection.
Hydraulic fracturing fluids (sometimes referred to as "fluid systems") are generally either water-
based or gel-based. Other fluid systems include foams or emulsions made with nitrogen, carbon
dioxide, or hydrocarbons; acid-based fluids; and others fMontgomerv. 2013: Saba etal.. 2012:
Gupta and Hlidek. 2009: Gupta and Valko. 2007: Halliburton. 1988). Water-based systems are used
more often with the most common type being "slickwater" formulations, which include polymers as
friction reducers and are typically used in very low permeability reservoirs such as shales (Barati
and Liang. 20141. Because slickwater fluids are thinner (have lower viscosity) they do not as easily
carry sand proppant into fractures, so larger volumes of water and greater pumping pressures are
required to effectively transport proppants into fractures. In contrast, gelled fluids (used in "gel
fracs") are more viscous, and more proppant can be transported with less water as compared to
slickwater fractures (Brannon et al.. 2009). Gel fracs are generally used in reservoirs with higher
permeability fBarati and Liang. 20141.
The composition of a typical water-based hydraulic fracturing fluid by volume is 90% to 97%
water, 2% to 10% proppant, and 2% or less additives (U.S. EPA. 2015a: OSHA. 2014a. b; Carter et
al.. 2013: Knappe and Fireline. 2012: Spellman. 2012: Siolander etal.. 2011: SWN. 2011). In a
detailed study, the EPA analysis of FracFocus 1.0 data for nearly 39,000 wells nationally in 2011
and 2012 indicates that the fracturing fluid injected into a well consists of nearly 90% water, 10%
proppant, and less than 1% additives (on a mass basis) (U.S. EPA. 2015al. The proportions of water,
proppant, and additives in the fracturing fluid, and the specific additives used, can vary depending
on a number of factors, including the rock type and the chemistry of the reservoir, whether oil or
gas is being produced, operator preference, and to some degree on local or regional availability of
3-21
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
chemicals fArthur etal.. 2014: Spellrnan. 2012: GWPC and ALL Consulting. 2009: Gupta and Valko.
20071. Hydraulic fracturing fluid composition and chemical use changes as processes are tested and
refined by companies and operators. These changes are driven by economics, scientific and
technological developments, and concerns about environmental and health impacts. Further detail
on hydraulic fracturing fluid systems is presented in Chapter 5.
Sources of water for hydraulic fracturing fluid include groundwater, surface water, and reused
wastewater fURS Corporation. 2011: Blauch. 2010: Kargbo et al.. 20101. The water may be brought
to the production well from an offsite regional source via trucks or piping, or it may be more locally
sourced (for example, pumped from a nearby river or a groundwater well). Selection of water
source depends upon availability, cost, water quality needs, and the logistics of delivering it to the
site. Figure 3-13 shows a row of water tankers storing water on a well site. Chapter 4 provides
additional details on water acquisition and the amounts of water used for hydraulic fracturing.
Figure 3-13. Water tanks (blue, foreground) lined up for hydraulic fracturing at a well site in
central Arkansas.
Photo credit: Martha Roberts (U.S. EPA).
Proppants are most commonly silicate minerals, primarily quartz sand fGWPC and ALL Consulting.
20091. Sand proppants can be coated with resins that make them more durable. Ceramic materials
are also sometimes used as proppants due to their high strength and resistance to crushing and
deformation fBeckwith. 20111.
Additives generally constitute less than 2.0% of hydraulic fracturing fluids f Carter etal.. 2013:
Knappe and Fireline. 2012: GWPC and ALL Consulting. 2009). The EPA analyzed additive data in the
EPA FracFocus 1.0 project database and estimated that chemicals used as additives were about
0.43% (the median value by mass) of the total amount of fluid injected for hydraulic fracturing (U.S.
EPA. 2015al. Given the total volume of hydraulic fracturing fluid, these small percentages of
chemicals in the fluid mean that a typical hydraulic fracturing job can handle, mix, and inject tens of
thousands of gallons of chemicals. Chapter 5 includes details on the number, types, and estimated
quantities of chemicals ty pically used in hydraulic fracturing.
3-22
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
3.3.3 Fluid Recovery, Handling, and Disposal or Reuse
At the end of the hydraulic fracturing process, the pressurized injection is stopped and the direction
of fluid flow reverses. Initially, the fluid flowing back into the well and to the surface is mostly the
injected fracturing fluid (sometimes referred to as flowback), The composition of the fluid changes
over time, though, and after the first few weeks or months the proportion of hydraulic fracturing
fluid flowing back into the well decreases and the proportion of formation water flowing into the
well and to the surface increases (NYSDEC. 2011). In this assessment, the water that flows from the
subsurface through oil and gas wells to the surface as a by-product of oil and gas production is
referred to as produced water. The amount of produced oil or gas flowing into the well gradually
increases until it is the primary constituent of the fluid emerging from the well at the surface.
Produced water continues to flow from the production well along with the oil or gas throughout the
operating life of the production well (Barbotetal.. 2013). See Chapter 7 for details, descriptions,
and discussions of the chemical composition and quantities of produced water recovered.
Produced water is sometimes referred to as hydraulic fracturing wastewater. Along with other
liquid waste collected from the well pad (such as rainwater runoff), it is typically stored
temporarily on-site in pits (Figure 3-14) or tanks. This wastewater can be moved offsite via truck or
pipelines for treatment and reuse or for disposal. Most hydraulic fracturing wastewater in the
United States is disposed of by injection into deep, porous geologic rock formations, often located
away from the production well site. This disposal-by-injection occurs not through oil and gas
production wells, but through wastewater injection wells regulated by EPA Underground Injection
Control (UIC) programs under the Safe Drinking Water Act.1 See Chapter 8 for a brief discussion of
wastewater injection.
Figure 3-14. A pit on the site of a hydraulic fracturing operation in central Arkansas.
Photo credit: Caroline E. Ridley (U.S. EPA).
1 States may be given federal EPA approval to run a UIC program under the Safe Drinking Water Act. Most oil- and gas-
related UIC programs are implemented by the states although some are implemented by the EPA.
3-23
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Other wastewater handling options include discharge to surface water bodies either with or
without treatment, evaporation or percolation pits, or reuse for subsequent fracturing operations
either with or without treatment fU.S. EPA. 2012h: U.S. GAP. 20121. Decisions regarding
wastewater handling are driven by factors such as cost (including costs of temporary storage and
transportation), availability of facilities for treatment, reuse, or disposal, and regulations
(Rassenfoss. 20111. Chapter 8 contains details of the treatment, reuse, and disposal of wastewater.
3.3.4 Oil and Gas Production
After the hydraulic fracturing activity is completed, the fracturing-related equipment is removed
and operators drain, fill in with soil, and regrade pits that are no longer needed unless multiple
wells are drilled and fractured on the same pad. The well pad size is reduced as the operation
moves toward the production phase fNYSDEC. 20111. Prior to and during production, the operator
runs production tests to determine the maximum flow rate that the well can sustain and to
determine optimum equipment settings (Hvne. 2012: Schlumberger. 20061. During production,
monitoring of mechanical integrity and performance (with pressure tests, corrosion monitoring,
etc.) can be conducted to ensure that the well is performing as intended. Such well tests and
monitoring may be required by state regulations.
Produced gas typically flows from the well through a pipe to a "separator" that separates the gas
from water and any liquid oil and gas fNYSDEC. 20111. The finished gas is typically piped to a
compressor station where it is pressurized and then piped to a main pipeline for sale fHvne. 20121.
Production at oil wells proceeds similarly, although oil/water or oil/water/gas separation typically
occurs on the well pad, no compressor is needed, and the oil can be hauled by truck or train, or
piped from the well pad to offsite storage and sale facilities.1
During the life of the well it may be necessary to repair components of the well and replace old
equipment Sometimes the well is re-fractured to boost production.2 Routine maintenance
activities, often referred to as "workovers," may be done with the well still in production
fVesterkiaer. 20021 or sometimes require stopping production and removing the wellhead to clean
out debris or repair components of the well fHvne. 20121. More extensive re-workings of a well,
sometimes referred to as "re-completions," can include making additional perforations in the well
in new sections to produce oil and/or gas from another production zone, lengthening the borehole,
or drilling new horizontal extensions (laterals) from an existing borehole.
3.3.4.1 Production Rates and Duration
The production life of a well depends on a number of factors, such as the amount of oil or gas in the
reservoir, the reservoir pressure, the rate of production, and the economics of well operations,
including the price of oil and gas. In hydraulically fractured wells in unconventional reservoirs,
1 In some oil production operations, the oil reservoir being tapped may include some natural gas that is extracted along
with oil through the production wells. In cases where no facilities or pipelines are in place to handle the natural gas or
move it to a market, the gas can be "flared" (ignited and burned at the well site] or vented into the atmosphere.
2 Sometimes boosting or reinvigorating production in a well is referred to as "well stimulation." In some cases, well
stimulation can refer to either the initial well hydraulic fracturing event or the re-fracturing of a well.
3-24
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
initial high production is typically followed by a rapid drop and then a slower decline in production
("Patzek etal., 2013"). The production phase may be 40 to 60 years in tight gas reservoirs ("Ross and
King, 2007") or range from 5 to 70 years in a gas- or liquids-rich shale ("King and Durham, 2015").
However, because the current hydraulic fracturing-led production surge is less than a decade old
with limited well production history, there is an incomplete picture of production declines and it is
not known how much and for how long these wells will ultimately produce ("Patzek etal., 2013").
3.3.5 Site and Well Closure
Once a well reaches the end of its useful life, it is removed from production and disconnected from
any pipelines that transferred produced oil or gas offsite. The well is then sealed to prevent any
movement of fluids inside or along the borehole. This is done by removing the wellhead, cutting the
casing off below ground surface, and then sealing portions of the well with one or more cement or
mechanical plugs placed permanently in sections of the well. Spaces between plugs may be filled
with a thick clay (bentonite) or drilling mud (NPC, 2011b"). State regulations identify plugging
locations within the borehole and the materials for plugging (Calvert and Smith, 1994], After
plugging and cementing, a steel plate is welded on top of the well casing to provide a complete seal
("API, 2010"). Permanently closing a well like this is called "plugging" a well. Some states require
formal notification of the location of these plugged wells. Proper plugging prevents fluids at the
surface from seeping down the borehole and migration of fluids through the borehole fNPC,
2011b"). See Chapter 6 for more details regarding fluid movement in wells and through the
borehole.
To complete site closure, any remaining production-related equipment is removed and the site land
cover and topography are restored to pre-well pad conditions to the extent possible. Some surface
structures from the former operations may be left in place for subsequent reuse.
3.4 How Widespread is Hydraulic Fracturing?
There is no national database or complete national registry of wells that have been hydraulically
fractured. However, hydraulic fracturing activity for oil and gas production in the United States is
substantial based on various reports and data sources. According to the Interstate Oil and Gas
Compact Commission (IOGCC), close to 1 million wells had been hydraulically fractured in the
United States by the early 2000s ("IOGCC, 2002"). A recent U.S. Geological Survey report estimated
approximately 1 million wells with 1.8 million hydraulic fracturing treatment records from 1947 to
2010 (more than one fracturing event, or treatment, can be conducted on a single well) ("Gallegos
and Varela, 2015"). Roughly a third of these 1 million wells were drilled and hydraulically fractured
between 2000 and 2013/2014 based on estimates from FracFocus (2016): Baker Hughes (2015):
Gallegos etal. (2015); Drillinglnfo (2014a); IHS Inc. (2014). This timeframe marks the beginning of
modern hydraulic fracturing (refer to Text Box 3-1). Figure 3-15 shows the location of the
approximately 275,000 oil and gas wells that were drilled and hydraulically fractured between
2000 and 2013 across the United States based on well and locational data from Drillinglnfo
(Drillinglnfo, 2014a).
3-25
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Hydraulically Fractured Well Locations
~ 2000-2013 Oil and Gas Wells
Source data credits: Drillinglnfo, Inc.
Basemap Credits: U.S. Census Bureau, Esri, DeLorme, 6EBC0, NOAA
NGDC, and other contributors
Figure 3-15. Locations of the approximately 275,000 wells drilled and hydraulically fractured
between 2000 and 2013.
Based on data from the Drillinglnfo Database.
The following two satellite photographs show hydraulic fracturing well sites in a regional context.
These Landsat images show the locations, number, and density of hydraulic fracturing well sites
across landscapes in northwest Louisiana (Figure 3-16) and western Wyoming (Figure 3-17). The
orange circles around some of the well sites identify them as operations for which well information
was reported to the FracFocus 1.0 registry and included in the EPA FracFocus 1.0 project database
(U.S. EPA. 2015c). Note that some of the well sites in the Landsat images, taken in 2014, are for
wells that were constructed after the development of the EPA FracFocus 1.0 project database.
3-26
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
vjKrIBun
Hydraulic fracturing sites
Well locations reported
to FracFocus
m Satellite imagery of well
g|i pads
0 2 4 Miles
Figure 3-16. Landsat photo showing hydraulic fracturing well sites near Frierson, Louisiana.
Imagery from USGS Earth Resources Observation and Science, Landsat 8 Operational Land Imager (scene
LC80250382014232LGN00) captured 8/20/2014, accessed 5/1/2015 from USGS's EarthExplorer
(http://earthexplorer.usgs.gov/). Inset imagery from United States Department of Agriculture National Agriculture
Imagery Program (entity M 3209351_NE 15_1_20130703_20131107) captured 7/3/2013, accessed 5/1/2015 from
USGS's EarthExplorer (http://earthexplorer.usgs.gov/). FracFocus well locations are from the EPA FracFocus 1.0
project database (U.S. EPA, 2015c).
3-27
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
Hydraulic fracturing sites
¦ Well sites in satellite
J image
Wells reported to
O FracFocus
0 2 4 Miles
Wyoming
Figure 3-17. Landsat photo showing hydraulic fracturing well sites near Pinedale, Wyoming
Imagery from USGS Earth Resources Observation and Science, Landsat 8 Operational Land Imager (scene
LC803703Q2014188LGN00) captured 7/7/2014, accessed 5/1/2015 from USGS's EarthExplorer
(http://earthexplorer.usgs.gov/). Inset imagery from United States Department of Agriculture National Agriculture
Imagery Program (entity M 4210927_NW 12_1_20120623_20121004) captured 6/23/2012, accessed 5/1/2015
from USGS's EarthExplorer (http://earthexplorer.usgs.gov/). FracFocus well locations are from the EPA FracFocus
1.0 project database (U.S. EPA, 2015c).
3-28
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
3.4.1 Number of Wells Fractured per Year
Approximately 25,000 to 30,000 new oil and gas wells were hydraulically fractured each year in the
United States between 2011 and 2014 based on data from several commercial data sets and
publicly available data from organizations that track drilling and hydraulic fracturing activities
(Table 3-1). These estimates do not include fracturing activity in older, existing wells (wells more
than one-year old that may or may not have been hydraulically fractured in the past). Likely
following the decline in oil prices (starting in about 2014) and gas prices (in about 2008), the
estimated number of new hydraulically fractured wells declined to about 20,000 in 2015 according
to well information submitted to FracFocus (FracFocus. 2016). Future drilling activity and the
annual number of new wells will be influenced by future oil and gas prices.
Table 3-1. Estimated number of new wells hydraulically fractured nationally by year from
various sources.
Data from FracFocus (2016); Baker Hughes (2015); Drillinglnfo (2014a); IHS Inc. (2014) as provided in Gallegos et al.
(2015).
Data Source
2011
2012
2013
2014
IHS
29,650
31,073
29,114
11,980 a
Drillinglnfo
23,144
22,865
15,903 b
NA
Baker Hughes
NA
24,948
25,368
26,548
FracFocusc
14,025
22,471
26,400
28,285
a The IHS well count for 2014 is incomplete as it represents data only for 8 months (January through August).
b The Drillinglnfo well count for 2013 is incomplete because some months are missing from some state data sets.
c The FracFocus 2011 and 2012 counts are underestimates because reporting well information to FracFocus was voluntary when
it began in 2011. The number of states requiring reporting to FracFocus has increased over time. See FracFocus discussion
below. The FracFocus well counts for 2011 and 2012 are from the EPA FracFocus 1.0 project database (U.S. EPA. 2015c)
developed from the FracFocus national registry, and the FracFocus counts for 2013 and 2014 are from (FracFocus. 2016).
The Information Handling Services (IHS) annual well count estimates presented in Table 3-1 are
from IHS data made available in a U.S. Geological Survey publication that evaluated well data from
2000 to 2014 (Gallegos etal.. 2015). The IHS data are compiled from a variety of public and private
sources and are commercially available from IHS Energy. A well is identified as a hydraulic
fracturing well apparently based on well operational information. Gallegos etal. f20151 estimated,
based on the IHS data, that approximately 371,000 wells were hydraulically fractured between
January 2000 and August 2014.
Drillinglnfo, another commercial database, is developed using data obtained from individual state
oil and gas agencies fDrillinglnfo. 2014al. Because Drillinglnfo data does not identify whether a
well has been hydraulically fractured, EPA relied on information about well orientation and the oil-
or gas-producing rock formation type to infer which wells were likely hydraulically fractured. This
is a similar approach to that used by the EPA for estimating oil and gas well counts for its
3-29
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
greenhouse gas inventory work (U.S. EPA. 2013c).1 Using this approach, we estimate from the
Drillinglnfo data the annual numbers presented in Table 3-1 above and also estimate that a total of
approximately 275,000 oil and gas wells were drilled and hydraulically fractured between 2000
and 2013.2
Well counts tracked by Baker Hughes provide another estimate of new wells fractured annually.
This field service company compiles new-well information based on its extensive field work in oil
and gas producing areas and through state agencies. Baker Hughes started compiling this publicly
available well count data in 2012, but stopped in 2014. The well count data are categorized into 14
basins containing reservoirs that are mostly unconventional (and, therefore, likely hydraulically
fractured wells) and one "other" category fBaker Hughes. 20151. The well count estimates in the
table above are for the 14 basins and, therefore, are considered estimates of new wells
hydraulically fractured in each year.
FracFocus is a national registry for operators of hydraulically fractured oil and gas wells to report
information about well location and depth, date of operations, and water and chemical use. The
registry, publicly accessible online fwww.fracfocus.org], was developed by the Groundwater
Protection Council and the Interstate Oil and Gas Compact Commission. Submission of information
to FracFocus was initially voluntary (starting in January 2011), but many states now require
reporting of hydraulic fracturing well activities to FracFocus. As of May 2015, 23 states required
reporting to FracFocus fKonschnik and Davalu. 20161. The annual well counts in the table above
are from the EPA FracFocus 1.0 project database for 2011 and 2012 fU.S. EPA. 2015cl and from the
FracFocus 2016 Quarterly Report for 2013 and 2014 (FracFocus. 2016). The well counts in the
earliest years are underestimates because not all states required oil and gas well operators to
submit hydraulic fracturing data to FracFocus.3 The FracFocus registry has undergone several
updates since its launch in 2011. For more details on FracFocus, see FracFocus f20161. Konschnik
and Davalu f20161. U.S. EPA f2015a1. U.S. EPA f2015c1. and DOE f2014a1.-*
In addition to these new well counts, some portion of existing wells are also re-fractured. Several
studies indicate that re-fracturing occurs in less than 2% of wells. Shires and Lev-On f20121
1 Using the Drillinglnfo data, EPA assumed that all horizontal wells were hydraulically fractured in the year they started
producing and assumed that all wells within a shale, coalbed, or low-permeability formation, regardless of well
orientation, were hydraulically fractured in the year they started producing. More details are provided in fU.S. EPA.
2013c]. Not all coalbed methane wells are hydraulically fractured, but coalbed methane wells represent gas production
that sometimes uses hydraulic fracturing. Given the small percent of coalbed methane wells relative to all hydraulically
fractured wells and the lack of data that distinguishes whether or not coalbed wells are hydraulically fractured, EPA
included coalbed production wells into all counts of wells that are hydraulically fractured.
2 The different well count totals from IHS and Drillinglnfo are likely due to different sources of data, different approaches
for defining hydraulically fractured wells in those sources, and somewhat different timeframes. The higher IHS count
likely includes hydraulically fractured vertical and deviated wells in conventional reservoirs (the Drillinglnfo estimate
does not] and covers a time period that is a year or more longer.
3 We compared state records of hydraulic fracturing wells in North Dakota, Pennsylvania, and West Virginia in 2011 and
2012 to those reported to FracFocus duringthose same years and found the FracFocus wells counts underestimated the
number of fracturing jobs in those states by approximately 30% on average. See Chapter 4, Text Box 4-1.
4 Analyses of the FracFocus data based on the EPA FracFocus 1.0 project database fU.S. EPA. 2015c] are presented in
Chapter 4 regarding water volumes and in Chapter 5 regarding chemical use.
3-30
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
suggested that the rate of re-fracturing in natural gas wells is about 1.6% whereas analysis for the
EPA's 2012 Oil and Gas Sector New Source Performance Standards indicated a re-fracture rate of
1% for gas wells fU.S. EPA. 2012f). The percentage of hydraulically fractured producing gas wells
that were re-fractured in a given year ranged from 0.3% to 1% across the 1990-2013 period
according to the EPA's Inventory of U.S. Greenhouse Gas Emissions and Sinks (U.S. EPA. 2015h).
The above rates are calculated by comparing the number of re-fractured wells in a single year to all
hydraulically fractured wells cumulatively over a multi-year time period. However, when
calculating the rates of wells that conduct re-fracturing in a given year compared to the total
number of wells in that same year, the re-fracturing rate is higher. Data provided to the EPA's
Greenhouse Gas Reporting Program (GHGRP) for 2011 to 2013 suggest that 9-14% ofthe gas wells
hydraulically fractured in each year were pre-existing wells undergoing re-fracturing fU.S. EPA.
2014b).1 Another rate presenting a somewhat different measure (estimated by an EPA review of
well records from 2009 to 2010) found that 16% of the surveyed wells had been re-fractured at
least once fU.S. EPA. 2016cl.2
In summary, a complete count of the number of hydraulically fractured wells in the United States is
hampered by a lack of a definitive and readily accessible source of information, and the fact that
existing well and drilling databases and registries track information differently and therefore are
not entirely comparable. There is also uncertainty about whether existing information sources are
representative of the nation (or parts of the nation), whether they include data for all production
well types, and to what degree they include activities in both conventional and unconventional
reservoirs. Taking these limitations into account, however, it is reasonable to conclude that
between approximately 25,000 and 30,000 new wells (and, likely, additional pre-existing wells)
were hydraulically fractured each year in the United States from about 2011 to 2014, and
approximately 20,000 wells were hydraulically fractured in 2015.
3.4.2 Hydraulic Fracturing Rates
Estimates of hydraulic fracturing rates, or the proportion of all oil and gas production wells that are
hydraulically fractured, also indicate widespread use of the practice. Data from IHS Inc. f20141
indicate that approximately 62% of all new oil and gas wells in 2013 were hydraulically fractured.
Data from Drillinglnfo (2014a). indicate a similar rate of 64% of all new production wells in 2012.
Estimates of hydraulic fracturing rates reported by states in response to an IOGCC survey tended to
be considerably higher. Of eleven oil and gas producing states that responded to the survey, ten
(Arkansas, Colorado, New Mexico, North Dakota, Ohio, Oklahoma, Pennsylvania, Texas, Utah, and
West Virginia) estimated that 78% to 99% of new wells in their states were hydraulically fractured
in 2012. Louisiana was the one exception, reporting a fracturing rate of 3.9% (IOGCC. 2015).
Hydraulic fracturing may be more prevalent in gas wells than in oil wells. A2010to2011 survey of
20 natural gas production companies reported that 94% of the gas wells that they operated were
1 The GHGRP reporting category that covers re-fracturing is "workovers with hydraulic fracturing." This re-fracturing data
is for gas wells only (and does not include oil wells].
2 This EPA report is based on a statistical survey so there is some uncertainty and a margin of error regarding the 16% re-
fracturing rate. This rate includes both oil and gas wells. For more details, see Chapter 6 and U.S. EPA ("20161
3-31
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
fractured (Shires and Lev-On. 20121. a rate that is higher than many of the reported statistics for oil
and gas together (presented in the previous paragraph). Recent EIA data on the portion of oil and
gas production attributable to hydraulically fractured wells also suggest possibly higher rates of
hydraulic fracturing for gas. In 2015, production from hydraulically fractured wells accounted for
an estimated 67% of natural gas production (EIA. 2016d] and 51% of oil production (EIA. 2016c],
3.5 Trends and Outlook for the Future
Future oil and gas drilling and production activities, including hydraulic fracturing will be
primarily affected by the cost of well operation (partly driven by technology) and the price of oil
and gas. Scenarios of increasing, stable, and decreasing hydraulic fracturing activity all appear to be
possible (Weijermars. 2014). The section below provides some discussion on trends and future
prospects for production quantities and locations.
Fossil fuels-oil, gas, and coal-have been dominant energy sources in the United States over the last
half century (Figure 3-18). The relative importance of oil, gas, and coal has changed several times,
with a significant recent shift starting in the mid-2000s as hydraulic fracturing transformed oil and
gas production. Coal, the leading fossil fuel from the mid-1980s to the mid-2000s, has experienced a
large decrease in production, dropping from approximately 33% of U.S. energy production in 2007
to approximately 20% (about 18 quadrillion Btus) by the end of 2015 (EIA. 2016a).1 In contrast,
natural gas production has risen to unprecedented levels, and oil production has resurged to levels
not seen since the 1980s. Oil accounted for 15% of U.S. energy production in 2007 and increased to
approximately 23% (about 20 quadrillion Btus) by the end of 2015, and natural gas as a portion of
domestic energy production went from 31% to 37% (about 33 quadrillion Btus) (EIA. 2016a).
35
Total Gas
30
25
Coal
™ 20
o
I—
~o
(0
D
a
Crude Oil
Nuclear
Renewable
1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Year
Figure 3-18. Primary U.S. energy production by source, 1950 to 2015.
Source: EIA (2016a).
1A Btu, or British thermal unit, is a measure of the heat (or energy] content of fuels. At the scale of national U.S.
production, the graph in Figure 3-18 presents Btus in quadrillions, or a thousand million million (which is 1015, or a 1
with 15 zeros].
3-32
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
The surge in both oil and gas production started in the mid- to late-2000s and was driven by market
forces (supply and demand) and the related developments in and expanded use of hydraulic
fracturing. Figure 3-19 focuses on the years 2000 to 2015 and presents data showing the steady
increase in the portion of oil and gas production coming from hydraulically fractured wells. Oil and
gas production associated with hydraulic fracturing was insignificant in 2000, but by 2015 it
accounted for an estimated 51% of US oil production and 67% of US gas production (Figure 3-19).
10
2000
2003
2006
2009
2012
2015
I From hydraulically
' fractured wells
I From all other wells
o ^
.2 ~o
u aj
=5 Q.
"O
O CD
qJ!
l/> u
£.15
3 o
—
z3
80
60
40
20
2000
2003
2006
2009
2012
2015
I From hydraulically
' fractured wells
From all other wells
Figure 3-19. U.S. production of oil (left) and gas (right) from hydraulically fractured wells from
2000 to 2015.
Source: EIA (2016c) (oil) and EIA (2016d) (gas), based on IHS Global Insight and Drillinglnfo, Inc.
Hydraulic fracturing activities are concentrated geographically in the United States. In 2011 and
2012 about half of hydraulic fracturing wells were located in Texas with another quarter located in
the four states of Colorado, Pennsylvania, North Dakota, and Oklahoma fU.S. EPA. 2015cl. The maps
in Figure 3-20 show changes starting in 2000 in the national geography of oil and gas production
through the increased use of horizontal drilling, which frequently is associated with hydraulic
fracturing. Some traditional oil- and gas-producing parts of the country, such as Texas, have seen an
expansion of historical production activity as a result of modern hydraulic fracturing. Pennsylvania,
a leading oil- and gas-producing state a century ago, has seen a resurgence in oil and gas activity.
Other states that experienced a steep increase in production, such as North Dakota, Arkansas, and
Montana, have historically produced less oil and gas.
3.5.1 Natural Gas
Drilling of new natural gas wells declined markedly as natural gas prices fell in 2008 (Figure 3-21).
Nevertheless, over the coming decades natural gas production is expected to increase and that
increase will be associated significantly with wells that are hydraulically fractured. Projections by
EIA indicate that gas production from shale (and tight oil reservoirs) will almost double from 2015
to 2040 when it will constitute nearly 70% of total natural gas production (EIA. 2016d). Slight
increases are projected for production from tight gas reservoirs and coalbed methane production is
expected to continue fairly steady at relatively low levels (EIA. 2016a) (Figure 3-22). These
projections are dependent on estimated future prices of natural gas and other assumptions, and
3-33
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
New well locations in
select years (number
of wells)
• 2000 (341)
• 2005 (1,809)
• 2012 (14,560)
EIA shale basins
Figure 3-20. Location of horizontal wells that began producing oil or natural gas in 2000, 2005,
and 2012.
Based on data from Drillinglnfo (2014a).
3-34
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
S B
3S 7
3
u ,
° 1
Q 1
8 0
19SB 1992 1996 2000
— — — Natural Gas Wellhead Free
1,500
1,400
1,200
ijn
&D
1,000 £
15
BOD
600
400
200
0
2004 2D0S 2012 2016
N aura I Gas Rotary Rigs in Operation
Figure 3-21. Natural gas prices and drilling activity, United States, 1988 to 2015.
Sources: EIA (2016b) and EIA(2016f).
Ol
n
E
2015
Shale gas and tight oil
Tight gas
Other ower 48 onshore
1990
2000
2010
2020
2030
2040
Coalbed methane
Lower 48 offshore
Alaska
Figure 3-22. Historic and projected natural gas production by source (trillion cubic feet).
Source: EIA (2016a).
the details are subject to change. Nonetheless, a continuing increase in production is generally
suggested and the locations of historical production identified in Figure 3-23 indicate areas of
continued and future hydraulic fracturing activities for natural gas production.
The geographic concentration and trends in shale gas production by play (and identified by state)
are shown in Figure 3-23. The Barnett Shale, where the modern hydraulic fracturing boom started,
was the largest producer of shale gas until about 2010, producing 1.5 trillion cubic feet (tcf) (42.5
billion cubic meters [bem]) that year and remains a significant producer. In 2009, the Marcellus and
Haynesville plays produced 0.12 and 0.43 tcf (3.4 and 12.2 bem), respectively, but by 2011,
production from the Haynesville play surpassed that in the Barnett play, and by 2013 the Marcellus
Shale surpassed both the Barnett and the Haynesville to become the play with the most production.
3-35
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
By 2014, the Marcellus play was producing 4.8 tcf (135.9 bcm) of gas annually with the Eagle Ford,
Haynesville, and Barnett each producing roughly 1.5 tcf (42.5 bcm). Estimates of technically
recoverable resources, a general indicator of potential future production, are noted for the
Marcellus (about 150 tcf [4.25 trillion cubic meters]), Haynesville (73 tcf [2.07 trillion cubic
meters]), Eagle Ford in Texas (55 tcf [1.56 trillion cubic meters]), and Utica in Ohio, Pennsylvania,
and West Virginia (55 tcf [1.56 trillion cubic meters]). This suggests that these four plays will be
active contributors of shale gas production for the foreseeable future fEIA. 20131.1 Other gas plays
with significant resources include the Fayetteville in Arkansas, the Woodford in Oklahoma, and the
Mancos in Colorado.
16
14
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
indicate a continuing, but lower rate of growth (as compared to the period from about 2005 to
2015). The locations of historical production identified in Figure 3-26 indicate areas of continued
and future hydraulic fracturing activities for oil.
100
cu BO
iii
CL
o
o
&0
40
20
1,600
1,400
1,200
i/i
CiD
1,000 ,5
BOD
600
400
200
0
(LI
.Q
E
19BS
1992
1996
2000
2004
20DB
2012
2016
— — — CrudeOilFrst Purchase Prize Cr udeOil Flotary Rgsin Operaion
Figure 3-24. Crude oil prices and drilling activity, United States, 1988 to 2015.
Sources: EIA (2016b) and EIA(2016e).
2015
12
Tight oil
_Q
Lower 48 offshore
Other ower 48 onshore
Alaska
2000
2010
2020
2030
2040
Figure 3-25. Historic and projected oil production by source (million barrels per day).
Source: EIA (2016a).
The geographic concentration and trends in tight oil production by play (and identified by state)
are shown in Figure 3-26. Early tight oil production in the United States was centered in the
Permian Basin in west Texas and New Mexico, at plays that included the Spraberry and the
Bonespring. After 2009, the Bakken play (centered in western North Dakota) and the Eagle Ford
play (in Texas) emerged as the largest-producing tight oil plays. Oil production in the Bakken
3-37
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
increased from 99 million bbls (16,000 million L) in 2009 to 394 million bbls (63,600 million L) in
2014 fEIA. 2016gl. Production from Eagle Ford increased from 12 million bbls (2,000 million L) in
2009 to 498 million bbls (79,100 million L) in 2014 fEIA. 2016gl.
General estimates of potential resources suggest that future tight oil production in the United States
will continue to be led by Texas and North Dakota. Technical recoverable resources are estimated
at about 23 billion bbls (3,600 billion L) for the Bakken, about 21 billion bbls (3,300 billion L) for
the Permian Basin, and about 10 billion bbls (1,600 billion L) for Eagle Ford fEIA. 20151. Other
plays with significant estimated resources include the Niobrara-Codell in Colorado and Wyoming
and the Granite Wash in Oklahoma and Texas fEIA. 20121.
>-
T3 4
1_
OJ
Q.
CO
_Q
O
1 2
(j
T3
o
0
2000 2002 2004 2006 2008 2010 2012 2014
Year
Figure 3-26. Production from U.S. tight oil plays, 2000-2014.
Source: EIA (2016g). The graph shows tight oil plays in the same vertical order as the legend.
Name of play
l Eagle Ford
(TX)
Bakken
(MT, ND)
I Bonespring & others
(TX, NM)
Spra berry
(TX, NM)
Niobrara-Codell
(CO, WY)
Granite Wash
(OK, TX)
Other major
US tight oil plays
3.6 Conclusions
Hydraulic fracturing is the injection of hydraulic fracturing fluids through the production well and
into the subsurface oil or gas reservoir under pressures great enough to fracture the reservoir rock.
The fractures allow for increased flow of oil and/or gas from the reservoir into the well. Water used
in the hydraulic fracturing fluid is typically obtained from sources in the vicinity of the well. Water
that naturally occurs in the oil and gas reservoir rocks often flows into the production well and
through the well to the surface as a byproduct of the oil and gas production process. This byproduct
water, commonly referred to as produced water, requires handling and management.
Many well site and operational activities are conducted to prepare a site and well for hydraulic
fracturing and oil and/or gas production. The actual hydraulic fracturing event is of relatively short
duration, usually several weeks or less, but it is also a phase of work with numerous complex
3-38
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
operational activities to handle, mix, and inject the hydraulic fracturing fluid under pressure
through the production well. The injected hydraulic fracturing fluid typically contains mostly water,
includes a proppant (commonly sand) which ensures that the fractures remain propped open after
injection, and contains less than two percent additives (chemicals) that improve the fluid
properties for fracturing. These small percentages of additives, given the total volume of hydraulic
fracturing fluids, mean that a typical hydraulic fracturing job can use tens of thousands of gallons of
chemicals.
Since about 2005, the combination of hydraulic fracturing and directional drilling pioneered in the
Barnett Shale in Texas has become widespread in the oil and gas industry. Hydraulic fracturing
combined with directional drilling is now a standard industry practice. It has significantly
contributed to the surge in United States oil and gas production, and accounted for slightly more
than 50% of oil production and nearly 70% of gas production in 2015. Hydraulic fracturing has
resulted in expanded production from unconventional shale and so-called tight oil or gas reservoirs
that had previously been largely unused. This hydraulic fracturing-based production activity is
geographically concentrated. About three-quarters of new hydraulic fracturing wells in 2011 and
2012 were located in five states (Texas, Colorado, Pennsylvania, North Dakota, and Oklahoma) with
about half of all wells located in Texas.
There is no national database or complete national registry of wells that have been hydraulically
fractured in the United States. Based on the data available from various commercial and public
sources, we estimate that 25,000 to 30,000 new wells were drilled and hydraulically fractured in
the United States annually between 2011 and 2014. In addition to these new wells, some existing
wells not previously fractured were fractured, and some that had been fractured in the past were
re-fractured. New drilling of hydraulic fracturing wells, influenced by oil and gas prices, peaked in
the United States between 2005 and 2008 for gas and between 2011 and 2014 for oil. Following
price declines, the number of new hydraulically fractured wells in 2015 was about 20,000. Future
drilling and production will be influenced by future gas and oil prices. Despite recent declines in
prices and new drilling, oil and gas production in the United States continues at historically high
levels with projections of continued growth in the medium and long term led by hydraulic
fracturing-based production from unconventional reservoirs.
3-39
-------
Chapter 3 - Hydraulic Fracturing for Oil and Gas in the United States
This page is intentionally left blank.
3-40
-------
Chapter 4- Water Acquisition
Chapter 4. Water Acquisition
Abstract
In this chapter, the EPA examined the potential impacts of water withdrawals for hydraulic fracturing on
drinking water resource quantity and quality, and identified common factors affecting the frequency and
severity of impacts. Groundwater and surface water resources used for hydraulic fracturing also
currently serve or in the future may serve as drinking water sources, and water withdrawals for
hydraulic fracturing can affect the quantity or quality of the remaining drinking water resource.
Hydraulic fracturing used a median of 1.5 million gallons (5.7 million liters) of water per well from 2011
through early 2013. Surface water supplies almost all water used for hydraulic fracturing in the eastern
United States, whereas surface water or groundwater is used in the West. Reuse of hydraulic fracturing
wastewater as a percentage of injected volume is generally low, with a median of 5% according to an
EPA literature survey. Greater reuse occurs where disposal options are limited (e.g., the Marcellus Shale
in Pennsylvania) and not necessarily where water availability is lowest.
Hydraulic fracturing generally uses and consumes a relatively small percentage of water when
compared to total water use, water consumption, and water availability at the national, state, and county
scale. There are exceptions, however. For example, EPA's analysis shows that counties in southern and
western Texas have relatively high hydraulic fracturing water withdrawals and low water availability.
These findings indicate where impacts are more likely to occur or be severe, but local information (i.e.,
at the scale of the drinking water resource) is needed to determine whether potential impacts have been
realized. In some example cases (e.g., the Eagle Ford Shale in Texas, the Haynesville Shale in Louisiana),
local impacts to drinking water resource quantity have occurred in areas with increased hydraulic
fracturing activity. In these instances, hydraulic fracturing water withdrawals contributed to local
impacts along with other water users and the lack of precipitation.
Drought or seasonal times of low water availability can increase the frequency and severity of
impacts, while water management practices such as the establishment of low-flow criteria (termed
passby flows), shifting from fresh to brackish water sources, or increasing the reuse of wastewater
for hydraulic fracturing can help protect drinking water resources.
Uncertainty about the extent of impacts on drinking water resources stems from the lack of
available data at the local scale. The EPA could assess the potential for impacts at the county scale,
but often could not determine whether impacts occurred at drinking water withdrawal locations.
Overall, hydraulic fracturing uses and consumes a relatively small percentage of water at the county
scale, but not always, and impacts can still occur depending on the local balance between withdrawals
and availability. Regional or local-scale factors, such as drought or water management, can modify this
balance to increase or decrease the frequency or severity of impacts.
4-1
-------
Chapter 4- Water Acquisition
This page is intentionally left blank.
4-2
-------
Chapter 4- Water Acquisition
4. Water Acquisition
4.1 Introduction
Water is a crucial component of nearly all hydraulic fracturing operations, generally making up 90 -
97% of the total fluid volume injected into a well (Chapter 5).1 Ground- and surface water resources
that serve as sources of water for hydraulic fracturing also provide water for public water supplies
or private drinking water wells. For this reason, water withdrawals for hydraulic fracturing can
impact drinking water resources by changing the quantity or quality of the remaining resource.2 In
this chapter, we consider potential impacts of water acquisition for hydraulic fracturing on both
drinking water resource quantity and quality, and, where possible, identify factors that affect the
frequency or severity of impacts.3
We define impacts broadly in this assessment to include any change in the quantity or quality of
drinking water resources; see Chapter 1 for more information. This definition applies reasonably
well to the subsequent chapters (Chapters 5-8); however, by this definition, even the smallest water
withdrawals would be considered impacts. Recognizing this, we focus on a smaller subset of
potential impacts, specifically water withdrawals that have the potential to limit the availability of
drinking water or alter its quality. Whether water withdrawals have this potential depends
primarily on the balance between water use and availability at the local scale.4 5 By "local" in this
chapter, we refer to the scale at which impacts to drinking water resources are expected to occur.
This usually means a given surface water (e.g., river or stream) or groundwater resource (i.e.,
aquifer), or a given watershed where we have detailed information about local water dynamics
(e.g., case studies). We note the scale at which data are available and findings are reported.
1 This range is based on multiple sources that either present hydraulic fracturing fluid composition as a function of
volume (e.g., 95% of the total volume injected] or as a function of mass (e.g., 90% of the total mass injected]. See Chapter
5 for additional information.
2 Surface water withdrawals can affect water quality by altering in-stream flow and decreasing the dilution of pollutants
or changing water chemistry (Section 4.5.3]. Groundwater withdrawals may alter water quality by inducing vertical
mixing of fresh groundwater with contaminated water from the land surface or underlying formations, or by promoting
changes in reduction-oxidation conditions and mobilizing chemicals from geologic sources (Section 4.5.1].
3 Water acquired for use in other oil and gas development steps besides hydraulic fracturing is beyond the scope of this
chapter, including the water used in well drilling and well pad preparation and water removal for the production of
coalbed methane. Furthermore, water released to the atmosphere via gas combustion is also outside the scope of this
chapter.
4 Throughout this chapter we use the terms "water use" and "water withdrawals" interchangeably to refer to the water
that is acquired for hydraulic fracturing operations.
5 There is no standard definition for water availability, and it has not been assessed recently at the national scale ("U.S.
GAP. 2014], Instead, a number of water availability indicators have been suggested (e.g.. Roy etal.. 2005], Here,
availability is most often used to qualitatively refer to the amount of a location's water that could, currently or in the
future, serve as a source of drinking water (U.S. GAP. 2014], which is a function of water inputs to a hydrologic system
(e.g., rain, snowmelt, groundwater recharge] and water outputs from that system occurring either naturally or through
competing demands of users. Where specific numbers are presented, we note the specific water availability indicator
used.
4-3
-------
Chapter 4- Water Acquisition
A variety of factors can modify the balance between water use and availability. For example,
multiple hydraulically fractured wells require more water than a single well, making it critical to
assess the cumulative effects of multiple wells over a given area or time period. Furthermore, the
combined effects of multiple water users pumping from the same aquifer can compound stress on
already declining groundwater supplies. Alternatively, locally high rates of hydraulic fracturing
wastewater reuse may help offset the need for fresh water withdrawals. These and other factors
are discussed throughout the chapter.
This chapter proceeds roughly in two halves. In the first half, we address water use and
consumption by hydraulic fracturing.1 We provide an overview of the types of water used for
hydraulic fracturing (Section 4.2); the amount of water used per well (Section 4.3); and then
estimates of hydraulic fracturing water use and consumption at the national, state, and county
scale, both in absolute terms and relative to total water use and consumption (Section 4.4).
Although most available data and literature pertain to water use, we discuss water consumption
because hydraulic fracturing consumes a substantial proportion of the water it uses, so that a
proportion of the water is lost from the local hydrologic cycle. See Section 4.4 and Chapter 2 for
more information.
In the second half of the chapter, we assess the potential for impacts by location in certain states
(and major oil and gas regions within select states) where hydraulic fracturing currently occurs
(Section 4.5; Appendix B.2). For each state and region, we discuss the water used and consumed by
hydraulic fracturing, and then compare it to water availability. We do this using several lines of
evidence: (1) literature information (both quantitative and qualitative) on state and regional
hydraulic fracturing water use and availability; (2) comparisons between our county level
estimates of hydraulic fracturing water use and an index of water availability; and (3) local case
studies from the Eagle Ford play in Texas, the Upper Colorado River Basin in Colorado, and the
Susquehanna River Basin in Pennsylvania.2 The use of case studies provides insight into the local,
sub-county scale, where impacts are most likely to be observed in both space and time.
Overall, this chapter provides a national assessment of where potential impacts to drinking water
quantity and quality are most likely due to water acquisition for hydraulic fracturing. We utilize
case studies where data are available to understand local dynamics and whether impacts are
indeed realized. In the absence of case studies, we use county level data to assess where potential
impacts are most likely. Finally, we identify the common factors affecting the frequency and
severity of impacts. We provide a synthesis of our findings in Section 4.6.
1 We refer specifically to "water consumption" when data are available or it is explicitly noted in the scientific literature.
However, when specific information is not available, we use "water use" or "water withdrawals" as general terms to refer
to both water use and consumption by hydraulic fracturing.
2 The EPA's Review of Well Operator Files for Hydraulically Fractured Oil and Gas Production Wells (i.e., the "Well F ile
Review;" see Text Box 6-1] was originally planned to inform the water acquisition stage of the hydraulic fracturing water
cycle, but did not yield any useable information on this topic, and is therefore not cited as a source of information in this
chapter. Although information in some well files was of good quality, the well files generally contained insufficient or
inconsistent information on nearby surface water and groundwater quality, injected water volumes, and wastewater
volumes and disposition; therefore, these data were not summarized ("U.S. EPA. 2015nl
4-4
-------
Chapter 4- Water Acquisition
4.2 Types of Water Used
The three major sources of water for hydraulic fracturing are surface water (i.e., rivers, streams,
lakes, and reservoirs), groundwater, and reused hydraulic fracturing wastewater.12 3 These sources
often vary in their initial water quality and in how they are provisioned to hydraulic fracturing
operations. In this section, we provide an overview of the sources (Section 4.2.1), water quality
(Section 4.2.2), and provisioning of water (Section 4.2.3) required for hydraulic fracturing. Detailed
information on the types of water used by state and locality is presented in Section 4.5.
4.2.1 Source
Whether water used in hydraulic fracturing originates from surface water or groundwater
resources is largely determined by the type of locally available water sources. Water transportation
costs can be high, so the industry tends to acquire water from nearby sources if available (Nicotet
al.. 2014: Mitchell etal.. 2013a: Kargbo etal.. 20101. Surface water supplies most of the water for
hydraulic fracturing in the eastern United States, whereas surface water or groundwater is used in
the more semi-arid to arid western states. In western states that lack available surface water
resources, groundwater generally supplies the majority of water needed for fracturing (Table 4-1).
Brackish sources of groundwater can be important for reducing demand on fresh groundwater
resources in certain regions (e.g., the Permian Basin and Eagle Ford Shale in Texas; see Section
4.5.1).4 Local policies also may direct the industry to seek withdrawals from designated sources
(U.S. EPA. 2013a): for instance, some states have encouraged the industry to seek water
withdrawals from surface water rather than groundwater due to concerns over aquifer depletion.
See Section 4.5.4 and Section 4.5.5 for more information.
1 We use the term "hydraulic fracturing wastewater" to refer to produced water that is managed using practices that
include, but are not limited to, reuse in subsequent hydraulic fracturing operations, treatment and discharge, and
injection into disposal wells. The term is being used in this study as a general description of certain waters and is not
intended to constitute a term of art for legal or regulatory purposes (see Chapter 8 and Appendix J, the Glossary, for more
detail].
2 Throughout this chapter we sometimes refer to "reused hydraulic fracturing wastewater" as simply "reused
wastewater," because this is the dominant type of wastewater reused by the industry. When referring to other types of
reused wastewater not associated with hydraulic fracturing (e.g., acid mine drainage, wastewater treatment plant
effluent], we specify the source of the wastewater.
3 We use the term "reuse" regardless of the extent to which the wastewater is treated (Nicot etal.. 2014]: we do not
distinguish between reuse and recycling except when specifically reported in the literature.
4 We use the term "fresh water" to qualitatively refer to water with relatively low TDS that is most readily and currently
available for drinking water. We do not use the term to imply an exact TDS limit.
4-5
-------
Chapter 4- Water Acquisition
Table 4-1. Estimated proportions of hydraulic fracturing source water from surface water and
groundwater.
Location
Surface water
Groundwater
Year or time
period of
estimate
Louisiana—Haynesville Shale
87%a
13%a
2009 - 2012
Oklahoma—Statewide
63%b
37%b
2011
Pennsylvania—Marcellus Shale, Susquehanna River Basin
92%c
8%c
2008 - 2013
Texas—Barnett Shale
50%d
50%d
2011 - 2013
Texas—Eagle Ford Shale
10%e
90%e
2011
Texas—TX-LA-MS Salt Basinf
30%e
70%e
2011
Texas—Permian Basin
0%e
100%e
2011
Texas—Anadarko Basin
20%e
80%e
2011
West Virginia—Statewide, Marcellus Shale
91%g
9%g
2012
a Percentages calculated from fracturing supply water usage data only. Rig supply water and other sources were excluded as
they fall outside the scope of this assessment. Data from October 1, 2009, to February 23, 2012, for 1,959 Haynesville Shale
natural gas wells (LA Ground Water Resources Commission. 2012).
b Proportion of surface water and groundwater permitted in 2011 by Oklahoma's 90-day provisional temporary permits for oil
and gas mining. Temporary permits make up the majority of water use permits for Oklahoma oil and gas mining (Taylor. 2012).
c Calculated from SRBC (2016) data from July 2008 to December 2013.
d Nicot etal. (2014).
e Nicot etal. (2012).
f Nicot et al. (2012) refer to this region of Texas as the East Texas Basin.
5 Estimated proportions are for 2012, the most recent estimate for a full calendar year available from West Vireinia PEP (2014).
Data from the West Virginia DEP show the proportion of water purchased from commercial brokers as a separate category and
do not specify whether purchased water originated from surface water or groundwater. Therefore, we excluded purchased
water in calculating the relative proportions of surface water and groundwater shown in Table 4-1 (West Virginia DEP. 2014).
The reuse of wastewater from past hydraulic fracturing operations reduces the need for
withdrawals of fresh surface water or groundwater.1 In a survey of literature values from 10 states,
basins, or plays, we found a median of 5% of the water used in hydraulic fracturing comes from
reused hydraulic fracturing wastewater, with this percentage varying by location (Table 4-2).2 3
1 Hydraulic fracturing wastewater may be stored on-site in open pits, which may also collect rainwater and runoff water.
We do not distinguish between the different types of water that are collected on-site during oil and gas operations, and
assume that most of the water collected on-site at well pads is hydraulic fracturing wastewater.
2 Throughout this chapter, we preferentially report medians where possible because medians are less sensitive to outlier
values than averages. Where medians are not available, averages are reported.
3 This chapter examines reused wastewater as a percentage of injected volume because reused wastewater may offset
total fresh water acquired for hydraulic fracturing. In contrast. Chapter 8 of this assessment discusses the total percentage
of the generated wastewater that is reused rather than managed by different means (e.g., disposal in Class II wells]. This
distinction is sometimes overlooked, which can lead to a misrepresentation of the extent to which wastewater is reused to
offset total fresh water used for hydraulic fracturing.
4-6
-------
Chapter 4- Water Acquisition
Table 4-2. Percentage of injected water volume that comes from reused hydraulic fracturing
wastewater in various states, basins, and plays.
See Section 4.5 and Appendix B.2 for additional discussion of reuse practices by state and locality and variation
overtime where data are available.
State, basin, or play
Estimate of the percentage
of injected water volume
that comes from reused
hydraulic fracturing
wastewater3
Year or time
period of
estimate
(NA = not
available)
California—Monterey Shale
13%b
2014
Colorado—Wattenberg Field, Denver-Julesburg Basin
0%c
NA
Pennsylvania—Statewide
19%d
2014
Pennsylvania-Marcellus Shale, Susquehanna River Basin
16%e
2008-2013
Texas—Barnett Shale
5%f
2011
Texas—Eagle Ford Shale
0%f
2011
Texas—TX-LA-MS Salt Basin8
5%f
2011
Texas—Permian Basin (far west portion)
0%f
2011
Texas—Permian Basin (Midland portion)
2%f
2011
Texas—Anadarko Basin
20%f
2011
West Virginia—Statewide
15%h
2012
Overall Mean1
8%
Overall Median'
5%
a All estimates in this table refer to the percentage of injected water volume that comes from reused hydraulic fracturing
wastewater. However, different literature sources used slightly different terminology when referring to this percentage. In the
table footnotes below, we reference the terminology reported in the literature source cited.
b Produced water as a percentage of total water volume for 480 well stimulations according to completion reports between
January 1, 2014, and December 10, 2014 (CCST. 2015a). All but two of these stimulations were conducted in Kern County,
California (the remaining two were completed in Ventura County, California). Well stimulations mostly consisted of hydraulic
fracturing operations, but also included smaller numbers of matrix acidizing and acid fracturing operations (CCST. 2015a).
c Reflects an assumption of reuse practices by Noble Energy in the Wattenberg Field of Colorado's Denver-Julesburg Basin, as
reported by Goodwin et al. (2014).
d Percentage of recycled water used in hydraulic fracturing in 2014 based on data from the Pennsylvania Bureau of Topographic
and Geologic Survey (Schmid and Yoxtheimer. 2015). This percentage was higher at 23% in 2013, but we present the most
recent estimate available in the above table. The slight decline to 19% in 2014 may be explained by the fact that some
completion reports had not yet been processed when these data were published, yet the data generally show an upward trend
overtime in reuse as a percentage of injected volume (Schmid and Yoxtheimer. 2015).
6 Flowback water as a percentage of total water injected from July 2008 to December 2013 (SRBC. 2016). This percentage was
22% in 2013 alone (SRBC. 2016).
f Estimated percentage of recycling/reused water in 2011 (Nicot et al.. 2012).
g Nicot et al. (2012) refer to this region of Texas as the East Texas Basin.
h Reused fracturing water as a percentage of total water used for hydraulic fracturing in 2012, calculated from data provided by
the West Virginia PEP (2014).
' Calculated based on the values presented in Table 4-2, excluding the value for Pennsylvania's Susquehanna River Basin to
avoid double counting with the statewide value. The overall mean is not weighted by the number of wells in a given state,
basin, or play.
' Calculated based on the values presented in Table 4-2, excluding the value for Pennsylvania's Susquehanna River Basin to
avoid double counting with the statewide value. The overall median is not weighted by the number of wells in a given state,
basin, or play.
4-7
-------
Chapter 4- Water Acquisition
Available data on reuse trends indicate increased reuse as a percentage of injected volume over
time in both Pennsylvania and West Virginia, likely due to the lack of nearby disposal options in
Class II injection wells regulated by the Underground Injection Control (UIC) Program (Section
4.5.3).
The reuse of wastewater for hydraulic fracturing is limited by the amount of water that returns to
the surface during production (Nicot etal.. 20121. In the first 10 days of well production, 5% to
almost 50% of the hydraulic fracturing fluid volume can be collected, with values varying across
geologic formations (Chapter 7, Table 7-1). Longer duration measurements are rare, but between
10% and 30% of the hydraulic fracturing fluid volume has been collected in the Marcellus Shale in
Pennsylvania over nine years of production, while over 100% has been collected in the Barnett
Shale in Texas over six years of production (Chapter 7, Table 7-2 ).x Assuming that 10% of hydraulic
fracturing fluid volume is collected in the first 30 days and 100% of the wastewater is reused, it
would take 10 wells to produce enough water to hydraulically fracture a new well. As more wells
are hydraulically fractured in a given area, the potential for wastewater reuse increases.
The decision to reuse hydraulic fracturing wastewater appears to be driven by economics and the
quality of the wastewater, and not concerns over local water availability (Section 4.2.2). Water
transportation costs (i.e., trucking, piping), the availability of Class II wells, and local regulations can
play a role in determining whether hydraulic fracturing wastewater is reused to offset the need for
fresh water withdrawals fSchmid and Yoxtheimer. 20151. Besides hydraulic fracturing wastewater,
other wastewaters may be reclaimed for use in hydraulic fracturing. These include acid mine
drainage, wastewater treatment plant effluent, and other sources of industrial and municipal
wastewater (Nicot etal.. 2014: Ziemkiewicz etal.. 20131. Limited information is available on the
extent to which these other wastewaters are used.
4.2.2 Quality
Water quality is an important consideration when sourcing water for hydraulic fracturing. Fresh
water is most often used to maximize hydraulic fracturing fluid performance and to ensure
compatibility with the geologic formation being fractured. This finding is supported by the EPA's
analysis of disclosures to the FracFocus Chemical Disclosure Registry (version 1.0; hereafter, the
EPA FracFocus report) fU.S. EPA. 2015bl. as well as by regional analyses from Texas f Nicot etal..
1 It is possible to collect over 100% of the hydraulic fracturing fluid volume because water from the formation returns to
the surface along with the injected water.
4-8
-------
Chapter 4- Water Acquisition
20121 and the Marcellus Shale (Mitchell etal.. 2013a!.1'2 Fresh water was the most commonly cited
water source by companies included in an analysis of nine hydraulic fracturing service companies
on their operations from 2005 to 2010 fU.S. EPA. 2013al. Three service companies noted that the
majority of their water was fresh, because it required minimal testing and treatment fU.S. EPA.
2013a).3 The majority of the nine service companies recommended testing for certain water quality
parameters (pH and maximum concentrations of specific cations and anions) in order to ensure
compatibility among the water, other fracturing fluid constituents, and the geologic formation fU.S.
EPA. 2013al.
The reuse of hydraulic fracturing wastewater may be limited to an extent by water quality. Over the
production life of a well, the quality of the wastewater produced begins to resemble the quality of
the water naturally found in the geologic formation and may be characterized by high
concentrations of total dissolved solids (TDS) f Goodwin et al.. 20141. High concentrations of TDS
and other individual dissolved constituents in wastewater, including specific cations (calcium,
magnesium, iron, barium, strontium), anions (chloride, bicarbonate, phosphate, and sulfate), and
microbial agents, can interfere with hydraulic fracturing fluid performance by producing scale in
the borehole or by interfering with certain additives in the hydraulic fracturing fluid (e.g., high TDS
may inhibit the effectiveness of friction reducers) (Gregory etal.. 2011: North Dakota State Water
Commission. 2010). Due to these limitations, wastewater can require treatment or blending with
fresh water to meet the level of water quality desired in the hydraulic fracturing fluid formulation.4
Options for treating hydraulic fracturing wastewater to facilitate reuse are available and being used
by the industry in some cases. For example, filter socks, centrifuge, dissolved air flotation, or
settling technologies can remove suspended solids, and physical/chemical precipitation or
electrocoagulation can remove dissolved metals (Schmid and Yoxtheimer. 2015). For more
information on treatment of hydraulic fracturing wastewater, see Chapter 8.
1 The FracFocus Chemical Disclosure Registry (often referred to as FracFocus; www.fracfocus.orgl is a national hydraulic
fracturing chemical disclosure registry managed by the Ground Water Protection Council and the Interstate Oil and Gas
Compact Commission. FracFocus was created to provide the public access to reported chemicals used for hydraulic
fracturing within their area. It was originally established in 2011 (version 1.0] for voluntary reporting by participating oil
and gas well operators. Six of the 20 states discussed in this assessment required disclosure to FracFocus at various
points between January 1,2011, and February 28,2013, the time period analyzed by the EPA; another three of the 20
states offered the choice of reporting to FracFocus or the state during this same time period (see Appendix Table B-5 for
states and disclosure start dates] fU.S. EPA. 2015bl
2 Of all disclosures reviewed that indicated a source of water for the hydraulic fracturing base fluid, 68% listed "fresh" as
the only source of water used. Note, 29% of all disclosures considered in the EPA's FracFocus report included information
on the source of water used for the base fluid (U.S. EPA. 201 Sb],
3 Service companies did not provide data on the percentage of fresh water versus non-fresh water used for hydraulic
fracturing fU.S. EPA. 2013a!
4 The EPA FracFocus report suggests that fresh water makes up the largest proportion ofthe base fluid when blended
with water sources of lesser quality ("U.S. EPA. 2015bl
4-9
-------
Chapter 4- Water Acquisition
4.2.3 Provisioning
Water for hydraulic fracturing is typically either self-supplied by the industry or purchased from
public water systems.1 Self-supplied water for fracturing generally refers to permitted direct
withdrawals from surface water or groundwater or the reuse of wastewater. Nationally,
self-supplied water is more common, although there is much regional variation (U.S. EPA. 2015b:
CCST. 2014: Mitchell etal.. 2013a: Nicot etal.. 20121. Water purchased from municipal public water
systems can be provided either before or after treatment fNicot etal.. 20141. Water for hydraulic
fracturing is also sometimes purchased from smaller private entities, such as local land owners
fNicot etal.. 20141.
4.3 Water Use Per Well
In this section, we provide an overview of the amount of water used per well during hydraulic
fracturing. We discuss water use in the life cycle of oil and gas operations (Section 4.3.1) and
national per well estimates and associated variability (Section 4.3.2). More detailed locality-specific
information on water use per well is provided in Section 4.5.
4.3.1 Hydraulic Fracturing Water Use in the Life Cycle of Oil and Gas
Water is needed throughout the life cycle of oil and gas production and use, including both at the
well for processes such as well pad preparation, drilling, and fracturing (i.e., the upstream portion),
and later for end uses such as electricity generation, home heating, or transportation (i.e., the
downstream portion) fliangetal.. 2014: Laurenzi and Tersev. 20131. Most of the upstream water
usage and consumption occurs during hydraulic fracturing (Jiang etal.. 2014: Clark etal.. 2013:
Laurenzi and Tersev. 20131.2 Water use per well estimates in this chapter focus on hydraulic
fracturing in the upstream portion of the oil and gas life cycle, as the downstream portion of the
lifecycle is outside the scope of this assessment.3
1 According to Section 1401(4] of the Safe Drinking Water Act, a public water system is defined as system that provides
water for human consumption from surface water or groundwater through pipes or other infrastructure to at least 15
service connections, or an average of at least 25 people, for at least 60 days per year. Public water systems may either be
publicly or privately owned.
2 Laurenzi and lersev ("20131 reported that hydraulic fracturing accounted for 91% of upstream water consumption,
based on industry data for 29 wells in the Marcellus Shale. (91% was calculated from their paper by dividing hydraulic
fracturing fresh water consumption (13.7 gal (51.9 L]/Megawatt-hour (MWh]] by total upstream fresh water
consumption (15.0 gal (56.8 L]/MWh] and multiplying by 100]. Similarly, liangetal. (2014] reported that 86% of water
consumption occurred at the fracturing stage for the Marcellus Shale, based on Pennsylvania Department of
Environmental Protection (PA DEP] data on 500 wells. The remaining water was used in several upstream processes (e.g.,
well pad preparation, well drilling, road transportation to and from the wellhead, and well closure once production
ended]. Clark et al. ("20131 estimated lower percentages (30%—80%] of water use at the fracturing stage for multiple
formations. Although their estimates for the fraction of water used at the fracturing stage may be low due to their higher
estimates for transportation and processing, the estimates bv Clark etal. ("20131 similarly illustrate the importance of the
hydraulic fracturing stage in water use, particularly in terms of the upstream portion of the life cycle.
3 When the full life cycle of oil and gas production and use is considered (i.e., both upstream and downstream water use],
most water is used and consumed downstream. For example, in a life cycle analysis of hydraulically fractured gas used for
electricity generation. Laurenzi and lersev ("20131 reported that only 6.7% of water consumption occurred upstream
(15.0 gal (56.8 L]/MWh], while 93.3% of fresh water consumption occurred downstream for power plant cooling via
4-10
-------
Chapter 4- Water Acquisition
4.3.2 National Estimates and Variability in Water Use Per Well for Hydraulic Fracturing
At its most basic level, the volume of water used per well for hydraulic fracturing equals the
concentration of water in the hydraulic fracturing fluid multiplied by the total volume of the fluid
injected. In turn, the total volume of fluid injected generally equals the volume of fluid in the
fractures, plus the volume of the well itself, plus any fluid lost due to "leakoff" or other unintended
losses.1
Nationally, most operators employ fracturing fluids with water as a base fluid, meaning the
concentration of water in the fluid is high (U.S. EPA. 2015b: Yangetal.. 2013: GWPC and ALL
Consulting. 20091. The EPA inferred that more than 93% of reported disclosures to FracFocus used
water as a base fluid fU.S. EPA. 2015bl. The median reported concentration of water in the
hydraulic fracturing fluid was 88% by mass, with 10th and 90th percentiles of 77% and 95%,
respectively. Only roughly 2% of disclosures (761 wells) reported the use of non-aqueous
substances as base fluids, typically either liquid-gas mixtures of nitrogen or carbon dioxide. Both of
these formulations still contained substantial amounts of water, as water made up roughly 60%
(median value) of the fluid in them (U.S. EPA. 2015b). Other formulations were rarely reported.
Fluid formulations are discussed further in Chapter 5.
On average, hydraulic fracturing requires more than a million gallons (3.8 million liters) of water
per well. Tackson etal. (2015) reported a national average of 2.4 million gal (9.1 million L) of water
per well, calculated from FracFocus disclosures between 2010 and 2013. According to the EPA's
project database of disclosures to FracFocus 1.0 (hereafter the EPA FracFocus 1.0 project database),
the median volume of water used per well was 1.5 million gal (5.7 million L) between 2011 and
early 2013, based on 37,796 disclosures nationally (U.S. EPA. 2015b. c).2 Data on reported
Information Handling Services well numbers and median volumes in Gallegos etal. (2015) show
that overall per well volumes have increased in recent years from approximately 1.5 million gal (5.7
million L) in 2011 to 2.7 million gal (10.2 million L) in 2014.3
The recent increase in water use per well has been driven primarily by the proportional increase in
horizontal wells f Gallegos etal.. 20151 (Figure 4-1). Increases in horizontal well length affect total
volumes injected primarily by allowing a larger fracture volume to be stimulated (Economides et
al.. 20131. As horizontal wells get longer, fracture, well, and total volumes all increase. Importantly,
increases in the well length and water use per well do not necessarily mean an increase in water
intensity (the amount of water used per unit energy extracted). Goodwin et al. (2014) found water
evaporation (209.0 gal (791.2 L]/MWh]. Similar results were found for gas extraction in the Eagle Ford Shale fScanlon et
al..2014bl.
1 Leakoff is the fraction of the hydraulic fracturing fluid that infiltrates into the formation (e.g., through an existing natural
fissure] and is not recovered during production. This water lost to the formation can be a substantial fraction of the water
injected fO'Mallev etal.. 20151. See Chapter 6 for more information about leakoff and some recent findings related to the
relationship between hydraulic fracturing fluid volume and fracture volume.
2 All water use data included in the EPA's FracFocus 1.0 project database were obtained from disclosures made to
FracFocus. Although disclosures were made on a per well basis, a small proportion of the wells were associated with
more than one disclosure (i.e., 876 out of 37,114, based on unique API numbers] fU.S. EPA. 2015cl. For the purposes of
this chapter, we discuss water use per disclosure in terms of water use per well.
3 Derived from supporting information in Gallegos etal. ("20151. Calculated by multiplying the median volume by the
number of wells for each well type, then summing volumes across well types, and dividing by the total number of wells.
4-11
-------
Chapter 4- Water Acquisition
intensity did not increase in the Denver Basin despite increases in well length and water use per
well.
Directional Oil
Directional Gas
Horizontal Oil
Horizontal Gas
•Vertical Oil
¦Vertical Gas
¦x— Estimated Median
Across All Well Types
Completion Year
Figure 4-1. Median water volume per hydraulically fractured well nationally, expressed by
well type and completion year.
Adapted using data from Gallegos et al. (2015). Note: shown in orange is the estimated median across all well
types, derived from Gallegos et al. (2015) supporting information Tables S2 and S3. Calculated by multiplying the
median volume by the number of wells for each well type, then summing volumes across well types, and dividing
by the total number of wells for each year. This estimated median across all well types reflects the central
tendency of the data, and was calculated because the individual data are proprietary and not published,
preventing the calculation of an overall median.
There is substantial variation around these per well estimates. For instance, the 10th and 90th
percentiles from the EPA FracFocus 1.0 project database are 74,000 gal and 6 million gal (280,000 L
and 23 million L) per well, respectively.1 Even in specific basins, plays, and within a single oil and
gas field, water use per well varies widely. For example, Laurenzi and Jersey (2013) reported
volumes ranging from 1 to 6 million gal (3.8 to 23 million L) per well (10th to 90th percentile) in the
Wattenberg Field in Colorado.
Of the major unconventional formation types discussed in Chapter 2 (shales, tight formations-
including tight sands or sandstones, and coalbeds), coalbeds generally require less water per well.
1 Although the EPA FracFocus report shows 5th and 95th percentiles, we report 10th and 90th percentiles throughout this
chapter to further reduce the influence of outliers.
4-12
-------
Chapter 4- Water Acquisition
Coalbed methane (CBM) comes from coal seams that often have a high initial water content and
tend to occur at much shallower depths fU.S. EPA. 2015kl. In part because of the shallower depths,
shorter well lengths result in lower water use per well, often by an order of magnitude or more
compared to operations in shales or tight formations (e.g.. Murray. 2013).
4.4 Hydraulic Fracturing Water Use and Consumption at the National, State,
and County Scale
In this section, we provide an overview of water use and consumption for hydraulic fracturing at
the national, state, and county scale. We then compare these values to total water use and
consumption at these scales. We do this to contextualize hydraulic fracturing water use and
consumption with total water use and consumption, and to illustrate whether hydraulic fracturing
is a relatively large or small user and consumer of water at these scales. Later, we compare
hydraulic fracturing water use to water availability estimates at the county scale (Text Box 4-2).
Water use is water withdrawn for a specific purpose, part or all of which may be returned to the
local hydrologic cycle. Water consumption is water that is removed from the local hydrologic cycle
following its use (e.g., via evaporation, transpiration, incorporation into products or crops,
consumption by humans or livestock), and is therefore unavailable to other water users (Maupin et
al.. 20141. Hydraulic fracturing water consumption can occur through evaporation from storage
ponds, the retention of water in the subsurface through imbibition, or disposal in Class II wells,
among other means.
Hydraulic fracturing water use is a function of the water use per well and the total number of wells
fractured at a given spatial scale during the time period analyzed, calculated from the EPA
FracFocus 1.0 project database (U.S. EPA. 2015c). Water consumption estimates are derived from
United States Geological Survey (USGS) water use data, and therefore both use and consumption
are presented with the published water use numbers being first
4.4.1 National and State Scale
Hydraulic fracturing uses and consumes billions of gallons of water each year in the United States,
but at the national and state scales, it is a relatively small user and consumer of water compared to
total water use and consumption. According to the EPA's FracFocus 1.0 project database, hydraulic
fracturing used 36 billion gal (136 billion L) of water in 2011 and 52 billion gal (197 billion L) in
2012, yielding an average annually of 44 billion gal (167 billion L) of water in 2011 and 2012 across
all 20 states in the project database fU.S. EPA. 2015b. c). National water use for hydraulic fracturing
can also be estimated by multiplying the water use per well by the number of wells hydraulically
fractured. If the median water use per well (1.5 million gal) (5.7 million L) from the EPA's
FracFocus 1.0 project database is multiplied by 25,000 to 30,000 wells fractured annually (Chapter
3), national water use for hydraulic fracturing is estimated to range from 38 to 45 billion gal (142 to
170 billion L) annually. Other calculated estimates have ranged higher than this, including
estimates of approximately 80 billion gal (300 billion L) fVengosh etal.. 20141 and 50 to 72 billion
gal (190-273 billion L) fU.S. EPA. 2015el. These estimates are higher due to differences in the
estimated water use per well and the number of wells used as multipliers. For example, Vengosh et
4-13
-------
Chapter 4- Water Acquisition
al. (20141 derived the estimate of approximately 80 billion gal (300 billion L) by multiplying an
average of 4.0 million gal (15 million L) per well (estimated for shale gas wells) by 20,000 wells
(the approximate total number of fractured wells in 2012).1
All of these estimates of water use for hydraulic fracturing are small relative to total water use and
consumption at the national scale. The USGS compiles national water use estimates every five years
in the National Water Census, with the most recent census conducted in 2010 (Maupin etal..
2014).2 The USGS publishes water use, not consumption estimates, yet by applying consumption
factors for each use category in the 2010 National Water Census, we derived estimates of total
water consumption. We also used a consumption factor to estimate hydraulic fracturing water
consumption from values in the EPA FracFocus 1.0 project database.3 Comparing these estimates,
average annual hydraulic fracturing water use in 2011 and 2012 was less than 1% of total 2010
annual water use for all of the 20 states combined where operators reported water use to
FracFocus in 2011 and 2012. Hydraulic fracturing water consumption followed the same pattern
when compared to total water consumption (Appendix Table B-l).4
At the state scale, hydraulic fracturing also generally uses billions of gallons of water, but accounts
for a low percentage of total water use or consumption. Of all states in the EPA FracFocus 1.0
project database, operators in Texas used the most water (47% of water use reported in the EPA
FracFocus 1.0 project database) (U.S. EPA. 2015c) (Appendix Table B-l). This was due to the large
number of wells in that state, since hydraulic fracturing water use is proportional to the number of
wells. Over 94% of reported water use occurred in just seven of the 20 states in the EPA FracFocus
1.0 project database (listed in order of highest statewide hydraulic fracturing water use): Texas,
Pennsylvania, Arkansas, Colorado, Oklahoma, Louisiana, and North Dakota (U.S. EPA. 2015c)
(Appendix Table B-l). Hydraulic fracturing is a small percentage when compared to total water use
(<1%) and consumption (<3%) in each individual state (Appendix Table B-l). Other studies have
shown similar results, with hydraulic fracturing water use and consumption ranging from less than
1 This could result in an overestimation because the estimate of 20,000 wells was derived in part from FracFocus, and
these wells are not necessarily specific to shale gas; they may include other types of wells that use less water (e.g., CBM].
The estimate of 1.5 million gal (5.7 million L] per well based on the U.S. EPA (~2015c1 FracFocus 1.0 project database likely
leads to a more robust estimate when used to calculate national water use for hydraulic fracturing because it includes
wells from multiple formation types (i.e., shale, tight sand, and CBM], some of which use less water than shale gas wells on
average.
2 The National Water Census includes uses such as public supply, irrigation, livestock, aquaculture, thermoelectric power,
industrial, and mining at the national, state, and county scale. The 2010 National Water Census included hydraulic
fracturing water use in the mining category; there was no designated category for hydraulic fracturing alone.
3 See footnotes for Appendix Table B-l or for Table 4-3 for a description of the consumption estimate calculations.
4 Water use percentages were calculated by averaging annual water use for hydraulic fracturing in 2011 and 2012 for a
given state or county ("U.S. EPA. 2015cl. and then dividing by 2010 USGS total water use ("Maupin et al.. 20141 and
multiplying by 100. Note, the annual hydraulic fracturing water use reported in FracFocus was not added to the 2010
total USGS water use value in the denominator, and is simply expressed as a percentage compared to 2010 total water use
or consumption. This was done because of the difference in years between the two datasets, and because the USGS 2010
Water Census (Maupin etal.. 2014) included hydraulic fracturing water use estimates in their mining category. This
approach is consistent with that of other literature on this topic; see Nicotand Scanlon (2012). Consumption estimates
were calculated in the same manner, except consumption, not use, values were employed. County level data from the
USGS 2010 Water Census are available online at http: //water.usgs.gov/watuse/data/2010/ ("accessed November 11,
2014],
4-14
-------
Chapter 4- Water Acquisition
1% of total use in West Virginia (West Virginia DEP. 20131. Colorado (Colorado Division of Water
Resources etal.. 20141. and Texas fNicotetal.. 2014: Nicotand Scanlon. 20121. to approximately
4% in North Dakota (North Dakota State Water Commission. 2014).
4.4.2 County Scale
Water use and consumption for hydraulic fracturing is also relatively small in most, but not all,
counties in the United States (Table 4-3; Figure 4-2; Figure 4-3a,b; and Appendix Table B-2). Based
on the EPA FracFocus 1.0 project database, reported fracturing water use in 2011 and 2012 was
less than 1% compared to 2010 USGS total water use in 299 of the 401 reporting counties (Figure
4-3a; Appendix Table B-2). However, hydraulic fracturing water use was 10% or more compared to
total water use in 26 counties, 30% or more in nine counties, and 50% or more in four counties
(Table 4-3; Figure 4-3a). McMullen County in Texas had the highest percentage at over 100%
compared to 2010 total water use.1 Total consumption estimates followed the same pattern, but
with more counties in the higher percentage categories (hydraulic fracturing water consumption
was 10% or more compared to total water consumption in 53 counties; 30% or more in 25
counties; 50% or more in 16 counties; and over 100% in four counties) (Table 4-3; Figure 4-3b).
Estimates based on the EPA's FracFocus 1.0 project database may form an incomplete picture of
hydraulic fracturing water use in a given state or county, because the majority of states with data in
the project database did not require disclosure to FracFocus during the time period analyzed (U.S.
EPA. 2015bl. We conclude that this likely does not substantially alter the overall patterns observed
in Figure 4-3a,b. See Text Box 4-1 for further details. These percentages also depend both upon the
absolute water use and consumption for hydraulic fracturing and the relative magnitude of other
water uses and consumption in that state or county. For instance, a rural county with a small
population might have relatively low total water use prior to hydraulic fracturing.2 Also, just
because water is used in a certain county does not necessarily mean it originated in that county.
The cost of trucking water can be substantial fSlutz etal.. 20121. and the industry tends to acquire
water from nearby sources when possible (Section 4.2.1); however, water can also be piped in from
more distant, regional supplies. Despite these caveats, it is clear that hydraulic fracturing is
generally a relatively small user (and consumer) of water at the county level, with the exception of a
small number of counties where water use and consumption for fracturing can be high relative to
other uses and consumption.
1 Estimates of use or consumption exceeded 100% when hydraulic fracturing water use averaged for 2011 and 2012
exceeded total water use or consumption in that county in 2010.
2 For example, McMullen County, Texas, mentioned above contains a small number of residents (707 people in 2010,
according to the U.S. Census Bureau ("20141.
4-15
-------
Chapter 4- Water Acquisition
Table 4-3. Average annual hydraulic fracturing water use and consumption in 2011 and 2012
compared to total annual water use and consumption in 2010, by county.
Only counties where hydraulic fracturing water was 10% or greater compared to 2010 total water use are shown
(for full table, see Appendix Table B-2). Average annual hydraulic fracturing water use data in 2011 and 2012 from
the EPA's FracFocus 1.0 project database (U.S. EPA, 2015c). Total annual water use data in 2010 from the USGS
(Maupin et al„ 2014). States listed by order of appearance in the chapter.
State
County
Total annual
water use in
2010 (millions
of gal)a
Average annual
hydraulic
fracturing water
use in 2011 and
2012
(millions of gal)b
Hydraulic
fracturing
water use
compared to
total water
use (%)c
Hydraulic fracturing
water consumption
compared to total
water consumption
(%)c-d
Texas
McMullen
657.0
745.9
113.5
350.4
Karnes
1861.5
1055.2
56.7
120.1
La Salle
2474.7
1288.7
52.1
93.7
Dimmit
4073.4
1794.2
44.0
81.3
Irion
1335.9
411.4
30.8
74.5
Montague
3989.5
925.3
23.2
77.8
De Witt
2394.4
546.6
22.8
48.6
Loving
781.1
138.4
17.7
94.1
San Augustine
1131.5
182.1
16.1
50.8
Live Oak
1916.3
294.0
15.3
40.1
Wheeler
6522.6
858.0
13.2
21.5
Cooke
4533.3
454.3
10.0
29.9
Pennsylvania
Susquehanna
1617.0
751.3
46.5
123.4
Sullivan
222.7
66.5
29.9
79.8
Bradford
4354.5
1059.4
24.3
78.2
Tioga
2909.1
566.3
19.5
47.3
Lycoming
5854.6
704.6
12.0
33.8
West Virginia
Doddridge
405.2
78.5
19.4
69.4
Ohio
Carroll
1127.9
152.7
13.5
37.3
North Dakota
Mountrail
1248.3
449.4
36.0
98.3
Dunn
1076.8
309.5
28.7
43.1
Burke
394.2
63.6
16.1
40.8
Divide
806.7
102.2
12.7
18.6
4-16
-------
Chapter 4- Water Acquisition
State
County
Total annual
water use in
2010 (millions
of gal)a
Average annual
hydraulic
fracturing water
use in 2011 and
2012
(millions of gal)b
Hydraulic
fracturing
water use
compared to
total water
use (%)c
Hydraulic fracturing
water consumption
compared to total
water consumption
(%)c-d
Arkansas
Van Buren
1587.8
899.6
56.7
168.8
Louisiana
Red River
1606.0
569.6
35.5
83.2
Sabine
1522.1
395.2
26.0
76.6
a County level data accessed from the USGS website (http://water.usgs.gov/watuse/data/2010/) on November 11, 2014. Total
water withdrawals per day were multiplied by 365 days to estimate total water use for the year (Maupin et al.. 2014).
b Average of water used for hydraulic fracturing in 2011 and 2012 calculated from the EPA FracFocus 1.0 project database (U.S.
EPA. 2015c).
c Percentages were calculated by averaging annual water use for hydraulic fracturing reported in FracFocus in 2011 and 2012 for
a given state or county (U.S. EPA. 2015c). and then dividing by 2010 USGS total water use (Maupin et al.. 2014) and multiplying
by 100.
d Consumption values were calculated with use-specific consumption rates predominantly from the USGS, including 19.2% for
public supply, 19.2% for domestic use, 60.7% for irrigation, 60.7% for livestock, 14.8% for industrial uses, 14.8% for mining
(Sollev et al.. 1998). and 2.7% for thermoelectric power (Diehl and Harris. 2014). We used rates of 71.6% for aquaculture from
Verdegem and Bosma (2009) ((evaporation per kg fish + infiltration per kg)/total water use per kg); and 82.5% for hydraulic
fracturing (consumption value calculated by taking the median value for all reported produced water/injected water
percentages in Tables 7-1 and 7-2 of this assessment and then subtracting from 100%). If a range of values was given, the
midpoint was used. Note, this aspect of consumption is likely a low estimate since much of this produced water (injected water
returning to the surface) is not subsequently treated and reused, but rather disposed of in Class II wells - see Chapter 8.
4-17
-------
Chapter 4 - Water Acquisition
Average Annual
Water Use in Mgal
(number of counties)
| >500(27)
| 100-500(60)
| 10-100(86)
1-10(115)
<1(113)
ElAshale basins
n State reporting
requirement
Figure 4-2. Average annual hydraulic fracturing water use in 2011 and 2012 by county.
Source: U.S. EPA (2015c). Water use in millions of gallons (Mgal). Counties shown with respect to major U.S. Energy Information Administration (EIA) shale
basins (EIA, 2015). Orange borders identify states that required some degree of reporting to FracFocus in 2011 and 2012.
4-18
-------
Chapter 4- Water Acquisition
Hydraulic fracturing water use
compared to total water use
(number of counties)
>100(1)
¦ 30%-100% (8)
10%-30% (17)
1%-10% (76)
< 1 (299)
| EIA shale basins
| State reporting requirement
(a)
Hydraulic fracturing water
consumption compared to total
water consumption (number
of counties)
| >100(4)
| 30%-100% (21)
| 10% -30% (28)
1%-10% (81)
< 1 (267)
EIA shale basins
n State reporting requirement
(b)
Figure 4-3. (a) Average annual hydraulic fracturing water use in 2011 and 2012 compared to
total annual water use in 2010, by county, expressed as a percentage; (b) Average annual
hydraulic fracturing water consumption in 2011 and 2012 compared to total annual water
consumption in 2010, by county, expressed as a percentage.
Average annual hydraulic fracturing water use data in 2011 and 2012 from the EPA's FracFocus 1.0 project
database (U.S. EPA, 2015c). Total annual water use data in 2010 from the USGS (Maupin et al., 2014). See Table 4-3
for descriptions of calculations for estimating consumption. Counties shown with respect to major U.S. EIA shale
basins (EIA, 2015). Orange borders identify states that required some degree of reporting to FracFocus in 2011 and
2012. Note: Values over 100% denote counties where the average annual hydraulic fracturing water use or
consumption in 2011 and 2012 exceeded the total annual water use or consumption in that county in 2010.
4-19
-------
Chapter 4- Water Acquisition
Text Box 4-1. Using the EPA's FracFocus 1.0 Project Database to Estimate Water Use for
Hydraulic Fracturing.
The FracFocus Chemical Disclosure Registry (often referred to as FracFocus; www.fracfocus.org1 is a national
hydraulic fracturing chemical disclosure registry managed by the Ground Water Protection Council and the
Interstate Oil and Gas Compact Commission. FracFocus was created to provide the public access to reported
chemicals used for hydraulic fracturing within their area. It was originally established in 2011 (version 1.0)
for voluntary reporting by oil and gas well operators. The EPA used the data available from FracFocus
between January 1, 2011 and February 28,2013 to develop the EPA FracFocus 1.0 project database; the
database and a related EPA report were both peer reviewed and published fU.S. EPA. 2015b. c). Six of the 20
states discussed in this assessment required disclosure to FracFocus at various points during this time;
another three of the 20 states offered the choice of reporting to FracFocus or the state during this same time
period (U.S. EPA. 2015b). Estimates based on the EPA's FracFocus 1.0 project database could form an
incomplete picture of hydraulic fracturing water use, because most states with data in the project database
(14 out of 20) did not require disclosure to FracFocus during the time period analyzed fU.S. EPA. 2015bl
Water use for fracturing is a function of the water use per well and the total number of wells fractured over a
given spatial area or a given period of time. For water use per well, we found seven literature values for
comparison with values from the EPA's FracFocus 1.0 project database. On average, water use estimates per
well in the project database were 77% of literature values (the median was 86%); Colorado's Denver Basin
was the only location where the project database estimate as a percentage of the literature estimate was low
(14%) (Appendix Table B-3). In general, water use per well estimates from the EPA's FracFocus 1.0 project
database appear to align closely with the literature estimates for most areas for which we have data, with the
exception of the Denver Basin of Colorado.
For the number of wells, we compared data in the EPA's FracFocus 1.0 project database to numbers available
in state databases from North Dakota, Pennsylvania, and West Virginia (Appendix Table B-4). These were the
state databases from which we could distinguish hydraulically fractured wells from other oil and gas wells.
On average, we found that the EPA FracFocus 1.0 project database included 67% of the wells listed in state
databases for 2011 and 2012 (Appendix Table B-4). Unlike North Dakota and Pennsylvania, West Virginia did
not require operators to report fractured wells to FracFocus during this time period, possibly explaining its
lower reporting rate. Multiplying the average EPA FracFocus 1.0 project database values of 77% for water use
per well and 67% for well counts yields 52%. Thus, the EPA FracFocus 1.0 project database estimates for
water use could be slightly over half of the estimates from these three state databases during this time period.
These values are based on small sample sizes (seven literature values and three state databases) and should
be interpreted with caution. Nevertheless, these numbers suggest that estimates based on the EPA's
FracFocus 1.0 project database likely form an incomplete picture of hydraulic fracturing water use during this
time period.
To assess how this might affect hydraulic fracturing water use estimates in this chapter, we doubled the
water use value in the EPA's FracFocus 1.0 project database for each county, an adjustment much higher than
any likely underestimation. Even with this adjustment, fracturing water use was still less than 1% compared
to 2010 total water use in the majority of the 401 U.S. counties represented in the EPA FracFocus 1.0 project
database (299 counties without adjustment versus 280 counties with adjustment). The number of counties
where hydraulic fracturing water use was 30% or more of 2010 total county water use increased from nine to
21 with the adjustment.
These results indicate that most counties have relatively low hydraulic fracturing water use relative to total
water use, even when accounting for likely underestimates. Since consumption estimates are derived from
use, these will also follow the same pattern. Thus, potential underestimates based on the EPA's FracFocus 1.0
project database likely do not substantially alter the overall pattern shown in Figure 4-3. Rather,
underestimates of hydraulic fracturing water use would mostly affect the percentages in the small number of
counties where fracturing already constitutes a higher percentage of total water use and consumption.
4-20
-------
Chapter 4- Water Acquisition
4.5 Potential for Impacts by Location
The potential for hydraulic fracturing water acquisition to impact drinking water availability or
alter its quality depends on the balance between water withdrawals and water availability at a
given location. Where water availability is high compared to the volume of water withdrawn for
hydraulic fracturing, this water use can be accomodated. However, where water availability is low
and hydraulic fracturing water use is high, these withdrawals are more likely to impact drinking
water resources. The balance between withdrawals and availability can vary greatly by geographic
location. Moreover, a combination of regional or site-specific factors can alter this balance, making
impacts more or less likely, or more or less severe. For these reasons, we discuss the various factors
and potential for impacts by geographic location in the following section.
We organize this discussion by state, addressing 15 states accounting for almost all disclosures
reported in the EPA FracFocus 1.0 project database fU.S. EPA. 2015cl: Texas (Section 4.5.1);
Colorado and Wyoming (Section 4.5.2); Pennsylvania, West Virginia, and Ohio (Section 4.5.3); North
Dakota and Montana (Section 4.5.4); Arkansas and Louisiana (Section 4.5.5), Oklahoma and Kansas
(Appendix B.2.1); and Utah, New Mexico, and California (Appendix B.2.2). We highlight the states
that best illustrate concepts relating to the potential for impacts, or factors that affect the frequency
or severity of these impacts in Section 4.5; the remaining states are discussed in Appendix B.2.
Within Section 4.5 and Appendix B, we address each state in order of most hydraulically fractured
wells to least, and combine states with similar geographies or activity. For certain states, we
address major oil and gas regions separately (e.g., the Permian Basin in Texas). Each section
describes the number of fractured wells in that state or region, the type of water used, water use
per well, and water use estimates at the county scale. We then discuss the potential for impacts by
comparing water use and water availability and addressing factors (e.g., drought or the amount of
water reused to offset fresh water use) that might alter the frequency or severity of impacts. As
noted in the chapter introduction, we use several lines of evidence to evaluate the potential for
impacts and factors for each location. We use the scientific literature, county level assessments, and
local case studies where available.
4.5.1 Texas
Hydraulic fracturing in Texas accounts for the bulk of the activity reported nationwide, comprising
48% of the disclosures in the EPA FracFocus 1.0 project database (U.S. EPA. 2015c) (Figure 4-4;
Appendix Table B-5). There are five major basins in Texas: the Permian, Western Gulf (includes the
Eagle Ford play), Fort Worth (includes the Barnettplay), TX-LA-MS Salt (includes the Haynesville
play), and the Anadarko (Figure 4-5); together, these five basins contain 99% of Texas' reported
wells (Appendix Table B-5).
4-21
-------
Chapter 4 - Water Acquisition
San
Joaquin
Montana
Thrust Belt
ikken
Vay
Bighorn ,
Powder
River
Denver
Michigan
Raton •+XWo n &
rmian Barneff Hayne$Mi?>gpsi
WBf
.Weste^i
Gulf
Black Warrior
Floyd-Neal
Play
>s/er TX-LA-MS
Salt
•
FracFocus Wells
EIA Plays
EIA Basins
Figure 4-4. Locations of wells in the EPA FracFocus 1.0 project database, with respect to U.S.
EIA shale plays and basins.
Note: Hydraulic fracturing can be conducted in geologic settings other than shale; therefore, some wells on this
map are not associated with any EIA shale play or basin (EIA, 2015; U.S. EPA, 2015c).
A ladarko
Fort Worth
TX-LA-MS
Salt Marsh
Permian
Barnetl
Play
Hay tics ville-Bossiir
jjfP Play (
Western Gulf
Figure 4-5. Major U.S. EIA shale plays and basins for Texas.
Source: EIA (2015).
4-22
-------
Chapter 4- Water Acquisition
Types of water used: What is known about water sources in Texas largely comes from direct surveys
and interviews with industry operators and water suppliers fNicot etal.. 2014: Nicot etal.. 20121.
Overall, groundwater is the dominant source throughout most of the state fNicot etal.. 2014: Nicot
etal.. 20121 (Table 4-1). The exception is the Barnett Shale, where both surface water and
groundwater are used in approximately equal proportions.
Hydraulic fracturing in Texas uses mostly fresh water fNicot etal.. 20121.1 The exception is the far
western portion of the Permian Basin, where brackish water makes up an estimated 80% of total
hydraulic fracturing water use. Brackish water is used to a lesser extent in the Anadarko Basin, the
Midland portion of the Permian Basin, and the Eagle Ford Shale (Table 4-4). Reuse of wastewater as
a percentage of total water use is generally low (5% or less) in all major basins and plays in Texas,
except for the Anadarko Basin in the Texas Panhandle, where it is 20% fNicot etal.. 20121 (Table
4-2).
Table 4-4. Estimated brackish water use as a percentage of total hydraulic fracturing water
use in the main hydraulic fracturing areas of Texas, 2011.a
Adapted from Nicot et al. (2012).b
Play
Percentage
Barnett Shale
3%
Eagle Ford Shale
20%
Texas portion of the TX-LA-MS Salt Basin0
0%
Permian Basin—Far West
80%
Permian Basin—Midland
30%
Anadarko Basin
30%
a Nicot et al. (2012) define brackish water as any water with a total dissolved solids (TDS) content of >1,000 mg/L, but <35,000
mg/L, although they often limit that range to between 1,000 and 10,000 mg/L.
b Nicot et al. (2012) present the estimated percentages of brackish, recycled/reused, and fresh water relative to total hydraulic
fracturing water use so that the percentages of the three categories sum to 100%.
c Nicot et al. (2012) refer to this region of Texas as the East Texas Basin.
The majority of water used in Texas for hydraulic fracturing is self-supplied via direct ground or
surface water withdrawals fNicot etal.. 20141. Less often, water is purchased from local
landowners, municipalities, larger water districts, or river authorities fNicot etal.. 20141.
Water use per well: Water use per well varies across Texas basins, with reported medians from
2011 to early 2013 of 3.9 million gal (14.8 million L) in the Fort Worth Basin, 3.8 million gal
(14.4 million L) in the Western Gulf, 3.3 million gal (12.5 million L) in the Anadarko, 3.1 million gal
(11.7 million L) in the TX-LA-MS Salt, and 840,000 gal (3.2 million L) in the Permian (Appendix
1 The EPA FracFocus report shows that "fresh" was the only source of water listed in 91% of all disclosures reporting a
source of water in Texas fU.S. EPA. 2015bl. Nineteen percent of Texas disclosures included information related to water
sources fU.S. EPA. 2015bl
4-23
-------
Chapter 4- Water Acquisition
Table B-5). Relatively low water use in the Permian Basin, which contains roughly half the reported
wells in the state, is due to the abundance of vertical wells, mostly for oil extraction fNicot etal..
20121.
Water use per well is increasing in most locations in Texas. In the Barnett Shale, water use per well
increased from approximately 3 million gal (11 million L) in the mid-2000's to approximately 5
million gal (19 million L) in 2011 as the horizontal lengths of wells increased (Nicot etal.. 20141.
Similar increases in lateral length and water use per well were reported for the Texas-Haynesville,
East Texas, and Anadarko basins, and most of the Permian Basin (Nicot etal.. 2012: Nicot and
Scanlon. 20121.1
Water use/consumption at the county scale: Water use and consumption for hydraulic fracturing can
be significant in some Texas counties. Texas contains five of nine counties nationwide where
operators used more than 1 billion gal (3.8 billion L) of water annually for hydraulic fracturing, and
five of nine counties where fracturing water use in 2011 and 2012 was 30% or more compared to
total water use in those counties in 2010 (Table 4-3, Figure 4-3a; Appendix Table B-2).2
According to detailed county level projections, water use for hydraulic fracturing is expected to
increase with oil and gas production in the coming decades, peaking around the year 2030 (Nicot et
al.. 20121. These projections were made before the recent decline in oil and gas prices, and so are
highly uncertain. If these projections hold, the majority of counties are expected to have relatively
low water use for fracturing in the future, but hydraulic fracturing water use could equal or exceed
10%, 30%, and 50% compared to 2010 total county water use in 30, nine, and three counties,
respectively, by 2030 (Appendix Table B-7).
Potential for impacts: Of all locations surveyed in this chapter, the potential for water quantity and
quality impacts due to hydraulic fracturing water acquisition appears to be highest in southern and
western Texas. This area includes the Anadarko, the Western Gulf (Eagle Ford play), and the
Permian Basins. According to Ceres (20141. 28% and 87% of the wells fractured in the Eagle Ford
play and Permian Basin, respectively, are in areas of high to extremely high water stress.3 A
comparison of hydraulic fracturing water use to water availability at the county scale also suggests
the potential for impacts in this region (Text Box 4-2).
1 It should be noted that energy production also increases with lateral lengths, and therefore, water use per unit energy
produced—typically referred to as water intensity—may remain the same or decline despite increases in per-well water
use fNicot etal.. 2014: Laurenzi and lersev. 20131
2 Texas also contains 10 of the 25 counties nationwide where hydraulic fracturing water consumption was greater than or
equal to 30% of 2010 total water consumption (Table 4-3]. Nicot and Scanlon f20121 found similar variation among
counties when they compared hydraulic fracturing water consumption to total county water consumption for the Barnett
play. Their consumption estimates ranged from 581 million gal (2.20 billion L] in Parker County to 2.7 billion gal (10.2
billion L] in Johnson County, representing 10.5% and 29.1% compared to total water consumption in those counties,
respectively. Fracturing in Tarrant County, part of the Dallas Fort-Worth area, consumed 1.6 billion gal (6.1 billion L] of
water, 1.4% compared to total county water consumption fNicot and Scanlon. 20121
3 Ceres T20141 compared well locations to areas categorized by a water stress index, characterized as follows: extremely
high (defined as annual withdrawals accounting for greater than 80% of surface flows]; high (40-80%) of surface flows];
or medium-to-high (20-40%) of surface flows].
4-24
-------
Chapter 4- Water Acquisition
Text Box 4-2. Hydraulic Fracturing Water Use as a Percentage of Water Availability Estimates.
Researchers at Sandia National Laboratories assessed county level water availability across the continental
United States fTidwell et al.. 20131 Assessments of water availability in the United States are generally
lacking at the county scale, and this analysis—although undertaken for siting new thermoelectric power
plants—can be used to assess potential impacts of hydraulic fracturing water withdrawals.
The authors generated annual water availability estimates for five categories of water: unappropriated
surface water, unappropriated groundwater, appropriated water potentially available for purchase, brackish
groundwater, and wastewater from municipal treatment plants fTidwell et al.. 20131 In the western United
States, water is generally allocated by the principle of prior appropriation—that is, first in time of use is first
in right. New development must use unappropriated water or purchase appropriated water from vested
users. In their analysis, the authors assumed 5% of appropriated irrigated water could be purchased; they
also excluded wastewater required to be returned to streams and the wastewater fraction already reused.
Given regulatory restrictions, they considered no fresh water to be available in California for new
thermoelectric plants. Their definition of brackish water ranged from 3,000 to 10,000 ppm TDS, and from 50
to 2,500 ft (15-760 m) below the surface.
Combining their estimates of unappropriated surface water and groundwater and appropriated water
potentially available for purchase, we derived a fresh water availability estimate for each county (except for
those in California) and then compared this value to reported water use for hydraulic fracturing in 2011 and
2012 (U.S. EPA. 2015c). We also added the estimates of brackish groundwater and wastewater from
municipal treatment plants to fresh water estimates to derive estimates of total water availability and did a
similar comparison. Since the water availability estimates already take into account current water use for oil
and gas operations, these results should be used only as indicator of areas where shortages might arise in the
future. Here we focus on hydraulic fracturing water use compared to water availability. If we compared
hydraulic fracturing water consumption to water availability, consumption would be lower relative to
availability since by definition, water consumption is less than water use. Hence, water use versus availability
acts as an upper-bound estimate, and includes consumption.
Overall, hydraulic fracturing water use represented less than 1% of fresh water availability in over 300 of the
395 counties analyzed (Figure 4-6a], This result suggests that there is ample water available at the county
scale to accommodate hydraulic fracturing in most locations. However, there was a small number of counties
where hydraulic fracturing water use was a relatively high percentage of fresh water availability. In 17
counties, fracturing water use actually exceeded the index of fresh water available; all of these counties were
located in the state of Texas and were associated with the Anadarko, Barnett, Eagle Ford, and Permian
basins/plays (Figure 4-5). In Texas counties with relatively high brackish water availability, hydraulic
fracturing water use represented a much smaller percentage of total water availability (fresh + brackish +
wastewater) (Figure 4-6b). This finding illustrates that potential impacts can be avoided or reduced in these
counties through the use of brackish water or wastewater for hydraulic fracturing; a case study in the Eagle
Ford play in southwestern Texas echoes this finding (Text Box 4-3).
(Text Box 4-2 is continued on the following page.)
4-25
-------
Chapter 4 - Water Acquisition
Average annual water use as a percentage of
fresh water available per Tidwell et al., 2013
(number of counties)
> 100% (17) gggg Data not available (8)
¦¦ 87 -100% (3) I I State reporting requirement
I,— 130 - 87% (15) Hi EIA Shale Basins
HZ) 15 - 30% (10)
H11-15% (36)
IH < 1% (312)
Text Box 4-2 (continued). Hydraulic Fracturing Water Use as a Percentage of Water
Availability Estimates.
(b)
Figure 4-6. Average annual hydraulic fracturing water use in 2011 and 2012 compared to (a) fresh water
available and (b) total water (fresh, brackish, and wastewater) available, by county, expressed as a percentage.
Counties shown with respect to major U.S. EIA shale basins (EIA. 2015). Orange borders identify states that required
some degree of reporting to FracFocus in 2011 and 2012. Data from U.S. EPA (2015c) and Tidwell et al. (2013': data from
Tidwell et al. (2013) supplied from the U.S. Department of Energy (DOE) National Renewable Energy Laboratory on
January 28, 2014 and available upon request from the U.S. DOE Sandia National Laboratories. The analysis by Tidwell et
al. (2013) was done originally for thermoelectric power generation. As such, it was assumed that no fresh water could be
used in California for this purpose due to regulatory restrictions, and therefore no fresh water availability data were
given for California. The total water available for California is the sum of brackish water plus wastewater only.
Average annual water use as
a percentage of total water
available per Tidwell et al.,
2013 (number of counties)
30 - 87% (2)
15- 30% (4)
gijggg Data not available (3)
|HJ EIA Shale Basins
~ State reporting requirement
||§g#jgM:S 2014 Esn
Copyright:^ 2014 Esfi
4-26
-------
Chapter 4- Water Acquisition
Surface water availability is generally low in southern and western Texas (Figure 4-7a), and both
fracturing operations and residents rely heavily on groundwater (Figure 4-7b). Similar to trends
nationally groundwater aquifers in Texas have experienced substantial declines caused by
withdrawals fKonikow. 2013: TWDB. 2012: George etal.. 20111. Groundwater in the Pecos Valley
Gulf Coast, and Ogallala aquifers in southern and western Texas is estimated to have declined by
roughly 5,11, and 44 mi3 (21, 45.5, and 182 km3), respectively, between 1900 and 2008 fKonikow.
2013).1
Runoff (in mm)
100 -200 >1000
200-300 No Dau
300-*00
Ratio of Ground Water Pumping
to Stream Flow ~ Pumping
<0.12 0 59-0 70
0.32 - 0.25 0 70-1.00
| 0.2S-0.37 2% No Data
(b)
Figure 4-7. (a) Estimated annual surface water runoff from the USGS; (b) Reliance on
groundwater as indicated by the ratio of groundwater pumping to stream flow and pumping.
Estimates for Figure 4-7a were calculated at the 8-digit hydrological unit code (HUC) scale by dividing annual
average daily stream flow (from October 1, 2012, to September 30, 2013) by HUC area. Data accessed from the
USGS (USGS, 2014c). Higher ratios (darker blues) in Figure 4-7b indicate greater reliance on groundwater. Figure
adapted from Tidwell et al. (2012), using data provided by the U.S. Department of Energy's Sandia National
Laboratories on December 12, 2014.
1 The estimate of total net volumetric groundwater depletion for the Gulf Coast aquifer is the sum of the individual
depletion estimates for the north (Houston area], central, and southern (Winter Garden area) parts of the Texas Gulf
Coast aquifer. Groundwater depletion from the Carrizo-Wilcox aquifer is included in the estimate for the southern portion
of the Gulf Coast aquifer fKonikow. 20131
4-27
-------
Chapter 4- Water Acquisition
Groundwater quality degradation associated with aquifer pumping and the cumulative effects of all
water users is well documented in the southern portion of the Ogallala aquifer. The quality of
groundwater used by many private, public supply, and irrigation wells is poorest in the aquifer's
southern portion, with elevated concentrations of TDS, chloride, nitrate, fluoride, manganese,
arsenic, and uranium (Chaudhuri and Ale. 2014a: Gurdaketal.. 2009: McMahon etal.. 2007).1
Extensive groundwater pumping can alter the quality of drinking water resources by inducing
vertical mixing of high-quality groundwater with recharge water from the land surface that has
been contaminated by nitrate or pesticides, or with lower-quality groundwater from underlying
geologic formations (Gurdak et al.. 2009: Konikow and Kendv. 2005). Pumping can also promote
changes in reduction-oxidation (redox) conditions and thereby mobilize chemicals from geologic
sources (e.g., uranium) (DeSimone etal.. 2014). Similar patterns of groundwater quality
degradation associated with prolonged aquifer depletion (i.e., salinization and contamination) have
also been observed in other Texas aquifers, notably the northwest Edwards-Trinity (plateau), Pecos
Valley, Carrizo-Wilcox, and southern Gulf Coast aquifers.2
The Texas Water Development Board (TWDB) estimates that overall demand for water (including
water for hydraulic fracturing) out to the year 2060 will outstrip supply in southern and western
Texas (TWDB. 2012). Furthermore, the TWDB expects groundwater supply in the major aquifers to
decline by 30% between 2010 and 2060, mostly due to declines in the Ogallala aquifer (TWDB.
2012.).3 4 Irrigated agriculture is by far the dominant user of water from the Ogallala aquifer
f Gurdak etal.. 20091. but fracturing operations, along with other uses, now contribute to the
aquifer's depletion.
The state has also experienced moderate to extreme drought conditions for much of the last decade,
and the second-worst and longest drought in Texas history between March 2010 and November
2014 fTWDB. 2016: National Drought Mitigation Center. 20151 (Figure 4-8). Sustained drought
conditions compound water availability concerns, and climate change is expected to place further
stress on groundwater both now and in the future (Aghakouchak et al.. 2014: Melillo etal.. 2014)
(Chapter 2). In their evaluation of the potential impact of climate change on groundwater recharge
in the western United States, Meixner et al. (2016) show the largest declines in recharge are
expected in specific aquifers in the southwestern United States, including the southern portion of
the Ogallala aquifer, which is expected to receive 10% less recharge through the year 2050.
1 Elevated levels of these constituents result from both natural processes and human activities, such as groundwater
pumping ("Chaudhuri and Ale. 2014a: Gurdak etal.. 20091
2 Persistent salinity has been observed in west Texas, specifically in the southern Ogallala, northwest Edwards-Trinity
(plateau], and Pecos Valley aquifers, largely due to prolonged irrigational groundwater pumping and ensuing alteration of
hydraulic gradients leading to groundwater mixing ("Chaudhuri and Ale. 2014bI High levels of groundwater salinization
associated with prolonged aquifer depletion have also been documented in the Carrizo-Wilcox and southern Gulf Coast
aquifers, underlying the Eagle Ford Shale in south Texas fChaudhuri and Ale. 2014b: Konikow. 2013: Boghici. 20091
Further, elevated levels of constituents, including nitrate, lead, fluoride, chloride, sulfate, iron, manganese, and TDS, have
been reported in the Carrizo-Wilcox aquifer fBoghici. 20091
3 TWDB (2012) defines groundwater supply as the amount of groundwater that can be produced given current permits
and existing infrastructure. By contrast, TWDB ("20121 defines groundwater availability as the amount of groundwater
that is available regardless of legal or physical availability. Total groundwater availability in Texas is expected to decline
by approximately 24% between 2010 and 2060 fTWDB. 20121.
4 This message is echoed in the 2017 Texas State Water Plan fTWDB. 20161.
4-28
-------
Chapter 4- Water Acquisition
Groundwater moves slowly, and natural recharge rates are lower during times of drought
fDeSimone etal.. 20141. Consequently, as water withdrawals continue to outpace the rate of
recharge, aquifer storage will decline further fUSGS. 19991. potentially impacting both drinking
water resource quantity and quality. For example, research from Steadman etal. f20151 in the
Eagle Ford play shows that hydraulic fracturing groundwater consumption exceeds estimated
recharge rates in the seven most active counties for drilling.
Percentage of
weeks in drought
Copyright:© 2014 Esri
Figure 4-8. Percentage of weeks in drought between 2000 and 2013 by county.
Drought for a given week is defined as any portion of a given U.S. county having a weekly classification of
moderate to exceptional drought (D1-D4 categorization) according to the National Drought Mitigation Center
(http://droughtmonitor.unl.edu); number of weeks = 731.
A case study in the Eagle Ford play in southwestern Texas compared water demand for hydraulic
fracturing with water supplies at the scale of the play, county, and 1 mi2 (2.6 km2) ("Scanlon etal.,
2014b). The authors observed generally adequate water supplies for hydraulic fracturing except in
specific locations, where they found excessive drawdown of groundwater locally in ~6% of the play
area, with estimated declines of ~100-200 ft (31-61 m) after hydraulic fracturing activity increased
in 2009 (Text Box 4-3).
4-29
-------
Chapter 4- Water Acquisition
Text Box 4-3. Case Study: Water Profile of the Eagle Ford Play, Texas.
Researchers from the University of Texas published a detailed case study of water supply and demand for
hydraulic fracturing in the Eagle Ford play in southwestern Texas (Scanlon et al.. 2014bl. This effort
assembled detailed information from state and local water authorities, and proprietary industry data on
hydraulic fracturing, to develop a portrait of water resources in this 16-county area.
Scanlon etal. f2014b1 compared water demand for hydraulic fracturing currently and over the projected play
life (20 years) relative to water supply from groundwater recharge, groundwater storage (brackish and
fresh), and stream flow. Using groundwater availability models developed by the Texas Water Development
Board, they reported that water demand for hydraulic fracturing in 2013 was 30% of annual groundwater
recharge in the play area, and over the 20-year play lifespan it was projected to be 26% of groundwater
recharge, 5-8% of fresh groundwater storage, and 1% of brackish groundwater storage. The dominant water
user in the play is irrigation (57 to 61% of water use, 62 to 65% of consumption), as compared with hydraulic
fracturing (13% of water use and 16% of consumption). At the county level, projected water demand for
hydraulic fracturing over the 20-year period was low relative to freshwater supply (ranging from 0.6-27% by
county, with an average of 7.3%). Similarly, projected total water demand from all uses was low relative to
supply, excluding two counties with high irrigation demands (Frio, Zavala), and one county with no known
groundwater supplies (Maverick).
Although supply was found to be sufficient even in this semi-arid region, there were important exceptions,
especially at sub-county scales. The researchers found no water level declines over much of the play area
assessed (69% of the play area), yet in some areas they estimated groundwater drawdowns of 50 ft (15 m) or
more (19% of the play area), 100 ft (31 m) or more (6% of the play area), and 200 ft (60 m) or more
(approximately 2% of the play area). This was corroborated with well monitoring data that showed a sharp
decline in water levels in several groundwater monitoring wells after hydraulic fracturing activity increased
in 2009.
The researchers further concluded that shifting toward brackish groundwater is feasible, as evidenced by
operators already doing so. This shift could further reduce impacts on fresh water resources and provide a
large source of water for future hydraulic fracturing. In a 2011 estimate, approximately 20% of water used in
the play came from brackish sources (Table 4-4), and anecdotal evidence suggests this practice has increased
since then (Scanlon et al.. 2014b). Projected hydraulic fracturing water use represents less than 1% of total
brackish groundwater storage in the play area. By contrast, Scanlon et al. (2014b) concluded there is limited
potential for reuse of wastewater in this play because of the small volumes that return to the surface during
production (less than or equal to 5% of hydraulic fracturing water requirements).
In contrast to southern and western Texas, the potential for water quantity and quality effects
appears to be lower in the north-central and eastern parts of the state, in areas including the
Barnett and Haynesville plays. Residents obtain water for domestic use—which includes use of
water for drinking—from a mixture of groundwater and surface water sources (Appendix Table B-
6). Counties encompassing Dallas and Fort Worth rely mostly on publicly-supplied surface water
fTWDB. 2012) (Appendix Table B-6). The Trinity aquifer in northeast Texas is projected to decline
only slightly between 2010 and 2060 fTWDB. 2012). Nevertheless, Bene etal. (2007) estimate that
hydraulic fracturing groundwater withdrawals will increase from 3% of total groundwater use in
2005 to 7%-13% in 2025, suggesting the potential for localized aquifer drawdown. Groundwater
quality degradation associated with aquifer drawdown has been documented in the Trinity and
Woodbine aquifers overlying much of the Barnett play, with both aquifers showing high levels of
salinization (Chaudhuri and Ale. 2013).
4-30
-------
Chapter 4- Water Acquisition
Overall, the potential for impacts appears higher in western and southern Texas, compared to the
northeast part of the state. Groundwater withdrawals for hydraulic fracturing, along with irrigation
and other uses, may contribute to water quality degradation associated with intensive aquifer
pumping in western and southern Texas. Areas with numerous high-capacity wells and large
amounts of sustained groundwater pumping are most likely to experience groundwater quality
degradation associated with withdrawals (Gurdak et al.. 2009: McMahon et al.. 20071. Further,
given that Texas is prone to drought conditions and groundwater recharge is limited, the already
declining aquifers in southern and western Texas are especially vulnerable to further groundwater
depletion and resulting impacts to groundwater quantity and quality (Gurdak etal.. 2009: Jackson
et al.. 2001). Impacts are likely to be localized drawdowns of groundwater, as shown by a detailed
case study of the Eagle Ford play (Text Box 4-3}. Scanlon etal. f2014b1 suggested that a shift
toward brackish water use could minimize potential future impacts to fresh water resources. This
finding is consistent with our county level data (Text Box 4-2).
4.5.2 Colorado and Wyoming
Colorado had the second highest number of disclosures in the EPA FracFocus 1.0 project database,
(13% of disclosures) (Figure 4-4 and Appendix Table B-5). We combine Colorado and Wyoming
because of their shared geology of the Denver Basin (including the Niobrara play) and the Greater
Green River Basin (Figure 4-9). There are three major basins reported for Colorado: the Denver
Basin; the Uinta-Piceance Basin; and the Raton Basin. Together these basins contain 99% of
reported wells in the state, although the bulk of the activity in Colorado is in the Denver Basin
(Appendix Table B-5). Fewer wells (roughly 4% of disclosures in the EPA FracFocus 1.0 project
database) are reported in Wyoming. There are two major basins reported for Wyoming (Greater
Green River and Powder River) that together contain 86% of activity in the state (Appendix Table
B-5).
Powder
River
Greater Green
River
Niobrara
Play
Uinta-Pi
Denver
Figure 4-9. Major U.S. EIA shale plays and basins for Colorado and Wyoming.
Source: EiA (2015).
4-31
-------
Chapter 4- Water Acquisition
Types of water used: Water for hydraulic fracturing in Colorado and Wyoming comes from both
groundwater and surface water, as well as reused wastewater f Colorado Division of Water
Resources etal.. 2014: BLM. 20131. Publicly available information on water sources for each state
generally comes in the form of a list of potential sources, and detailed information on the types of
water used for hydraulic fracturing is not readily accessible.1 In northwestern Colorado's Garfield
County (Uinta-Piceance Basin), the U.S. EPA (2015e) reports that any fresh water used for
fracturing comes from surface water sources. In the Denver Basin (Niobrara play) of southeastern
Wyoming, qualitative information suggests that groundwater supplies much of the water used for
fracturing, although no data were available to characterize the ratio of groundwater to surface
water withdrawals (AMEC Environment & Infrastructure. 2014: BLM. 2013: Tyrrell. 2012).
Non-fresh water sources, including industrial and municipal wastewater, brackish groundwater,
and reused hydraulic fracturing wastewater, are sometimes listed as potential alternatives to fresh
water for fracturing in both Colorado and Wyoming (Colorado Division of Water Resources et al..
2014: BLM. 2013): no data are available to show the extent to which these non-fresh water sources
are used at the state or basin level. Based on discussions with industry, the U.S. EPA f2015el
reports that fresh water is used solely for drilling and reused wastewater supplies nearly all the
water for hydraulic fracturing in Colorado's Garfield County. This estimate of reused wastewater as
a percentage of injected volume is markedly higher than in other locations and likely results from
the geologic characteristics of the Piceance tight sand formation, which has naturally high water
content and produces large volumes of relatively high-quality wastewater fU.S. EPA. 2015el
In contrast, a study by Goodwin et al. (2014) assumed no reuse of wastewater for hydraulic
fracturing operations by Noble Energy in the Denver-Julesburg Basin of northeastern Colorado
(Table 4-2). It is unclear whether this assumption is indicative of reuse practices of other
companies in the Denver-Julesburg Basin. The difference in reused wastewater rates reported by
the U.S. EPA (2015e) and Goodwin etal. (2014) may indicate an east-west divide in Colorado (i.e.,
low reuse in the east versus high reuse in the west), due at least in part to differences in wastewater
volumes available for reuse. However, further information is needed to adequately characterize
reuse patterns in Colorado.
Water use per well: Water use per well varies across Colorado, with median values of 1.8 million,
400,000, and 96,000 gal (6.8 million, 1.5 million, and 360,000 L) in the Uinta-Piceance, Denver, and
Raton Basins, respectively, according to the EPA FracFocus 1.0 project database (Appendix Table B-
5). Relatively low water volumes per well are reported in Wyoming (Appendix Table B-5). Low
volumes reported for the Raton Basin of Colorado and the Powder River Basin of Wyoming are
likely due to the prevalence of CBM extraction in these locations (U.S. EPA. 2015k: Sando etal..
2014).
More difficult to explain are the low volumes reported for the Denver Basin in the EPA FracFocus
1.0 project database. These values are lower than volumes reported in other non-CBM basins
1 The Colorado Oil and Gas Conservation Commission collects information on the sources and quality of water used for
hydraulic fracturing, including reused wastewater, with Form 5A, and has done so since June 2012; however, these data
are in PDFs linked to individual wells and are not aggregated into a searchable database.
4-32
-------
Chapter 4- Water Acquisition
included in Appendix Table B-5. Goodwin etal. (20141 report much higher water use per well in the
Denver Basin from 2010 to 2013, with a median of 2.8 million gal (10.6 million L) (although only
usage for the Wattenberg Field was reported). Indeed, the 10th-90th percentiles (2.4-3.8 million gal)
(9.1-14.4 million L) from Goodwin etal. T20141 are almost completely above those from the EPA
FracFocus 1.0 project database for the Denver Basin (Appendix Table B-5).1 However, it is difficult
to draw clear conclusions because of differences in scale (i.e., field in Goodwin et al. (2014) versus
basin in the EPA FracFocus 1.0 project database) and operators (i.e., Noble Energy in Goodwin et al.
(2014) versus all in the EPA FracFocus 1.0 project database).
Trends in water use per well are generally lacking for Colorado, with the exception of those
reported by Goodwin etal. f2014I They found that water use per well is increasing with well
length in the Denver Basin; however, they also observed that water intensity (gallons of water per
unit energy extracted) did not change, since energy recovery increased along with water use.
Water use/consumption at the county scale: Hydraulic fracturing operations in Colorado use billions
of gallons of water, but this amount is a small percentage compared to total water used or
consumed at the county scale. In both Garfield and Weld Counties, located in the Uinta-Piceance and
Denver Basins, respectively, hydraulic fracturing used more than 1 billion gal (3.8 billion L)
annually. Fracturing water use and consumption in these counties exceeded those in all other
Colorado counties combined (Appendix Table B-2), but the water used for hydraulic fracturing in
Garfield and Weld counties was less than 2% and 3% compared to 2010 total water use and
consumption, respectively. In comparison, irrigated agriculture accounts for over 90% of the water
used in both counties (Maupin etal.. 2014). Overall, hydraulic fracturing accounts for less than 2%
compared to 2010 total water use in all Colorado counties represented in the EPA FracFocus 1.0
project database (Appendix Table B-2). Water use estimates based on the EPA FracFocus 1.0
project database may be low relative to literature and state estimates (Text Box 4-1), but even if
estimates from the project database were doubled, hydraulic fracturing water use and consumption
would still be less than 4% and 6% compared to 2010 total water use and consumption,
respectively, in each Colorado county.
In Wyoming, reported water use for hydraulic fracturing is small compared to Colorado (Appendix
Table B-l). Fracturing water use and consumption did not exceed 1% of 2010 total water use and
consumption, respectively, in any county (Appendix Table B-2). Unlike Colorado, Wyoming did not
require disclosure to FracFocus during the time period analyzed by the EPA fU.S. EPA. 2015bl
(Appendix Table B-5).
Colorado Division of Water Resources etal. (2014) projected that annual water use for hydraulic
fracturing in the state would increase by approximately 16% between 2012 and 2015, but demand
in later years is unclear. Even with an increase of 16% or more, hydraulic fracturing would still
remain a relatively small user of water at the county scale in Colorado.
1 Different spatial extents might explain these differences, since Goodwin et al. ("20141 focus on 200 wells in the
Wattenberg Field of the Denver Basin; however. Weld County is the center of activity in the Wattenberg Field, and the EPA
FracFocus 1.0 project database contains 3,011 disclosures reported in Weld County, with a median water use per of
407,442 gal (1,542,340 L], similar to that for the basin as a whole.
4-33
-------
Chapter 4- Water Acquisition
Potential for impacts: The potential for water quantity and quality impacts due to hydraulic
fracturing water withdrawals appears to be low at the county scale in Colorado and Wyoming
because fracturing accounts for a low percentage of total water use and consumption (Figure
4-3a,b). This conclusion is also supported by the comparison of hydraulic fracturing water use to
water availability at the county scale (Text Box 4-2; Figure 4-6a,b). However, counties in Colorado
and Wyoming are large in their spatial extents, and any potential impacts will depend on site-
specific factors affecting the balance between water use and availability at the local scale (i.e., at a
given withdrawal point). In a multi-scale case study in the Upper Colorado River Basin, the U.S. EPA
(2015e) did not identify any locations where fracturing currently contributed to locally high water
use intensity due to the high rates of wastewater reuse reported. They did conclude, however, that
future effects may be possible (Text Box 4-4).
Text Box 4-4. Case Study: Impact of Water Acquisition for Hydraulic Fracturing on Local Water
Availability in the Upper Colorado River Basin.
The U.S. EPA f~2015e1 conducted a case study to explore the impact of hydraulic fracturing water demand on
water availability at the river basin, county, and local scales in the semi-arid Upper Colorado River Basin
(UCRB) of western Colorado. The study area overlies the Piceance geologic basin with natural gas in tight
sands. Water withdrawal impacts were quantified using a water use intensity index (i.e., the ratio between
the volume of water withdrawn at a site for hydraulic fracturing and the volume of available water).
Researchers obtained detailed site-specific data on hydraulic fracturing water usage from state and regional
authorities, and estimated available water supplies using observations at USGS gage stations and empirical
and hydrologic modeling.
They found that water supplies accessed for oil and gas demand were concentrated in Garfield County, and
most fresh water withdrawals were concentrated within the Parachute Creek watershed (198 mi2). However,
fresh water makes up a small proportion of the total water used for fracturing due to large quantities of high-
quality wastewater produced from the Piceance tight sands. Based on discussions with industry, the U.S. EPA
(2015e) reports that fresh water is used solely for drilling and reused wastewater supplies nearly all the
water for hydraulic fracturing in Garfield County. Due to the high reuse rate, the U.S. EPA (2015e) did not
identify any locations in the Piceance play where fracturing contributed to locally high water use intensity.
Scenario analyses demonstrated a pattern of increasing potential impact with decreasing watershed size in
the UCRB. The U.S. EPA (2015e) examined hydraulic fracturing water use intensity under the current rates of
both directional (S-shaped) and horizontal drilling. They showed that for the more water-intensive horizontal
drilling, watersheds had to be larger to meet the same index of water use intensity (0.4) as that for directional
drilling (100 mi2 for horizontal drilling, as compared to 30 mi2 for directional drilling). To date, most wells
have been drilled directionally into the Piceance tight sands, although a trend toward horizontal drilling is
expected to increase annual water use per well by about four times. Despite this increase, total hydraulic
fracturing water use is expected to remain small relative to other users. Currently, irrigated agriculture is the
largest water user in the UCRB.
Greater water demand could occur in the future if the water-intensive oil shale extraction industry becomes
economically viable in the region. Projections for oil shale water demand indicate that the industry could
increase water use for energy extraction in Garfield and Rio Blanco counties.
4-34
-------
Chapter 4- Water Acquisition
East of the Rocky Mountains in the Denver Basin, the potential for localized impacts exists given the
combination of high hydraulic fracturing activity and low water availability (e.g., Weld County,
Colorado), but lack of available data and literature at the local scale limits our ability to assess the
potential for impacts in this location. Ceres f20'141 concludes that all fractured wells in the Denver
Basin are in high or extremely high water-stressed areas. Furthermore, the development of the
Niobrara Shale in southeast Wyoming occurs in areas already impacted by high agricultural water
use from the Ogallala aquifer, including the state's only three groundwater control areas, which
were established as management districts in the southeast portion of the state in response to
declining groundwater levels (AMEC Environment & Infrastructure. 2014: Wyoming State
Engineer's Office. 2014: Tyrrell. 2012: Bartos and Hallberg. 20111. Groundwater withdrawals for
hydraulic fracturing may have the potential to contribute to water quality degradation in these
areas, depending on site-specific factors that may alter the balance between water use and
availability.
Overall, the potential for impacts appears low at the county scale in Colorado and Wyoming but
local effects are certainly possible particularly east of the Rocky Mountains in the Denver Basin.
Lack of available data and literature at the local scale limits our ability to assess the potential for
impacts in this location.
4.5.3 Pennsylvania, West Virginia, and Ohio
Pennsylvania had the third most disclosures in the EPA FracFocus 1.0 project database (6.5% of
disclosures) (Appendix Table B-5; Figure 4-4). We combine West Virginia and Ohio with
Pennsylvania because they share similar geology overlying the Appalachian Basin (including the
Marcellus, Devonian, and Utica stacked plays) (Figure 4-10); however, much less activity is
reported in these two states (Appendix Table B-5).
Marcellus, Devonian,
and Utica Plays
Appalachian
Figure 4-10. Major U.S. EIA shale plays and basins for Pennsylvania, West Virginia, and Ohio.
Source: EIA (2015).
4-35
-------
Chapter 4- Water Acquisition
Types of water used-. Surface water is the primary water source for hydraulic fracturing in
Pennsylvania, West Virginia, and Ohio fSRBC. 2016: Schmid and Yoxtheimer. 2015: West Virginia
DEP. 2014: Mitchell et al.. 2013a: West Virginia DEP. 2013: Ohio EPA. 2012b: STRONGER. 2011 hi
(Table 4-1). Further, the water used for hydraulic fracturing is most often fresh water in all three
states. In both Pennsylvania's Susquehanna River Basin and throughout West Virginia, most water
for hydraulic fracturing is self-supplied via direct withdrawals from surface water and groundwater
fU.S. EPA. 2015e: West Virginia DEP. 20131. Operators also purchase water from public water
systems, which may include a variety of commercial water brokers fWest Virginia DEP. 2014: SRBC.
2013: West Virginia DEP. 2013). Municipal supplies are also used, particularly in urban areas of
Ohio fSTRONGER. 2011bl.
Reused hydraulic fracturing wastewater as a percentage of total water used for fracturing was 19%
in 2014 in Pennsylvania, and 15% in 2012 in West Virginia f Schmid and Yoxtheimer. 2015: West
Virginia DEP. 2014) (Table 4-2). Available data indicate an increasing trend in reuse of wastewater
over time in this region, likely due to the lack of nearby disposal options in Class II wells. Reused
wastewater as a percentage of injected water volume ranged from approximately 2% to 19% in
Pennsylvania (statewide) from 2009-2014 fSchmid and Yoxtheimer. 20151. This upward trend is
also shown in Pennsylvania's SRB, where reuse as a percentage of total water injected reached 22%
in 2013; the average reuse rate for 2008-2013 in the SRB was 16% (SRBC. 2016) (Table 4-2). In
West Virginia, reuse as a percentage of injected volume ranged from 6% to 15% from 2010-2012
fWest Virginia DEP. 20141. In Ohio's Marcellus and Utica Shales, reuse of wastewater is reportedly
uncommon f STRONGER. 2011bl. likely due to the prevalence of disposal wells in Ohio. See Chapter
8 for more information.
Aside from reused hydraulic fracturing wastewater, other types of wastewaters reused for
hydraulic fracturing may include wastewater treatment plant effluent, treated acid mine drainage,
and rainwater collected at various well pads (West Virginia DEP. 2014: SRBC. 2013: West Virginia
DEP. 2013: Ziemkiewicz etal.. 2013: Ohio EPA. 2012b). No data are available on the frequency of
use of these other wastewaters.
Water use per well: Operators in these three states reported the third, fourth, and fifth highest
median water use per well of the states we considered from the EPA FracFocus 1.0 project
database, with 5.0, 4.2, and 3.9 million gal (18.9,15.9, and 14.8 million L) in West Virginia,
Pennsylvania, and Ohio, respectively (Appendix Table B-5). Hansen etal. f20131 report similar
water use estimates for Pennsylvania and West Virginia for 2011 and 2012 (Appendix Table B-5).
This correspondence is not surprising, as these estimates are also based on FracFocus data (via
Skytruth). For 2011, the year overlapping with the time frame of the EPA FracFocus report (U.S.
EPA. 2015b). Mitchell etal. (2013a) report an average of 2.3 million gal (8.7 million L) for vertical
wells (54 wells) and 4.6 million gal (17.4 million L) for horizontal wells (612 wells) in the
Pennsylvania portion of the Upper Ohio River Basin, based on records from PA DEP. The weighted
average water use per well was 4.4 million gal (16.7 million L), similar to results based on the EPA
FracFocus 1.0 project database listed above. In Pennsylvania's SRB, the long-term average water
use per well from 2008-2013 was 4.3 million gal (16.3 million L). In 2013, the average water use
per well increased to approximately 5.1 to 6.5 million gal (19.3 to 24.6 million L) due to increasing
4-36
-------
Chapter 4- Water Acquisition
lengths of laterals in horizontal drilling fSRBC. 20161. Across the entire state of Pennsylvania, water
use per well has increased over time, which may be explained by increasing horizontal well length,
depth, and length of the completed interval (Schmid and Yoxtheimer. 2015).
Water use/consumption at the county scale: In this tri-state region, the highest water use for
hydraulic fracturing is in northeastern Pennsylvania counties. On average, operators in Bradford
County reported over 1 billion gal (3.8 billion L) used annually in 2011 and 2012 for fracturing;
operators in three other counties (Susquehanna, Lycoming, and Tioga Counties) reported
500 million gal (1.9 billion L) or more used annually in each county (Table 4-3). On average,
hydraulic fracturing water use is 3.2% compared to 2010 total water use for counties with
disclosures in the EPA FracFocus 1.0 project database in these three states (Table 4-3;
Appendix Table B-2). Susquehanna County in Pennsylvania has the highest percentages relative to
2010 total water use (47%) and consumption (123%).
Potential for impacts: Water availability is higher in Pennsylvania, West Virginia, and Ohio than in
many western states, reducing the likelihood of impacts to drinking water resource quantity and
quality. At the county scale, water supplies appear adequate to accommodate this use (Text Box
4-2; Figure 4-6a,b). However, impacts could still occur at the local scale (i.e., specific withdrawal
points) as high water availability in a region does not preclude water stress, particularly if water
withdrawals occur during seasonal low-flow periods (Entrekin etal.. 2015). Without management
of the rate and timing of withdrawals, surface water withdrawals for hydraulic fracturing have the
potential to affect both drinking water quantity and quality fMitchell etal.. 2013al. For instance,
withdrawals may alter natural stream flow regimes, potentially decreasing a stream's capacity to
dilute contaminants (Gallegos etal.. 2015: Mitchell etal.. 2013a: Entrekin etal.. 2011: NYSDEC.
2011: van Vliet and Zwolsman. 2008: IPCC. 2007: Environment Canada. 2004: Murdoch etal..
20001.
In a second, multi-scale case study, EPA showed that the potential for water acquisition impacts to
drinking water resource quantity and quality increases at finer temporal and spatial resolutions
fU.S. EPA. 2015el. They concluded that individual streams in Pennsylvania's SRB can be vulnerable
to typical hydraulic fracturing water withdrawals depending on stream size, as defined by
contributing basin area (U.S. EPA. 2015e) (Text Box 4-5). They observed infrequent (in less than
1% of withdrawals) high ratios of hydraulic fracturing water consumption to stream flow (high
consumption-to-stream flow events). Further research from Barth-Naftilan etal. f20151 in
Pennsylvania's Marcellus Shale (SRB and Ohio River Basin (ORB)) confirmed that stream flow
alteration due to hydraulic fracturing surface water withdrawals increases at finer spatial scales
(i.e., smaller watershed area). They showed that streams with drainage areas under 50 mi2 (130
km2) are the most vulnerable to stress induced by flow alteration (Barth-Naftilan etal.. 20151.
4-37
-------
Chapter 4- Water Acquisition
Text Box 4-5. Case Study: Impact of Water Acquisition for Hydraulic Fracturing on Local Water
Availability in the Susquehanna River Basin.
The U.S. EPA (2015e) conducted a second case study analogous to that in the UCRB (Text Box 4-4), to explore
the impact of hydraulic fracturing water demand on water availability at the river basin, county, and local
scales in the SRB in northeastern Pennsylvania. The study area overlies the Marcellus Shale gas reservoir.
Water withdrawal impacts were quantified using a water use intensity index (Text Box 4-4). Researchers
obtained detailed site-specific data on hydraulic fracturing water usage from state and regional authorities,
and estimated available water supplies using observations at USGS gage stations and empirical and
hydrologic modeling.
Most water for fracturing in the SRB is self-supplied by operators from rivers and streams with withdrawal
points distributed throughout a wide geographic area. Public water systems provide a relatively small
proportion of the water needed. Reuse of wastewater as a percentage of hydraulic fracturing fluid volume
averaged 16% from 2008-2013, and has increased over time, reaching 22% in 2013 (SRBC. 2016) (Table
4-2). The Susquehanna River Basin Commission (SRBC) regulates water acquisition for hydraulic fracturing
and issues permits that set limits on the volume, rate, and timing of withdrawals at individual withdrawal
points; passby flow thresholds (hereafter, passby flows) halt water withdrawals during low flows.
The U.S. EPA (2015e) demonstrated that streams can be vulnerable from hydraulic fracturing water
withdrawals depending on their size, as defined by contributing basin area. Small streams have the potential
for impacts (i.e., high water use intensity) for all or most of the year. The U.S. EPA (2015e) showed an
increased likelihood of impacts in small watersheds in the SRB (less than 10 mi2 or 26 km2). Furthermore,
they showed that in the absence of passby flows, even larger watersheds (up to 600 mi2 or 1,554 km2) could
be vulnerable during maximum withdrawal volumes and infrequent droughts. However, high water use
intensity calculated from observed hydraulic fracturing withdrawals occurred at only a few withdrawal
locations in small streams; local high water use intensity was not found at the majority of withdrawal points.
Detailed studies and state reports available throughout the Marcellus Shale region help provide an
understanding of the potential impacts of hydraulic fracturing water withdrawals in both space and
time at the local scale (SRBC. 2016: Barth-Naftilan etal.. 2015: U.S. EPA. 2015e). In the SRB and
ORB, water for hydraulic fracturing is taken from both large rivers and small headwater streams,
with a considerable fraction of the water taken from small streams of small watersheds (Barth-
Naftilan etal.. 2015). The SRBC reports that most natural gas development in the SRB is focused in
rural, headwater areas, where withdrawals have the potential to alter natural stream flow regimes
(SRBC. 2016). In an analysis of the effects of water withdrawals on twelve streams in the SRB,
Shank and Stauffer (2015) found that the largest withdrawals relative to stream size were from
headwater streams, where daily withdrawals averaged 6.8% of average daily flows. However, they
found water management in the form of low flow protections helped limit the potential for impacts.
Compared to conventional energy extraction, hydraulic fracturing consumes more water in a highly
concentrated period of time (Patterson etal.. 2016): thus, the cumulative impact of multiple wells
withdrawing water from small streams, particularly during drought or seasonal low flows, has the
potential to impact the quantity and quality of drinking water resources (Patterson etal.. 2016). For
instance, in modeling the potential future impact of hydraulic fracturing in the Delaware River
Basin (DRB), Habichtetal. (2015) showed that under maximum well development, hydraulic
fracturing water withdrawals from small streams could remove up to 70% of water during periods
4-38
-------
Chapter 4- Water Acquisition
of low stream flow, and less than 3% during periods of normal stream flow.1 Unlike groundwater
withdrawals, any impacts to drinking water resource quantity and quality associated with surface
water withdrawals are likely to persist for a shorter time period since the rate of replenishing
water removed from the system is greater in surface water than groundwater fAlley etal.. 19991
(Section 4.5.1).
The potential for water acquisition impacts to drinking water resource quality in this region is also
greatest in small, unregulated streams, particularly under drought conditions or during seasonal
low flows (U.S. EPA. 2015e: Vengosh etal.. 2014: Mitchell etal.. 2013a: Vidic etal.. 2013: Rahm and
Riha. 2012: Rolls etal.. 2012: Kargbo etal.. 2010: McKay and King. 20061. Surface water quality
impacts may be of concern if a pollution discharge point (e.g., sewage treatment plant, agricultural
runoff, or chemical spill) is immediately downstream of a hydraulic fracturing withdrawal point
fU.S. EPA. 2015e: NYSDEC. 20111.2 Potential water quality impacts associated with reduced water
levels may also include possible interference with the efficiency of drinking water treatment plant
operations, as increased contaminant concentrations in drinking water sources may necessitate
additional treatment and ultimately impact drinking water quality fWater Research Foundation.
2014: Benotti etal.. 2010).3
Water management policies in place in this region can help reduce the potential for impacts
associated with hydraulic fracturing water withdrawals, including excessive lowering of water
levels, unreliable water supplies, and degradation of water quality fSRBC. 2016: Barth-Naftilan et
al.. 2015: U.S. EPA. 2015el (Text Box 4-5). For instance, the SRBC manages the quantity, location,
and timing of withdrawals, using site-specific information to set instantaneous and daily
withdrawal limits for all approved surface water and groundwater withdrawals. They also set low
flow protections, known as passby flows, for most approved surface water withdrawals that require
withdrawals to cease when stream flow drops below a prescribed threshold level fSRBC. 20161.
Passby flows can reduce the frequency of high consumption-to-stream flow events, particularly in
the smallest streams (Shank and Stauffer. 2015: U.S. EPA. 2015e).
Overall, there appears to be adequate surface water for hydraulic fracturing in Pennsylvania, West
Virginia, and Ohio, but there is still the potential for impacts to both drinking water resource
quantity and quality, particularly in small streams, if the rate and timing of withdrawals are not
managed (U.S. EPA. 2015e). These potential impacts are expected to be localized in space (i.e.,
1 Presently there is a moratorium on hydraulic fracturing in the DRB, which spans Pennsylvania, Delaware, New Jersey,
and New York. Habichtetal. ("20151 modeled the potential future environmental impact of hydraulic fracturing in the DRB
should the moratorium be lifted, allowing hydraulic fracturing to expand into this region in the future.
2 Aside from direct surface water withdrawals, unmanaged withdrawals from public water systems can cause cross-
contamination if there is a loss of pressure, allowing the backflow of pollutants from tank trucks into the distribution
system. The state of Ohio has issued a fact sheet relevant to this potential concern, intended specifically for public water
systems providing water to oil and gas companies ("Ohio EPA. 2012al. To prevent potential cross-contamination, Ohio
requires a backflow prevention device at cross-connections. For example, bulk loading stations that provide public supply
water directly to tank trucks are required to have an air-gap device at the cross-connection to prevent the backflow of
contaminants into the public water system ("Ohio EPA. 2012al.
3 For instance, an increased proportion of organic matter entering a treatment plant may increase the formation of
trihalomethanes, byproducts of the disinfection process formed as chlorine reacts with organic matter in the water being
treated fWater Research Foundation. 20141.
4-39
-------
Chapter 4 - Water Acquisition
occurring at specific withdrawal points], and time (e.g., low flow periods). Passby flows appear to
be an effective water management tool for reducing the potential for impacts from surface water
withdrawals.
4.5.4 North Dakota and Montana
North Dakota was fourth in the number of disclosures in the EPA FracFocus 1.0 project database
(5.9% of disclosures] (Appendix Table B-5; Figure 4-4). We combine Montana with North Dakota,
because both overlie the Williston Basin (which contains the Bakken play, shown in Figure 4-11),
although many fewer wells are reported for Montana (Appendix Table B-5). The Williston Basin is
the only basin with significant activity reported for either state, though other basins are also
present in Montana (e.g., the Powder River Basin).
Figure 4-11. Major U.S. EIA shale plays and basins for North Dakota and Montana.
Source: EIA (20151
Types of water used: Hydraulic fracturing in the Bakken play depends on both ground and surface
water resources. Surface water from the Missouri River system provides the largest source of fresh
water in the center of Bakken oil development (North Dakota State Water Commission. 2014: EERC.
2011. 2010: North Dakota State Water Commission. 20101. Apart from the Missouri River system,
regional surface waters (e.g., smaller streams) do not provide a consistent supply of water for the
oil industry due to seasonal stream flow variations. Sufficient stream flows generally occur only in
the spring after snowmelt f EERC. 20111 Groundwater from glacial and bedrock aquifer systems
has traditionally supplied much of the water needed for Bakken development, but concerns over
limited groundwater supplies have led to limits on the number of new groundwater withdrawal
permits issued fCeres. 2014: Plummer etal.. 2013: EERC. 2011. 2010: North Dakota State Water
Commission. 2010).
The water used for Bakken development is mostly fresh. The EPA FracFocus report shows that
"fresh" was the only source of water listed in almost all disclosures reporting a source of water in
North Dakota (U.S. EPA. 2015b).1 Reuse of Bakken wastewater is limited due to its high TDS, which
1 Twenty-five percent of North Dakota disclosures included information related to water sources ("U.S. EPA. 2015bl.
Montana
Thrust Belt
Bakken
Play
Williston
Powder
River
Bighorn
EIA Basins
4-40
-------
Chapter 4- Water Acquisition
presents challenges for treatment and reuse (Gadhamshetty etal.. 20151. Industry is currently
researching treatment technologies for reuse of this wastewater fCeres. 2014: EERC. 2013. 20111.
Water for hydraulic fracturing is commonly purchased from municipalities or other public water
systems in the region. The water is often delivered to trucks at water depots or transported directly
to well pads via pipelines fEERC. 20111.
Water use per welt Water use per well is intermediate compared with other areas, with a median of
2.0 and 1.6 million gal (7.6 and 6.1 million L) per well in the Williston Basin in North Dakota and
Montana, respectively, according to the EPA's FracFocus 1.0 project database (Appendix Table B-5).
The North Dakota State Water Commission reports similar volumes (2.2 million gal (8.3 million L)
per well on average for North Dakota) in a summary fact sheet fNorth Dakota State Water
Commission. 2014).1 Scanlon et al. (2016) show that average water use per well in the Bakken play
has increased over time, from 580,000 gal (2.2 million L) in 2005 to 3.7 million gal (14.1 million L)
in 2014, due in part to the increasing lengths of laterals in horizontal drilling.
In addition to water for hydraulic fracturing, Bakken wells may require "maintenance water"
(Scanlon et al.. 2016: Scanlon etal.. 2014a). This extra water is reportedly needed because of the
relatively high salt content of Bakken brine, potentially leading to salt buildup, pumping problems,
and restriction of oil flow. Based on estimates from the North Dakota Department of Mineral
Resources, Scanlon etal. f 20161 report that approximately 400 - 600 gal (1,500 - 2,300 L) per day
per each well may be required for well maintenance. Assuming a 15-year lifetime for wells, this
could add up to 3.3 million gal (12.5 million L) per well of additional water (Scanlon etal.. 2016).
Water use/consumption at the county scale: Water use for fracturing in this region is greatest in the
northwestern corner of North Dakota (Gadhamshettv etal.. 2015). Hydraulic fracturing water use
in 2011 and 2012 averaged approximately 123 million gal (466 million L) per county in the two-
state area, with use in McKenzie and Williams Counties in North Dakota exceeding 500 million gal
(1.9 billion L) (Appendix Table B-2). There were four counties where 2011 and 2012 average
hydraulic fracturing water use was 10% or more of 2010 total water use. Mountrail and Dunn
Counties showed the highest percentages (36% and 29%, respectively). Outside of North Dakota's
northwest corner, hydraulic fracturing used much less water in the rest of the state and Montana
(Table 4-3; Appendix Table B-2).
Potential for impacts: In this region, there are concerns about over-pumping groundwater
resources, but the potential for impacts appears to be low provided the Missouri River is
determined to be a sustainable and usable source. This finding of a low potential for impacts is also
supported by the comparison of hydraulic fracturing water use to water availability at the county
scale (Text Box 4-2; Figure 4-6a,b). This area is primarily rural, interspersed with small towns.
Residents rely on a mixture of surface water and groundwater for domestic use depending on the
county, with most water supplied by local municipalities (Appendix Table B-6).
1 The fact sheet is a stand-alone piece, and it is not accompanied by an underlying report.
4-41
-------
Chapter 4- Water Acquisition
The state of North Dakota and the U.S. Army Corps of Engineers concluded that groundwater
resources in western North Dakota are not sufficient to meet the needs of the oil and gas industry
fU.S. Army Corps of Engineers. 2011: North Dakota State Water Commission. 20101. All users
combined currently withdraw approximately 6.2 billion gal (23.5 billion L) of water annually in an
11-county region in western North Dakota, already stressing groundwater supplies (U.S. Army
Corps of Engineers. 20111. By comparison, the total needs of the oil and gas industry are projected
to range from approximately 2.2 and 8.8 billion gal (8.3 and 33.3 billion L) annually by the year
2020 (U.S. Army Corps of Engineers. 2011).
Due to concerns for already stressed groundwater supplies, the state of North Dakota limits
industrial groundwater withdrawals, particularly from the Fox Hills-Hell Creek aquifer f Ceres.
2014: Plummer etal.. 2013: EERC. 2011. 2010: North Dakota State Water Commission. 20101.
Currently, the oil industry is the largest industrial user of water from the Fox Hills-Hell Creek
aquifer (North Dakota State Water Commission. 2010). Many farms, ranches, and some
communities in western North Dakota rely on flowing wells from this artesian aquifer, particularly
in remote areas that lack electricity for pumping; however, low recharge rates and withdrawals
throughout the last century have resulted in steady declines in the formation's hydraulic pressure
(North Dakota State Water Commission. 2010). Declines in hydraulic pressure do not appear to be
associated with impacts to groundwater quality; rather, the state is concerned with maintaining
flows for users fNorth Dakota State Water Commission. 20101.
To reduce demand for groundwater, the state is encouraging the industry to seek surface water
withdrawals from the Missouri River system. The North Dakota State Water Commission concluded
the Missouri River and its dammed reservoir, Lake Sakakawea, are the only plentiful and
dependable water supplies for the oil industry in western North Dakota (North Dakota State Water
Commission. 20101. In 2011, North Dakota authorized the Western Area Supply Project, by which
Missouri River water (via the water treatment plant in Williston, North Dakota) will be supplied to
help meet water demands, including for oil and gas development, of the state's northwest counties
(WAWSA. 2011). In July 2012, the U.S. Army Corps of Engineers made available approximately 32.6
billion gal (123 billion L) of water per year from Lake Sakakawea for municipal and industrial water
demands over the next ten years fU.S. Army Corps of Engineers. 20111. The Army Corps estimated
that the oil and gas industry could use up to 8.8 billion gal (33.3 billion L) annually during this time
period in the 11-county surrounding area, and included this as part of the 32.6 billion gal total (123
billion L) to be made available (U.S. Army Corps of Engineers. 2011). For context, annual water use
for hydraulic fracturing in all North Dakota counties combined was approximately 2.2 billion gal
(8.3 billion L) per year in 2011 and 2012 according to EPA's FracFocus 1.0 project database
(Appendix Table B-2). As such, Lake Sakakawea appears to be an adequate resource to meet the
water demands of hydraulic fracturing in the region at least in the near term.
4.5.5 Arkansas and Louisiana
Arkansas and Louisiana were ranked seventh and tenth in the number of disclosures in the EPA
FracFocus 1.0 project database, respectively (Appendix Table B-5). Hydraulic fracturing activity in
Louisiana occurs primarily in the TX-LA-MS Salt Basin, which contains the Haynesville play; activity
in Arkansas is dominated by the Arkoma Basin, which contains the Fayetteville play (Figure 4-12).
4-42
-------
Chapter 4- Water Acquisition
Fayetteville Play
Arkoma
Haynt svilfe-Bossier
Western
Gulf
EIA Basins
Figure 4-12. Major U.S. EIA shale plays and basins for Arkansas and Louisiana.
Source: EIA (2015).
Types of water used: Surface water is reported as the primary source of water for hydraulic
fracturing operations in both Arkansas and Louisiana fANRC. 2014: LA Ground Water Resources
Commission. 2012: STRONGER. 20121. Quantitative information is lacking for Arkansas on the
proportion of water sourced from surface versus groundwater. However, data are available for
Louisiana, where an estimated 87% of water for hydraulic fracturing in the Haynesville Shale is
from surface water fLA Ground Water Resources Commission. 20121 (Table 4-1). In 2008, during
the early stages of development, hydraulic fracturing in Louisiana relied heavily on groundwater
from the Carrizo-Wilcox aquifer, and concerns for the sustainability of groundwater resources
prompted the state to encourage surface water withdrawals (LA Ground Water Resources
Commission, 20121.
The EPA FracFocus report suggests that significant reuse of wastewater may occur in Arkansas to
offset total fresh water used for hydraulic fracturing; 70% of all disclosures reporting a water
source indicated a blend of "recycled/surface," whereas 3% of disclosures reporting a water source
noted "fresh" as the exclusive water source fU.S. EPA. 2015b!.1 According to Veil (20111. Arkansas'
1 Ninety-three percent of Arkansas disclosures included information related to water sources fU.S. EPA, 2015b").
4-43
-------
Chapter 4- Water Acquisition
Fayetteville Shale wastewater is of relatively good quality (i.e., lowTDS), facilitating reuse.1 Data
are generally lacking on the extent to which hydraulic fracturing wastewater is reused in Louisiana.
Water use per well: Arkansas and Louisiana have the highest median water use per well of the states
we considered from the EPA FracFocus 1.0 project database, at 5.3 million and 5.1 million gal (20.1
million and 19.3 million L), respectively (Appendix Table B-5).2
Water use/consumption at the county scale: On average, hydraulic fracturing uses 408 million gal
(1.54 billion L) of water each year in Arkansas counties reporting activity, or 9.3% of 2010 total
county water use (26.9% of total county consumption) (Appendix Table B-2). In 2011 and 2012,
five counties dominated fracturing water use in Arkansas: Cleburne, Conway, Faulkner, Van Buren,
and White Counties (Appendix Table B-2). Van Buren, which is sparsely populated and thus has
relatively low total water use and consumption, is by far the Arkansas county highest in hydraulic
fracturing water use and consumption relative to 2010 total water use and consumption (56% and
168%, respectively) (Table 4-3).
In Louisiana, hydraulic fracturing water use is concentrated in six parishes in the far northwestern
corner of the state, associated with the Haynesville play.3 On average in 2011 and 2012, hydraulic
fracturing used 117 million gal (443 million L) of water annually per parish, representing
approximately 3.6% and 10.8% of 2010 total water use and consumption, respectively (Appendix
Table B-2). Operators in DeSoto Parish used the most water (over 1 billion gal (3.8 billion L)
annually). Hydraulic fracturing water use and consumption was highest relative to 2010 total water
use and consumption (35.5% and 83.2%, respectively) in Red River Parish (Table 4-3). These
numbers may be low estimates, since Louisiana required disclosures to the state or FracFocus, and
Arkansas required disclosures to the state but not FracFocus, during the time period analyzed (U.S.
EPA. 2015bl (Appendix Table B-5).
Potential for impacts: Water availability is generally higher in Arkansas and Louisiana than in states
farther west, reducing the potential for impacts to drinking water quantity and quality (Figure 4-6a,
Figure 4-7a; Text Box 4-2). However, generally high water availability in this region does not
preclude the potential for impacts at the local scale, particularly if surface water withdrawals occur
during seasonal low flow periods. For instance, precipitation is highest in Arkansas in the late
autumn and winter, with little rainfall occurring in the late spring and summer; thus, most small
streams do not flow year round (Entrekin etal.. 2015). Hydraulic fracturing surface water
withdrawals from small streams during seasonal low flows have the potential to impact the
quantity and quality of drinking water resources.
Additionally, in northwestern Louisiana, there are concerns about over-pumping of groundwater
resources. Prior to 2008, most operators in the Louisiana portion of the Haynesville Shale used
groundwater, withdrawing from the Carrizo-Wilcox, Upland Terrace, and Red River Alluvial aquifer
1 Veil ("20111 reports a range of 20,000-25,000 ppm TDS for Fayetteville Shale wastewater.
2 According to STRONGER (2012) and STRONGER (2011a). both states require disclosure of information on water use per
well, but this has not been synthesized into state level reports to date.
3 Louisiana is divided into parishes, which are similar to counties in other states.
4-44
-------
Chapter 4- Water Acquisition
systems (LA Ground Water Resources Commission. 20121. To mitigate stress on groundwater, the
state issued a water use advisory to the oil and gas industry that recommended Haynesville Shale
operators seek alternative water sources to the Carrizo-Wilcox aquifer, which is predominantly
used for public supply fLDEO. 20081. Operators then transitioned to mostly surface water, with a
smaller groundwater component (approximately 13% of all fracturing water used) (LA Ground
Water Resources Commission. 20121. Of this groundwater component, the majority (approximately
74%) still came from the Carrizo-Wilcox aquifer fLA Ground Water Resources Commission. 20121.
Although the potential for hydraulic fracturing withdrawals to affect water supplies and water
quality in the aquifer was reduced, it was not entirely eliminated. Despite Louisiana's water use
advisory, a combination of drought conditions and higher than normal withdrawals (for all uses,
not solely hydraulic fracturing) from the Carrizo-Wilcox and Upland Terrace aquifers caused
several water wells to go dry in July 2011 fLA Ground Water Resources Commission. 20121. In
August 2011, a groundwater emergency was declared for southern Caddo Parrish (LA Ground
Water Resources Commission. 20121. Hydraulic fracturing withdrawals contributed to these
conditions, alongside other users of water and the lack of precipitation.
4.6 Chapter Synthesis
In this chapter, we examined the potential for water acquisition for hydraulic fracturing to impact
the quantity and quality of drinking water resources, and identified factors affecting the frequency
or severity of impacts. Whether impacts occur from water acquisition for hydraulic fracturing
depends on the local balance between water withdrawals and availability, and this balance can be
modified by a combination of site or regional-specific factors. For this reason, information is needed
at the local scale to determine whether impacts actually occur, yet this information is not available
in many locations where hydraulic fracturing takes place; see Section 4.6.3 on Uncertainties below.
Despite these limitations, our chapter used the scientific literature, county level assessments, and,
where available, local case studies to point to areas with a higher potential for impacts; understand
local dynamics, including example cases of impacts; and identify common factors that increase or
decrease the frequency or severity of impacts. In this section, we summarize our major findings
regarding hydraulic fracturing water acquisition activities, potential impacts, and these common
factors (4.6.1 and 4.6.2). We then discuss uncertainities (4.6.3), and provide final conclusions
(4.6.4).
4.6.1 Major Findings
The first half of this chapter focused on water acquisition activities, providing an overview of the
types of water used (including sources, quality, and provisioning), water use per well, and water
use and consumption at the national, state, and county scale. The three major types of water used
for hydraulic fracturing are surface water, groundwater, and reused hydraulic fracturing
wastewater. Because trucking can be a major expense, operators tend to use water sources as close
to the well pad as possible. Operators usually self-supply surface water or groundwater directly,
but may also obtain water from public water systems or other suppliers. Hydraulic fracturing
operations in the eastern United States rely predominantly on surface water, whereas operations in
4-45
-------
Chapter 4- Water Acquisition
more semi-arid to arid western states use either surface water or groundwater. There are areas of
the country that rely entirely on groundwater supplies (e.g., western Texas).
Reuse of wastewater reduces the demand on fresh water sources, which currently supply the vast
majority of water used for hydraulic fracturing. The proportion of the water used in hydraulic
fracturing that comes from reused hydraulic fracturing wastewater is generally low; in a survey of
literature values from 10 states, basins, or plays, we found a median value of 5%, with this
percentage varying by location (Table 4-2).1 Available data on reuse trends indicate increasing
reuse of wastewater over time in both Pennsylvania and West Virginia, likely due to the lack of
nearby disposal options in Class II wells. Reuse as a percentage of water injected is typically lower
in other areas of the United States, likely in part because of the availability of disposal wells; see
Chapter 8 for more information.
The median amount of water used nationally per hydraulically fractured well was approximately
1.5 million gal (5.7 million L) in 2011 through early 2013 based on the EPA analysis of FracFocus
disclosures (U.S. EPA. 2015b, c). This increased to approximately 2.7 million gal (10.2 million L) in
2014, driven by a proportional increase in horizontal wells (estimated from data in Gallegos etal.,
20151. These national estimates represent a variety of fractured well types, including types
requiring much less water per well than horizontal shale gas wells. Thus, published estimates for
horizontal shale gas wells are typically higher (e.g., approximately 4 million gal (15 million L) per
well (Vengosh et al.. 2014), and should not be applied to all fractured wells to derive national
estimates. There was also wide variation within and among states and basins in the median per
well water volumes reported in 2011 and 2012, from more than 5 million gal (19 million L) in
Arkansas and Louisiana to less than 1 million gal (3.8 million L) in Colorado, Wyoming, Utah, New
Mexico, and California (U.S. EPA. 2015c). This variation can result from several factors, including
geologic formation, well length, and fracturing fluid formulation.
Hydraulic fracturing uses billions of gallons of water every year at the national and state scales, and
even in some counties. When expressed relative to total water use or consumption at these scales,
however, hydraulic fracturing generally accounts for only a small percentage, usually less than 1%.
These percentages are higher though in specific counties. Annual hydraulic fracturing water use
was 10% or more compared to 2010 total water use in 6.5% of counties with FracFocus disclosures
in 2011 and 2012 in the EPA FracFocus 1.0 project database, 30% or more in 2.2% of counties, and
50% or more in 1.0% of counties (Appendix Table B-2). Consumption estimates follow the same
pattern, with higher percentages in each category: hydraulic fracturing water consumption was
10%, 30%, and 50% or more of 2010 total water consumption in 13.5%, 6.2%, and 4.0% of counties
with FracFocus disclosures in the EPA FracFocus 1.0 project database (Appendix Table B-2). Thus,
hydraulic fracturing represents a relatively large user and consumer of water in these counties.
Whether water quantity or quality impacts occur from water acquisition for hydraulic fracturing
depends on the local balance between water withdrawals and availability. From our survey of the
literature and our county level assessments, southern and western Texas appear to have the
1 Note that reused water as a percentage of total water injected differs from the percentage of wastewater that is reused.
See Section 4.2 and Chapter 8 for more information.
4-46
-------
Chapter 4- Water Acquisition
highest potential for impacts of the areas assessed in this chapter, given the combination of high
hydraulic fracturing water use, relatively low water availability, intense periods of drought, and
reliance on declining groundwater resources; see Section 4.6.2 on Factors below. Importantly, our
results do not preclude the possibility of local water impacts in areas with comparatively lower
potential, nor do they necessarily mean impacts have occurred in the high potential areas. Our
survey, however, provides an indicator of areas with higher potential for impacts, and could be used
to target resources or future studies.
In two example cases, local impacts to drinking water resources occurred in areas with increased
hydraulic fracturing activity. In a detailed case study, Scanlonetal. (2014b) observed generally
adequate water supplies for hydraulic fracturing in the Eagle Ford play in southern Texas, except in
specific locations. They found excessive drawdown of groundwater locally, with estimated declines
of ~100-200 ft (30-60 m) in a small proportion of the play (~6% of the area) after hydraulic
fracturing activity increased in 2009. In 2011, drinking water wells in an area overlapping with the
Haynesville Shale ran out of water due to higher than normal groundwater withdrawals and
drought fLA Ground Water Resources Commission. 20121. Hydraulic fracturing water withdrawals
contributed to these conditions, along with other water users and the lack of precipitation. By
contrast, two EPA case studies in the Upper Colorado and the Susquehanna River Basins found
minimal impacts from hydraulic fracturing withdrawals currently (U.S. EPA. 2015e) (Sections 4.5.2,
4.5.3).
These site-specific findings emphasize the need to focus on regional and local dynamics when
considering the impacts from hydraulic fracturing water withdrawals. The case studies and the
scientific literature as a whole suggest some common factors that increase or decrease the
frequency or severity of impacts. These are summarized in the section below.
4.6.2 Factors Affecting Frequency or Severity of Impacts
The potential for impacts depends on the combination of water withdrawals and water availability
at a given withdrawal location. Where water withdrawals are relatively low compared to water
availability, impacts are unlikely to occur. Where water withdrawals are relatively high compared
to water availability, impacts are more likely.
Areas reliant on declining groundwater are particularly vulnerable to more frequent and severe
impacts from cumulative water withdrawals, including withdrawals for hydraulic fracturing.
Groundwater recharge rates can be extremely low, and groundwater pumping is exceeding
recharge rates in many areas of the country (Konikow. 2013). When pumping exceeds recharge, the
cumulative effects of withdrawals are manifested in declining water levels. For this reason, water
levels in many aquifers in the United States have declined substantially over the last century
fKonikow. 20131. Cumulative drawdowns can affect surface water bodies since groundwater can be
the source of base flow in streams (Winter et al.. 1998). and alter groundwater quality by
mobilizing chemicals from geologic sources, among other means (DeSimone et al.. 2014: Alley etal..
1999). Although in many of these areas (e.g., the Ogallala aquifer), irrigated agriculture is the
dominant user of groundwater, hydraulic fracturing withdrawals now also contribute to declining
groundwater levels. Hydraulic fracturing groundwater consumption, for example, exceeds
4-47
-------
Chapter 4- Water Acquisition
estimated recharge rates in the seven most active hydraulic fracturing counties in the Eagle Ford
Shale in southern Texas fSteadman etal.. 20151. When necessary, state and local governments have
encouraged or mandated industry to use surface water over groundwater, as evidenced in both
Louisiana and North Dakota.
Among surface water sources, smaller streams, even in humid areas, are more vulnerable to
frequent and severe impacts from withdrawals. A detailed EPA case study found that streams with
the smallest contributing areas in northeastern Pennsylvania were particularly vulnerable to
withdrawals (U.S. EPA. 2015e). Protecting smaller streams from excessive withdrawals is probably
most important for aquatic life, but may also protect drinking water quantity and quality in certain
instances.
Seasonal or long-term drought can also make impacts more frequent and severe for surface water
and groundwater sources. Hot, dry weather depletes surface water bodies and reduces or prevents
groundwater recharge, while water demand often increases simultaneously (e.g., for irrigation).
The EPA case study in Pennsylvania found that even large streams could be vulnerable to
withdrawals during times of low flows fU.S. EPA. 2015el. Much of the western United States has
experienced prolonged periods of drought over the last decade (Figure 4-8). This dynamic will
likely be magnified by future climate change in certain locations fMeixner etal.. 20161.
By contrast to the above factors, consumption of water for hydraulic fracturing does not appear to
substantially influence the frequency or severity of impacts. There are concerns that hydraulic
fracturing permanently removes water from the hydrologic cycle, posing a threat to long-term
water supplies. Since impacts occur locally and depend on the local water balance, impacts can
occur regardless of whether the water is withdrawn and returned to the larger hydrologic cycle
elsewhere or whether it is permanently sequestered underground. We acknowledge that whether
the water is returned to the larger hydrologic cycle may make a difference for the water budget of a
larger area, such as on the state, regional, or national scale. For example, water converted to steam
during thermoelectric cooling in one location may condense and fall as precipitation in an adjacent
state or region. At these larger scales, however, hydraulic fracturing water consumption is a very
small fraction of total water availability.1 Plus, at these scales, there are other larger factors that can
affect regional water budgets, but which are out of scope for this assessment.2 For these reasons,
focusing on consumption distracts from the more salient issue that impacts depend upon the spatial
and temporal balance between local water withdrawals and availability.
1 For example, hydraulic fracturing used approximately 3.3 billion gal (12.5 billion L] of water on average annually in all
Colorado counties with hydraulic fracturing activities combined according to FracFocus disclosures in 2011 and 2012
(Appendix B-l]. Using the consumption rate of 82.5% yields a consumption estimate of approximately 2.7 billion gal (10.2
billion L]. This would be approximately 0.1% of the fresh water and total water availability metrics used in Textbox 4-2
for all of those same counties combined (approximately 2.6 trillion gal (9.8 trillion L] of fresh water and total water
available].
2 The combustion of methane produced by hydraulic fracturing, for example, adds water molecules to the environment,
and at large scales, this may affect regional water budgets. However, quantifying this is outside the scope of this
assessment. Similarly, there are other larger factors (e.g., water used for cooling thermoelectric power plants] that can
affect regional water budgets, but these are also outside the scope of this assessment.
4-48
-------
Chapter 4- Water Acquisition
There are also factors that can decrease the frequency and severity of any impacts from water
withdrawals. The literature suggests that water management, particularly wastewater reuse, the
use of brackish groundwater, the use of passby flows, and transitioning from limited groundwater
sources to more abundant surface water sources can reduce impacts. Reuse is not a universal
solution, since in many areas of the country wastewater volumes from one well are often a small
percentage of the water needed to fracture the next well. In the Marcellus Shale, for instance, 100%
reuse of the wastewater produced from one well means reducing fresh water demand by 10 or 30%
for the next (Section 4.2.1; Chapter 7). Nevertheless, reuse can be an important local factor reducing
fresh water demand.
Switching to brackish water is another means by which fresh water demand can be—and is in some
locations—reduced. This is a source of alternative water in western and southern Texas, for
example. In these areas, use of brackish water is currently reducing impacts to fresh water sources,
and could with continued use reduce future impacts (Scanlon et al.. 2014b: Nicot etal.. 2012). Our
county level estimates suggest that brackish water could readily meet the volume demanded by
hydraulic fracturing in Texas.
Water management also includes passby flows, a low stream flow threshold below which
withdrawals are not allowed. Evidence suggests passby flows can be effective in protecting streams
from hydraulic fracturing water withdrawals (U.S. EPA. 2015e). Finally, as evidenced by examples
in both North Dakota and Louisiana, water management may include transitioning from declining
groundwater sources to surface water, if available.
4.6.3 Uncertainties
There are several uncertainties inherent in our assessment of the potential impacts of water
acquisition for hydraulic fracturing. The largest uncertainties stem from the lack of literature and
data on this subject at local scales. Because impacts occur at a given withdrawal point, our
assessment could assess the potential for impacts, but often could not determine if potential
impacts were realized in the absence of local data. The exceptions were local case studies from the
Eagle Ford play in Texas, the Upper Colorado River Basin in Colorado, and the Susquehanna River
Basin in Pennsylvania. Moreover, it is also not clear if local impacts, for example a drinking water
well going dry, are likely to be documented in the scientific literature.
Other uncertainties arise from data limitations on the volume and types of water used or consumed
for hydraulic fracturing, future water use projections, and water availability estimates. There are no
nationally consistent data sources, and therefore, water use estimates must be based on multiple,
individual pieces of information. For example, in their National Water Census, the USGS includes
hydraulic fracturing in the broader category of "mining" water use, but hydraulic fracturing water
use is not reported separately fMaupin et al.. 20141. There are locations where average annual
hydraulic fracturing water use in 2011 and 2012 in the EPA FracFocus 1.0 project database
exceeded total mining water use in 2010, and one county where it exceeded all water use (U.S. EPA.
2015c: Maupin etal.. 20141. This could be due to a rapid increase in hydraulic fracturing water use,
differences in methodology between the two databases (i.e., the USGS 2010 National Water Census
and the EPA FracFocus 1.0 project database), or both.
4-49
-------
Chapter 4- Water Acquisition
We used the EPA FracFocus 1.0 project database for water use estimates, which itself has
limitations. Many states in the project database did not require disclosure to FracFocus during the
time period analyzed fU.S. EPA. 2015bl We conclude that this likely does not change the overall
hydraulic fracturing water use patterns observed across the United States (Text Box 4-1), but could
affect particular county level estimates. Also, the database covered the time period of 2011 through
early 2013. Thus, changes in the industry since then are not reflected in these data.
Hydraulic fracturing water use data that are often provided as water use associated with a
particular well. While this is valuable information, the potential impacts of water acquisition for
hydraulic fracturing could be better assessed if data were also available at the withdrawal point If
the total volume, date, location, and type (i.e., surface water or groundwater; and fresh, brackish, or
reused wastewater) of each water withdrawal were documented, effects on availability could be
better estimated. For example, surface withdrawal points could be aggregated by watershed or
aquifer to estimate effects on downstream flow, groundwater levels, and water quality. Some of this
information is available in disparate forms, but the lack of nationally consistent data on water
withdrawal locations, timing, and amounts—data that are publicly available, easy to access and
analyze—limits our assessment of potential impacts. The Susquehanna River Basin Commission
collects this type of detailed data on hydraulic fracturing water withdrawals, but this type of
information is not widely available across the nation.
Future hydraulic fracturing water use is also a source of uncertainty. Because water withdrawals
and potential impacts are concentrated in certain localized areas, water use projections need to
match this scale. Projections are available for Texas at the county scale, but more information at the
county or sub-county scale is needed in other states with hydraulic fracturing activity and water
availability concerns (e.g., northwest North Dakota, eastern Colorado). Due to a lack of data, we
generally could not assess future water use and the potential for impacts in most areas of the
country, nor could we examine these in combination with other relevant factors (e.g., climate
change or population growth).
4.6.4 Conclusions
With notable exceptions, hydraulic fracturing uses and consumes a relatively small percentage of
water when compared to total use, consumption, and availability at the national, state, and county
scale. Despite this, impacts on drinking water resource quantity and quality from hydraulic
fracturing water acquisition can occur at the local scale, because hydraulic fracturing water
withdrawals are often concentrated in space and time, and impacts depend upon the local balance
between withdrawals and availability. In two example cases, local impacts to drinking water
resource quantity occurred in areas with increased hydraulic fracturing activity (e.g., in Texas's
Eagle Ford play, and in Louisiana's Haynesville Shale). Declining groundwater resources, especially
in the western United States, are particularly vulnerable to withdrawals, as are smaller streams,
even in the more humid East. Finally, there are factors that increase or decrease the frequency and
severity of impacts—included in this are times of low water availability, such as during drought,
which can increase the frequency and severity of impacts, or conversely water management
practices (e.g., shifting to brackish water, or passby flows), which can help protect drinking water
resources.
4-50
-------
Chapter 5 - Chemical Mixing
Chapter 5. Chemical Mixing
Abstract
This chapter provides an analysis of the potential impacts on drinking water resources during the
chemical mixing stage of the hydraulic fracturing water cycle and the factors governing the frequency
and severity of these impacts. The chemical mixing stage includes the mixing of base fluid (90% to 97%
by volume, typically water), proppant (2% to 10% by volume, typically sand), and additives (up to 2%
by volume) on the well pad to make hydraulic fracturing fluid. This fluid is engineered to create and
extend fractures in the targeted formation and to carry proppant into the fractures. Concentrated
additives are delivered to the well pad and stored on site, often in multiple, closed containers, and
moved around the well pad in hoses and tubing.
Changes in drinking water quality can occur if spilled fluids reach groundwater or surface water
resources. In this assessment, a spill is considered to be any release of fluids. The EPA's analysis found
that spills and releases of chemicals and fluids have occurred during the chemical mixing stage and have
reached soil and surface water receptors. Spills of hydraulic fracturing fluids or additives included in the
analysis had a median spill volume of 420 gal (1,590 L), with a range of 5 to 19,320 gal (9 to 72,130 L).
Spills were caused most often by equipment failure or human error. The potential for spilled fluids to
reach, and therefore impact, groundwater or surface water resources depends on the composition of the
spilled fluid, spill characteristics, spill response activities, and the fate and transport of the spilled fluid.
The movement of spilled hydraulic fracturing fluids and chemicals through the environment is difficult
to predict, because spills are site- and chemical-specific, and because hydraulic fracturing-related spills
are typically complex mixtures of chemicals. Physicochemical properties, which depend on the
molecular structure of a chemical, govern whether spilled chemicals volatilize, sorb, transform, and
travel. Spill prevention practices and spill response activities can prevent spilled fluids from reaching
ground or surface drinking water resources.
The severity of potential impacts on water quality from spills of additives or hydraulic fracturing fluids
depends on the identity and amount of chemicals that reach ground or surface water resources, the
hazards associated with the chemicals, and the characteristics of the receiving water body. The lack of
monitoring following spills, along with the lack of publicly available information on the composition of
additives and fracturing fluids, containment and mitigation measures in use, the proximity of chemical
mixing to drinking water resources, and the fate and transport of spilled fluids limits the EPA's ability to
fully assess potential impacts on drinking water resources and their frequency and severity. This
chapter shows that spills of additives and hydraulic fracturing fluids during the chemical mixing stage of
the hydraulic fracturing water cycle have occurred and have reached and impacted drinking water
resources.
5-1
-------
Chapter 5 - Chemical Mixing
This page is intentionally left blank.
5-2
-------
Chapter 5- Chemical Mixing
5. Chemical Mixing
5.1 Introduction
This chapter provides an analysis of the potential impacts on drinking water resources during the
chemical mixing stage of the hydraulic fracturing water cycle and the factors governing the
frequency and severity of these impacts. Chemical mixing is a complex process that requires the use
of specialized equipment and a range of different additives to produce the fluid that is injected into
a well to fracture the formation. This fluid, the hydraulic fracturing fluid, generally consists of a
base fluid (typically water), a proppant (typically sand), and additives (chemicals), although there is
no standard or single composition of hydraulic fracturing fluid used. The number, type, and amount
of chemicals used to create the hydraulic fracturing fluid vary from well to well based on site- and
operator-specific factors. Spills may occur at any point in the chemical mixing process.1 The
potential for spilled fluids to reach, and therefore impact, ground or surface water resources
depends on the composition of the spilled fluid, spill characteristics, spill response activities, and
the fate and transport of the spilled fluid. This chapter is structured around these concepts.
The chapter starts by discussing the characteristics of hydraulic fracturing fluids (Sections 5.2 to
5.4). This includes an introductory overview of the chemical mixing process (Section 5.2), a
description of the different components of the hydraulic fracturing fluid (Section 5.3), the range of
different chemicals used and their classes, the most frequently used chemicals nationwide, and
volumes used (Section 5.4).2 (Appendix H provides a list of chemicals that the EPA identified as
being used in hydraulic fracturing fluids.)
The chapter continues with a discussion on how chemicals are managed on the well pad, the
characteristics of spills when they occur, and spill response activities (Sections 5.5 to 5.7). This
includes a description on how potential impacts of a spill on drinking water resources depends
upon chemical management practices, such as storage, on-site transfer, and equipment
maintenance (Section 5.5). A summary analysis of reported spills and their common causes at
hydraulic fracturing sites is then presented (Section 5.6). Then, there is a discussion on the different
efforts of spill prevention, containment, and mitigation (Section 5.7).
Next, the fate and transport of spilled chemicals is discussed (Section 5.8). This section includes
how a chemical can move through the environment and transform, and what governs exposure
concentrations of chemicals in the environment. Due to the complexities of the processes and the
site-specific and chemical-specific nature of spills, it is difficult to develop a full assessment of their
fate and transport This section provides a general overview and discusses how the fate and
transport of a chemical depends on site conditions, environmental conditions, physicochemical
1 In this assessment, a spill is considered to be any release of fluids. Spills can result from accidents, fluid management
practices, or illegal dumping.
2 Chemical classes are groupings of different chemicals based on similar features, such as chemical structure, use, or
physical properties. Examples of chemical classes include hydrocarbons, alcohols, acids, and bases.
5-3
-------
Chapter 5 - Chemical Mixing
properties of the released chemicals, fluid composition, volume of the release, the proximity to a
drinking water resource, and the characteristics of the drinking water resource that is the receptor.
Next is an overview of on-going changes in chemical use in hydraulic fracturing, with an emphasis
on industry efforts to reduce potential impacts from surface spills by using fewer and safer
chemicals (Section 5.9). The chapter concludes by providing a synthesis, including a summary of
findings, factors that affect frequency and severity of potential impacts, and a discussion of
uncertainties and data gaps (Section 5.10).
Due to the limitations of available data and the scope of this assessment, it is not possible to provide
a detailed analysis of all of the factors listed above. Data limitations preclude a quantitative analysis
of the likelihood or severity of chemical spills or impacts. Spills that occur off-site, such as those
during transportation of chemicals to the site or storage of chemicals in staging areas, are out of the
scope of this assessment. This chapter qualitatively characterizes the potential for impacts on
drinking water resources given the current understanding of overall operations and specific
components of the chemical mixing process.
5.2 Chemical Mixing Process
Understanding the chemical mixing process is necessary to understand how, why, and when spills
might occur. This section provides a general overview of the chemical mixing stage of the hydraulic
fracturing water cycle fCarter etal.. 2013: Knappe and Fireline. 2012: Spellman. 2012: Arthur et al..
2008). Figure 5-1 shows a hydraulic fracturing site during the chemical mixing process. In our
discussion, we focus on the types of additives used at each phase of the process. While similar
processes are used to fracture horizontal and vertical wells, a horizontal well treatment is
described here. Horizontal well treatments are likely to be more complex and therefore illustrative
of the variety of practices that have become more prevalent over time with advances in technology
(Chapter 3). A water-based system is described, because water is the most commonly used base
fluid, appearing in more than 93% of FracFocus 1.0 disclosures between January 1, 2011 and
February 28, 2013 (U.S. EPA. 2015a).1 While the number and types of additives may vary widely,
the basic chemical mixing process and the on-site layout of hydraulic fracturing equipment are
similar across sites fBT Services Company. 20091. Equipment used in the chemical mixing process
typically consists of chemical storage trucks, water supply tanks, proppant supply, slurry blenders,
a number of high-pressure pumps, a manifold, surface lines and hoses, and a central control unit
Detailed descriptions of specific additives and the equipment used in the process are provided in
Sections 5.3 and 5.5, respectively.
1 FracFocus f www.fracfocus.orgl is a registry of information of water and chemical use in wells in which hydraulic
fracturing is conducted. More details are provided in Text Box 5-1.
5-4
-------
Chapter 5 - Chemical Mixing
Figure 5-1. Representative hydraulic fracturing site showing equipment used on-site during
the chemical mixing process.
The frac well head is located in the center bottom (green), the manifold runs down the middle, and high pressure
pumps lead into the manifold from either side. Source: Schlumberger.
At a newly-drilled production well, the chemical mixing process begins after the drilling, casing, and
cementing processes are finished and hydraulic fracturing equipment has been setup and
connected to the well. The process can generally be broken down into one or more sequential
stages with specific chemicals added at different phases during each stage phase to achieve a
specific purpose fKnappe and Fireline. 2012: Fink. 20031. The process for water-based hydraulic
fracturing is described in Figure 5-2 below.
The first phase is the cleaning and preparation of the well. The fluid used in this phase is often
referred to as the pre-pad fluid, pre-pad volume, or spearhead. Acid is typically the first chemical
introduced. Acid, with a concentration of 3% to 28% (by volume, typically hydrochloric acid, HC1),
is used to clean any cement left inside the well from cementing the casing and dissolve any pieces of
rock that may remain in the well that could block the perforations.1 Acid is typically pumped
directly from acid storage tanks or tanker trucks, without being mixed with other additives. The
first, or pre-pad, phase may also involve mixing and injection of additional chemicals to facilitate
the flow of fracturing fluid introduced in the next phase of the process. These additives may include
biocides, corrosion inhibitors, friction reducers, and scale inhibitors (Carter etal.. 2013: King. 2012:
Knappe and Fireline. 2012: Spellman. 2012: Arthur et al.. 20081.
1 Prior to the injection of the pad fluid, for wells that are cased in the production zone, the well casing is typically
perforated to provide openings through which the pad fluid can enter the formation. A perforating gun is typically used to
create small holes in the section of the well being fractured. The perforating gun is lowered into position in the horizontal
portion of the well. An electrical current is used to set off small explosive charges in the gun, which creates holes through
the well casing and out a short distance into the formation f Gupta and Valko. 20071
5-5
-------
Chapter 5 - Chemical Mixing
Chemical Mixing Phases
Hydraulic Fracturing Stages
V
Pre-Pad: Acid
and/or chemical
may be injected
to clean the well
Phase repeated for each stage:
Perforation: To create openings in the well to access targeted geologic formation
Pad: Inject base fluid and additives into the formation to initiate fractures
Proppant: Inject proppant with base fluid and additives to grow and prop open fractures
Flush: Fluid injected to push proppant remaining in well into fractures
Stage 1: Chemicals &
Proppant Injected
Stage 2: Chemicals &
Proppant Injected
Proppant
gradient
Stage 3: Chemicals
& Proppant Injected
V
Stage 4: Chemicals
& Proppant Injected
Proppant
gradient
JL
<
rD
QL
U)
a>
n
r-f-
o'
Horizontal Well Section
Proppant
gradient
Proppant
gradient
Perforations
Figure 5-2. Overview of a chemical mixing process of the hydraulic fracturing water cycle.
This figure outlines the chemical mixing process for a generic water-based hydraulic fracture of a horizontal well.
The chemical mixing phases outline the steps taken at the surface in the overall fracturing job, while the hydraulic
fracturing stages outline how each section of the horizontal well would be fractured beginning with the toe of the
well, shown on left-side. The proppant gradient represents how the proppant size may change within each stage of
fracturing as the fractures are elongated. The chemical mixing process is repeated depending on the number of
stages used for a particular well. The number of stages is determined in part by the length of the horizontal leg. In
this figure, four stages are represented, but typically, a horizontal fracturing treatment would consist of 10 to 20
stages per well (Lowe et al„ 2013). Fracturing has been reported to be done in as many as 59 stages (Pearson et al„
2013).
In the second phase, a hydraulic fracturing fluid, typically referred to as the pad or pad volume, is
mixed, blended, and pumped down the well under high pressure to create fractures in the
formation.1 The pad is a mixture of base fluid, typically water, and additives and is designed to
create, elongate, and enlarge fractures in the targeted geologic formation when injected under high
pressure fGupta and Valko. 20071 (see Section 6.3 for additional information on fracture growth
following injection). A typical pad consists of, at minimum, a mixture of water and friction reducer.
A typical pad consists of, at minimum, a mixture of water and friction reducer. Other additives (see
U.S. EPA (2015ai and Table 5-1) may be used to facilitate flow and kill bacteria (Carter et al.. 2013:
King. 2012: Knappe and Fireline. 2012: Spellman. 2012: Arthur etal.. 2008). The pad is pumped
into the formation through perforations or sliding sleeves in the well casing.
1 In terms of chemical mixing, "pad" is a term used to describe hydraulic fracturing fluid without solid at the start of the
fracturing of the formation. In terms of the entire hydraulic fracturing process, the "well pad" or "pad" is the area of land
where drilling occurs.
5-6
-------
Chapter 5- Chemical Mixing
In the third phase, proppant, typically sand, is mixed into the hydraulic fracturing fluid. The
proppant volume, as a proportion of the injected fluid, is increased gradually until the desired
concentration in the fractures is achieved. Gelling agents, if used, are also mixed with the proppant
and base fluid in this phase to increase the viscosity to help carry the proppant. Additional
chemicals may be added to gelled fluids, initially to maintain viscosity and later to break down the
gel and decrease viscosity, so the hydraulic fracturing fluid can more readily flow back out of the
formation and through the well to facilitate production from the fractured formation fCarter etal..
2013: King. 2012: Knappe and Fireline. 2012: Spellman. 2012: Arthur etal.. 2008).
A final flush or clean-up phase may be conducted after the stage is fractured, with the primary
purpose of maximizing well productivity. The flush is a mixture of water and additives that work to
aid the placement of the proppant, clean out the chemicals injected in previous phases, and prevent
microbial growth in the fractures (Knappe and Fireline. 2012: Fink. 2003).
The second, third, and fourth phases are repeated multiple times in a well with multi-stage
hydraulic fracturing. For each stage, the well is typically perforated and fractured beginning at the
end, or toe, of the well and proceeding backwards toward the bend or heel of the well, near the
vertical section. In vertical wells, stages typically begin in deeper portions of the well and proceed
shallower. Each fractured stage is isolated before the next stage is fractured. The number of stages
sets how many times the chemical mixing process is repeated at the site surface (Figure 5-2). The
number of stages increases with longer intervals of the well subjected to hydraulic fracturing
(Carter etal.. 2013: King. 2012: Knappe and Fireline. 2012: Spellman. 2012: Arthur etal.. 2008).
The number of stages per well can vary, with several sources suggesting between 10 and 20 stages
is typical fGNB. 2015: Lowe etal.. 20131.1 The full range reported in the literature is much wider,
with one source documenting between 1 and 59 stages per well f Pearson etal.. 20131 and others
reporting values within this range (NETL. 2013: STO. 2013: Allison etal.. 2009). The number of
stages per well seems to have increased over time. One study reports that the average number of
stages per horizontal well rose from approximately 10 in 2008 to 30 in 2012 fPearson etal.. 20131.
As more stages are used, the total volume of hydraulic fracturing fluid and chemicals increase. This
increases the potential, frequency, and severity of surface spills associated with chemical mixing
and thus potential impacts on drinking water resources.
In each of these phases, water is usually the primary component of the hydraulic fracturing fluid,
though the exact composition of the fluid injected into the well changes over the duration of each
stage. In water-based hydraulic fracturing, the composition, by volume, of a typical hydraulic
fracturing fluid is 90% to 97% water, 2% to 10% proppant, and 2% or less additives (Carter etal..
2013: Knappe and Fireline. 2012: SWN. 20111.2
1 The number of stages has been reported to be 6 to 9 in the Huron in 2009 (Allison etal.. 2009). 13 to 32 in the Marcellus
fNETL. 20131. and up to 40 by STO f20131.
2 This range is based on a compilation of sources. Sources present compositions as by mass, by volume, or without
specificity. Because of non-additive volumes, the composition by volume can be different before and after mixing. By
mass: 90% water, 8-9% proppant, 0.5 to 1.5% additives ("Knappe and Fireline. 20121: 88% water, 11% proppant, <1%
5-7
-------
Chapter 5 - Chemical Mixing
5.3 Overview of Hydraulic Fracturing Fluids
Hydraulic fracturing fluids are formulated to perform specific functions: create and extend the
fracture and transport and place the proppant in the fractures fMontgomerv. 2013: Spellman. 2012:
Gupta and Valko. 20071.1 The hydraulic fracturing fluid generally consists of three parts: (1) the
base fluid, which is the largest constituent by volume, (2) the additives, and (3) the proppant
Additives, which can be a single chemical or a mixture of chemicals, are chosen to serve a specific
purpose in the hydraulic fracturing fluid (e.g., friction reducer, gelling agent, crosslinker, biocide)
fSpellman. 20121. Throughout this chapter, "chemical" is used to refer to an individual chemical
substance (e.g., methanol, petroleum distillates).2 Proppants are small particles, usually sand, mixed
with fracturing fluid to hold fractures open so that the target hydrocarbons can flow from the
formation through the fractures and up the wellbore. The combination of additives, and the mixing
and injection process, varies based on a number of factors as discussed below. The additive
combination determines the amount and type of equipment required for storage and, therefore,
contributes to the determination of the potential for spills and impacts of those spills.
The particular composition of a hydraulic fracturing fluid is designed based on empirical
experience, the geology and geochemistry of the production zone, economics, goals of the fracturing
process, availability of the desired chemicals, and preference of the service company or operator
(Montgomery. 2013: ALL Consulting. 2012: Klein etal.. 2012: Ely. 1989).3 No single set of specific
chemicals is used at every site. Multiple types of fracturing fluids may be appropriate for a given
site, and any given type of fluid may be appropriate at multiple sites. For the same type of fluid
formulation, there can be differences in the additives, chemicals in those additives, and the
concentrations selected. There are broad criteria for hydraulic fracturing fluid selection based on
the targeted production zone temperature, pressure, water sensitivity, and permeability (Gupta and
Valko. 2007: Elbel and Britt. 2000). Figure 5-3 provides a general overview of the types of decisions
to determine which fluid can be used for different situations. Similar fluids may be appropriate for
different formations. For example, crosslinked fluids with 25% nitrogen foam (titanate or zirconate
crosslink + 25% nitrogen) can be used in both gas and oil wells with high temperatures and
additive (as median maximum concentration] ("U.S. EPA. 2015al. 94% water, 6% proppant, <1% additive fSiolander etal..
20111.88% water, 11% proppant, <1% additive fOSHA. 2014a. bl. By volume: 95% water, 5% proppant, <1% additive
(before mixing], 97% water, 2% proppant, <1% additive (after mixing] fSiolander etal.. 20111.90% water, 10%
proppant, <1% additive (before mixing], 95% water, 5% proppant, <1% additive (after mixing] fOSHA. 2014a. b], 98-
99.5%, water and sand 0.5 to 2% additives fSpellman. 20121. Not specified: 99.9% water and sand, 0.1% chemicals fSWN.
2011], 98-99%o water and proppant, 1-2 % additives (Carter etal.. 2013],
1 We use "hydraulic fracturing fluid" to refer to the fluid that is injected into the well and used to create and hold open
fractures the formation.
2 In this chapter, because of the way many chemicals are reported, we use the word "chemical" to refer to any individual
chemical or chemical substance that has been assigned a CASRN (Chemical Abstracts Service Registry Number]. A CASRN
is a unique identifier for a chemical substance, which can be a single chemical (e.g., hydrochloric acid, CASRN 7647-01-0]
or a mixture of chemicals (e.g., hydrotreated light petroleum distillates (CASRN 64742-47-8], a complex mixtures of C9 to
C16 hydrocarbons]. For simplicity, we refer to both pure chemicals and chemical substances that are mixtures, which
have a single CASRN, as "chemicals."
3 Empirical experience tends to provide better result as operators gain experience at a new site or geology increases.
When an operator moves to a new basin geology, there may be less than optimal results. With experience and
understanding of the geology increases, the empirical evidence will inform what hydraulic fracturing fluid composition
works better than others.
5-8
-------
Chapter 5- Chemical Mixing
Gas Weil
less
175°F
; Low pressure
: ; Of
No *•*•! water sensitive
No
Low pressure
or
water sensitive ~i Yes
less
150°F
More
Foamed
fluids
Surface-delafei
borate
crosslinked guar
or HPS
j ten H 30Q8F ,J More
, Less .
!250°f
1 70% - 75%
! quality or
: low pH
:i crosslinked
5 +25% CO,
; Delayed borate, ' Zirconate
«titanate or zirconate I crosslinked
-*r
! Low pH
; crosslinked : ? Titanate
+
25% COj
Super
crosslinked guar
HP6
borate
+ N2
or
; zirconate
: crosslinked
: +
:• 25% u,
: No ij kfW > 1000 md-ft + xf > 300 ft yM ;
: Linear fluids :
Oil Wei
; No ;
Lass
Low
pressure
Very water sensitive __j more
m
3 Less ;•
¦ ¦ ¦ '.I- '
i;
: ! Oil; ;
i i; polymer
{emulsion
\ Yes i
No
low
pressure
Yes
See gas
well guide
Gelled oil
Gelled oil + Na
tow
pressure
* ?es .
: 200°F ,
j More j : Lest -« 200°F -i More ;
i; -
Zirconate
issllnfo
HPG
tess K 200 f *1 Mora
I
Low pH
+
25% CO,
Titanate or zirconate
25% N,
Figure 5-3. Example hydraulic fracturing fluid decision tree for gas and oil wells.
This decision tree figure serves as an example of the factors that determine the type of hydraulic fracturing fluid chosen to fracture a given formation, depending on whether the
well will produce oil or gas. Factors include water sensitivity, formation temperature, and pressure. HPG is hydroxypropylguar, guar derivatized with propylene oxide.
Parameters are: kf, fracture permeability, w is the fracture width, and xf is the fracture half-length. This figure was chosen to represent the differences between oil and gas wells
and the types of decisions involved with choosing a fluid. This is adapted from Elbel and Britt (2000) and, as such, is dated to that time period. Since then, slickwater has become
increasingly popular due to its simplicity and cheaper cost, and slickwater has often replaced linear and crosslinked gelled fluids, especially in shales. Other decision tree figures
may exist. © 2000 Schlumberger. First published by John Wiley & Sons Ltd. All rights reserved.
5-9
-------
Chapter 5 - Chemical Mixing
variation in water sensitivity.12 One of the most important properties in designing a hydraulic
fracturing fluid is the viscosity fMontgomerv. 20131.3
Table 5-1 provides a list of common types of additives, their functions, and the most frequently
used chemicals for each purpose based on the EPA's analysis of disclosures to FracFocus 1.0 fU.S.
EPA. 2015a, hereafter referred to as the EPA FracFocus 1.0 report), the EPA's project database of
disclosures to FracFocus 1.0 fU.S. EPA. 2015c, hereafter referred to as the EPA FracFocus 1.0
project database), and other literature sources.4 Additional information on more additives can be
found in U.S. EPA C2015al.
A general description of typical hydraulic fracturing fluid formulations nationwide is difficult,
because fracturing fluids vary from well to well. Based on the EPA FracFocus 1.0 report, the median
number of chemicals reported for each disclosure was 14, with the 5th to 95th percentile ranging
from four to 28 (see Appendix H for a list of hydraulic fracturing fluid chemicals). The median
number of chemicals per disclosure was 16 for oil wells and 12 for gas wells fU.S. EPA. 2015a).
Other sources have stated that between three and 12 additives and chemicals are used
(Schlumberger. 2015: Carter etal.. 2013: Spellman. 2012: GWPC and ALL Consulting. 2009).5
Water, the most commonly used base fluid for hydraulic fracturing, is inferred to be used as a base
fluid in more than 93% of EPA FracFocus 1.0 disclosures fU.S. EPA. 2015cl. Alternatives to water-
based fluids, such as hydrocarbons and gases, including carbon dioxide and nitrogen-based foam,
may also be used based on formation characteristics, cost, or preferences of the well operator or
service company (ALL Consulting. 2012: GWPC and ALL Consulting. 2009). Non-aqueous base fluid
ingredients were identified in 761 (2.2%) of EPA FracFocus 1.0 disclosures fU.S. EPA. 2015al. Gases
and hydrocarbons may be used alone or blended with water; more than 96% of the disclosures
identifying non-aqueous base fluids are blended fU.S. EPA. 2015al. There is no standard method to
categorize the different fluid formulations (Patel etal.. 2014: Montgomery. 2013: Spellman. 2012:
Gupta and Valko. 2007). Therefore, we broadly categorize the fluids as water-based or alternative
fluids.
1A crosslinked fluid is a fluid that has polymers that have been linked together through a chemical bond. A crosslink
chemical is added to have the polymer chains linked together to form larger chemical structures with higher viscosity.
The increased fracturing fluid viscosity allows the fluid to carry more proppant into the fractures. The fracturing fluid
remains viscous until a breaking agent is introduced to break the cross-linked polymer.
2 Water sensitivity refers to when a formation's physicochemical properties are affected in the presence of water. An
example of a water sensitive formation would be one where the soil particles swell when water is added, reducing the
permeability of the formation.
3 Viscosity is a measure of the internal friction of fluid that provides resistance to shear within the fluid, informally
referred to as how "thick" a fluid is. For example, custard is thick and has a high viscosity, while water is runny with a low
viscosity. Sufficient viscosity is needed to create a fracture and transport proppant (Gupta and Valko. 2007). In lower-
viscosity fluids, proppant is transported by turbulent flow and requires more hydraulic fracturing fluid. Higher-viscosity
fluids allows the fluid to carry more proppant, requiring less fluid but necessitating the reduction of viscosity after the
proppant is placed fRickman et al.. 2008: Gupta and Valko. 20071.
4 A disclosure refers to all data submitted for a specific oil and gas production well for a specific fracture date.
5 Sources may differ based on whether they are referring to additives or chemicals.
5-10
-------
Chapter 5- Chemical Mixing
Table 5-1. Examples of common additives, their function, and the most frequently used
chemicals reported to FracFocus for these additives.
The list of examples of common additives was developed from information provided in multiple sources (U.S. EPA,
2015a, c; Stringfellow et al., 2014; Montgomery, 2013; Vidic et al., 2013; Spellman, 2012; GWPC and ALL
Consulting, 2009; Arthur et al., 2008; Gupta and Valko, 2007; Gidlev et al., 1989). The additive functions are based
on information the EPA received from service companies (U.S. EPA, 2013a).
Additives
Function
Chemicals reported in 20% or more of
disclosures in the EPA FracFocus 1.0 project
database for given additivea'b
Acid
Dissolves cement, minerals, and clays to
reduce clogging of the pore space
Hydrochloric acid
Biocide
Controls or eliminates bacterial growth,
which can be present in the base fluid and
may have detrimental effects on the long
term well productivity
Glutaraldehyde;
2,2-dibromo-3-nitrilopropionamide
Breaker
Reduces the designed increase in viscosity
of specialized treatment fluids such as gels
and foams after the proppant has been
placed and flowback commences to clean
up the well
Peroxydisulfuric acid diammonium salt
Clay control
Prevents the swelling and migration of
formation clays that otherwise react to
water-based fluids
Choline chloride
Corrosion
inhibitor
Protects the iron and steel components in
the wellbore and treating equipment from
corrosive fluids
Methanol; propargyl alcohol; isopropanol
Crosslinker
Increases the viscosity of base gel fluids by
connecting polymer molecules
Ethylene glycol; potassium hydroxide; sodium
hydroxide
Emulsifier
Facilitates the dispersion of one immiscible
fluid into another by reducing the interfacial
tension between the two liquids to achieve
stability
2-Butoxyethanol;
polyoxyethylene(10)nonylphenyl ether; methanol;
nonyl phenol ethoxylate
Foaming agent
Generates and stabilizes foam fracturing
fluids
2-Butoxyethanol; Nitrogen, liquid; isopropanol;
methanol; ethanol
Friction reducer
Reduces the friction pressures experienced
when pumping fluids through tools and
tubulars in the wellbore
Hydrotreated light petroleum distillates
Gelling agent
Increases fracturing fluid viscosity allowing
the fluid to carry more proppant into the
fractures and to reduce fluid loss to the
reservoir
Guar gum; hydrotreated light petroleum distillates
Iron control
agent
Controls the precipitation of iron
compounds (e.g., Fe203) from solution
Citric acid
5-11
-------
Chapter 5 - Chemical Mixing
Additives
Function
Chemicals reported in 20% or more of
disclosures in the EPA FracFocus 1.0 project
database for given additivea'b
Nonemulsifier
Separates problematic emulsions generated
within the formation
Methanol; isopropanol; nonyl phenol ethoxylate
pH control
Affects the pH of a solution by either
inducing a change (pH adjuster) or
stabilizing and resisting change (buffer) to
achieve desired qualities and optimize
performance
Carbonic acid, dipotassium salt; potassium
hydroxide; sodium hydroxide; acetic acid
Resin curing
agents
Lowers the curable resin coated proppant
activation temperature when bottom hole
temperatures are too low to thermally
activate bonding
Methanol; nonyl phenol ethoxylate; isopropanol;
alcohols, C12-14-secondary, ethoxylated
Scale inhibitor
Controls or prevents scale deposition in the
production conduit or completion system
Ethylene glycol; methanol
Solvent
Controls the wettability of contact surfaces
or prevents or breaks emulsions1
Hydrochloric acid
a Chemicals (excluding water and quartz) listed in the EPA FracFocus 1.0 project database in more than 20% of disclosures for a
given purpose when that purpose was listed as used on a disclosure (U.S. EPA, 2015c). These are not necessarily the active
ingredients for the purpose, but rather are listed as being commonly present for the given purpose. Chemicals may be disclosed
for more than a single purpose (e.g., 2-butoxyethanol is listed as being used as an emulsifier and a foaming agent).
b Analysis considered 32,885 disclosures and 615,436 ingredient records that met selected quality assurance criteria, including:
completely parsed (parsing is the process of analyzing a string of symbols to identify and separate various components); unique
combination of fracture date and API well number; fracture date between January 1, 2011, and February 28, 2013; valid CASRN;
valid concentrations; and valid purpose. Disclosures that did not meet quality assurance criteria (5,645) or other, query-specific
criteria were excluded from analysis.
5.3.1 Water-Based Fracturing Fluids
The advantages of water-based fracturing fluids are low cost, ease of mixing, and ability to recover
and reuse the water. The disadvantages are that they have low viscosity, they create narrow
fractures, and they may not provide optimal performance in water-sensitive formations
(Montgomery. 2013: Gupta and Valko. 20071 (Section 5.3.2). Water-based fluids can be as simple as
water with a few additives to reduce friction, such as "slickwater," or as complex as water with
crosslinked polymers, clay control agents, biocides, and scale inhibitors fSpellman. 20121. (See
Figure 5-4 for a slickwater example.)
Gels may be added to water-based fluids to increase viscosity, which assists with proppant
transport and results in wider fractures. Gelling agents include natural polymers, such as guar,
starches, and cellulose derivatives, which require the addition of biocide to minimize bacterial
growth (Spellman. 2012: Gupta and Valko. 2007). Gels may be linear or crosslinked. Crosslinking
1 Wettability is the ability of a liquid to maintain contact with a solid surface. When wettability is high, a liquid droplet will
lie flat across a surface, maximizing the area of contact between the liquid and the solid. When wettability is low, a liquid
droplet will approach a spherical shape, minimizing the area of contact between the liquid and solid.
5-12
-------
Chapter 5- Chemical Mixing
increases viscosity without adding more gel. Gelled fluids require the addition of a breaker, which
breaks down the gel after it carries in the proppant, to reduce fluid viscosity to facilitate fluid
flowing back after treatment fSpellman. 2012: Gupta and Valko. 20071. The presence of residual
breakers may make it difficult to reuse recovered water (Montgomery. 2013).
5.3.2 Alternative Fracturing Fluids
Alternative hydraulic fracturing fluids can be used for water sensitive formations (i.e., formations
where permeability is reduced when water is added) or as dictated by production goals
fHalliburton. 19881. Examples of alternative fracturing fluids include acid-based fluids; non-
aqueous-based fluids; energized fluids, foams, or emulsions; viscoelastic surfactant fluids; gels;
methanol; and other unconventional fluids (Montgomery. 2013: Saba etal.. 2012: Gupta and Hlidek.
2009: Gupta and Valko. 2007: Halliburton. 19881.
Acid fracturing is generally used in carbonate formations without the use of a proppant. Fractures
are initiated with a hydraulic fracturing fluid, and acid (gelled, foamed, or emulsified) is added to
irregularly etch the wall of the fracture. The etching serves to prop open the formation, for a high-
conductivity fracture fSpellman. 2012: Gupta and Valko. 20071.
Non-aqueous fluids, like petroleum distillates and propane, are used in water-sensitive formations.
Non-aqueous fluids may also contain additives, such as gelling agents, to improve performance
f Gupta and Valko. 20071. The use of non-aqueous fluids has decreased due to safety concerns, and
because water-based and emulsion fluid technologies have improved fMontgomery. 2013: Gupta
and Valko. 2007). Methanol, for example, was previously used as a base fluid in water-sensitive
reservoirs beginning in the early 1990s, but was discontinued in 2001 for safety concerns and cost
(Saba etal.. 2012: Gupta and Hlidek. 2009: Gupta and Valko. 2007). Methanol is still widely used as
an additive or in additive mixtures in hydraulic fracturing fluid formulations.
Energized fluids, foams, and emulsions minimize fluid leakoff in low pressure targeted geologic
formations, have high proppant-carrying capacity, improve fluid recovery, and are sometimes used
in water-sensitive formations (Barati and Liang. 2014: Gu and Mohantv. 2014: Spellman. 2012:
Gupta and Valko. 2007: Martin and Valko. 20071.1 However, these treatments tend to be expensive,
can require high pressure, and pose potential health and safety concerns (Montgomery. 2013:
Spellman. 2012: Gupta and Valko. 2007). Energized fluids (see Figure 5-4 for an example of an
energized fluid composition) are mixtures of liquid and gas (Patel etal.. 2014: Montgomery. 2013).
Nitrogen (N2) or carbon dioxide (CO2), the gases used, make up less than 53% of the fracturing fluid
volume, typically ranging from 20% to 30% by volume fMontgomerv. 2013: Gupta and Valko. 2007:
Mitchell. 1970). Energized foams are liquid-gas mixtures, with nitrogen or carbon dioxide gas
comprising more than 53% of the fracturing fluid volume, with a typical range of 65% to 80% by
volume fMontgomerv. 2013: Mitchell. 19701. Emulsions are liquid-liquid mixtures, typically a
1 Leakoff is the fraction of the injected fluid that infiltrates into the formation (e.g., through an existing natural fissure]
and is not recovered during production fHconomides etal.. 20071 See Chapter 6, Section 6.3 for more discussion on
leakoff.
5-13
-------
Chapter 5 - Chemical Mixing
hydrocarbon (e.g., condensate or diesel) with water.1 Both water-based fluids, including gels, and
non-aqueous fluids can be energized fluids or foams.
Foams and emulsions break easily using gravity separation and are stabilized by using additives
such as foaming agents (Gupta and Valko. 20071. Emulsions may be used to stabilize active chemical
ingredients or to delay chemical reactions, such as the use of carbon dioxide-miscible, non-aqueous
fracturing fluids to reduce fluid leakoff in water-sensitive formations fTavlor etal.. 20061.
Other types of fluids not addressed above include viscoelastic surfactant fluids, viscoelastic
surfactant foams, crosslinked foams, liquid carbon dioxide-based fluid, and liquid carbon dioxide-
based foam fluid, and hybrids of other fluids fKing. 2010: Brannon et al.. 2009: Curtice et al.. 2009:
Tudor etal.. 2009: Gupta and Valko. 2007: Coulter etal.. 2006: Bover etal.. 2005: Fredd etal.. 2004:
MacDonald et al.. 20031.
5.3.3 Tracers
Some chemicals are added to the fluid to act as tracers. Tracers are added to hydraulic fracturing
fluid to assess the efficiency of fracturing and proppant placement. As an example, the efficiency of
oil production from multistage fracturing was assessed by using 17 oil soluble tracers. Each tracer
was used to assess production from a specific interval of the well fCatlettetal.. 20131. although the
specific compounds used were not identified (Table 5-2). Chemical classes of tracers and individual
examples show a range of compounds employed including both inorganic and organic, and
including radioactive elements, although only a few specific chemicals have been revealed. Of these,
examples are proppant tracers and fluorocarbons. Although radioactive fluids have also been used
for proppant tracing, a commonly-used approach has the short-half-life elements Antimony124,
Iridium192, and Scandium46 bound to the proppants and gamma emissions are subsequently
measured by a neutron-logging device (Sonnenfield et al.. 2016: Odegard etal.. 2015: Lowe etal..
2013: Osborn and Mcintosh. 2011: McDaniel etal.. 2010).23 Of the organic tracers, 14 fluorinated
organics have been identified through an analysis of FracFocus 2.0 disclosures (Konschnik and
Davalu. 20161. Three fluorinated tracers and Antimony124 were identified in produced water
(Maguire-Bovle and Barron. 2014) (Appendix Table H-4).
Table 5-2. Classes and specifically identified examples of tracers used in hydraulic fracturing
fluids.
Class
Specific Chemical3
Reference
Thiocyanates (SCN )
ND
Dugstad (2007)
Fluorobenzoic acids
ND
Dugstad (2007)
1 Diesel is a mixture typically of C8 to C21 hydrocarbons. The shorthand "C8" is used to represent a hydrocarbon with 8
carbons. Thus "C21" represents a hydrocarbon with 21 carbons. Octane has 8 carbons and is thus a C8, and is a
component of gasoline.
2 Antimony124: 60.2 days. Iridium192 : 74 days, Scandium46:83.8 days.
3 Gadolinium155 and Gadolinium157 have been suggested as bound proppant tracers because of their high-gamma-capture
cross-sections ("Liu etal.. 20151.
5-14
-------
Chapter 5- Chemical Mixing
Class
Specific Chemical3
Reference
Radioactive tracers
Tritiated Water
Dugstad (2007)
Tritiated Methanol
Dugstad (2007)
Antimony124
Silber et al. (2003)
Iridium192
Silber et al. (2003)
Scandium46
Silber et al. (2003)
Fluorocarbons
2,2,3,3,4,4,4-heptafluorobutyl undecylate
Maguire-Bovle and Barron (2014)
2,3,4-Trifluorobenzoic acid
Konschnikand Davalu (2016)
2,4,5-Trifluorobenzoic acid
Konschnikand Davalu (2016)
2,4-Difluorobenzoic acid
Konschnikand Davalu (2016)
2,6-Difluorobenzoic acid
Konschnikand Davalu (2016)
2-Chloro-4-fluorobenzoic acid
Konschnikand Davalu (2016)
2-Fluorobenzoic acid
Konschnikand Davalu (2016)
2-Trifluoromethylbenzoic acid
Konschnikand Davalu (2016)
3-Trifluoromethylbenzoate
Konschnikand Davalu (2016)
4-(Trifluoromethyl)benzoic acid
Konschnikand Davalu (2016)
4-Chloro-2-fluorobenzoic acid
Konschnikand Davalu (2016)
4-fluoro-2-(trifluoromethyl)benzoic acid
Konschnikand Davalu (2016)
4-Fluoro-3-(trifluomethyl)benzoic acid
Konschnikand Davalu (2016)
Benzoic acid, 3,5-difluoro-
Konschnikand Davalu (2016)
c/s-4-ethyl-5-octyl-2,2-b/s(trifluoromethyl)-l,3
dioxolane
Maguire-Bovle and Barron (2014)
p-Fluorobenzoic acid
Konschnikand Davalu (2016)
tri-fluoromethyl tetradeculate
Maguire-Bovle and Barron (2014)
Oil soluble alkyl esters
ND
Deans (2007)
Unstable emulsions
ND
Catlett et al. (2013)
Controlled-release
polymers and solid
tracers
ND
Salman et al. (2014)
a ND = none disclosed.
5-15
-------
Chapter 5 - Chemical Mixing
A different set of tracers have been proposed for identifying environmental impacts from hydraulic
fracturing fluids fKurose. 20141. These tracers are designed so that the fluids from individual wells
are identifiable while having no environmental impact themselves. DNA and nanoparticles with
magnetic properties made specifically for each well have been proposed for this purpose fKurose.
20141.
5.3.4 Proppants
Proppants are small particles carried down the well and into fractures by hydraulic fracturing fluid.
They hold the fractures open after the injection pressure has been released and the hydraulic
fracturing fluid has been removed (Brannon and Pearson. 20071. The propped fractures provide a
path for the hydrocarbon to flow from the reservoir. The EPA's analysis of FracFocus 1.0 data
showed that 98% of disclosures reported sand as the proppant, making sand (i.e., quartz) the most
commonly reported proppant (U.S. EPA. 2015a). Other proppants include man-made or specially
engineered particles, such as high-strength ceramic materials or sintered bauxite (Schlumberger.
2014: Brannon and Pearson. 20071. Proppant types can be used individually or in combinations.
5.3.5 Example Hydraulic Fracturing Fluids
There is no standard composition of hydraulic fracturing fluid used across the United States, and
the literature does not present any typical hydraulic fluid composition. In Figure 5-4, we present
two examples of hydraulic fracturing fluid mixtures based on analyses conducted on the EPA
FracFocus 1.0 project database fU.S. EPA. 2015cl. These examples representtwo different types of
fluids used at two different wells. The first is a slickwater, and the second is an energized fluid.1
Details of each fluid are presented in the figure along with pie charts of their composition, as given
by maximum percent by mass of the total hydraulic fracturing fluid.
The first hydraulic fracturing fluid (Figure 5-4a), the slickwater, is composed of 87% water, 13%
sand, and 0.05% chemicals, by mass. The fluid is 71% fresh water and 16% reused produced water,
with a total water volume of 4,763,000 gal (18,030,000 L). The chemical composition consists of six
different additive types (acid, friction reducer, biocide, scale inhibitor iron control, and corrosion
inhibitor) and a total of 13 chemicals.
The second hydraulic fracturing fluid (Figure 5-4b), the energized fluid, is more complex and
consists of 58% water, 28% nitrogen gas, 13% sand, and 1.5% additives, by mass, with a total water
volume of 105,000 gal (397,000 L). The hydraulic fracturing fluid composition consists of 10
additives (acid, surfactant, foamer, corrosion inhibitor, biocide, friction reducer, breaker, scale
inhibitor, iron control, and clay stabilizer) and a total of 28 chemicals.
1A slickwater is a hydraulic fracturing fluid designed to have a low viscosity to allow pumping at high rates. The critical
additive in a slickwater is the friction reducer, which makes the fluid "slick."
5-16
-------
Chapter 5 - Chemical Mixing
(a) Slickwater
0.01% Friction Reducer (1)
Bradford County, Pennsylvania
Well depth = 7,255 feet
Total water volume = 4,763,000 gallons
16%* Reused
Wastewater
13% Sand
0.03% Acid 1
71% Fresh Water
0.006% Biocide (3)
0.002% Scale Inhibitor (2)
0.0009% Iron
Control (1)
0.0006% Corrosion
Inhibitor (5)
0.05% Additives (13 Chemicals)
*Maximum percent by mass of the total hydraulic fracturing fluid
28%* Nitrogen (gas)
58% Water
0.006%
0.004%
0.1% Acid (1)
1.2% Clay
Control (1)
0.08% Surfactant (3)
0.05% Foamer (2)
0.03% Corrosion
Inhibitor (11)
Biocide (4)
Friction
Reducer (1)
Breaker(1)
Scale
Inhibitor (4)
Iron Control (1)
1.5% Additives (28 Chemicals)
* Maximum percent by mass of the total hydraulic fracturing fluid
(b) Energized Fluid
Rio Arriba County, New Mexico
Well depth = 7,640 feet
Total water volume = 105,000 gallons
Figure 5-4. Example hydraulic fracturing fluids.
Example compositions of (a) slickwater and (b) energized fluid. The base fluid and proppants are on the left, and
the additive breakdown is on the right. The number in parentheses represents the number of chemicals in that
additive. See Table 5-1 for the function of different additives and the most common chemicals in those additives
reported as based on the analysis of the EPA FracFocus 1.0 project database (U.S. EPA, 2015c).
These two examples give an idea of the difference in the compositions of two example hydraulic
fracturing fluids. These compositions are the final mixture as if the entire fluid were mixed at once;
they are generally not the actual composition at any given point in time. These compositions
provide the potential composition of a spilled hydraulic fracturing fluid during the chemical mixing
stage. Any of these ingredients (e.g., biocide) could be released by itself or mixed with the base fluid
with other additives. The variability of hydraulic fracturing fluids from well to well and site to site
makes it difficult to assess the potential of hydraulic fracturing additive or fluid release.
5.4 Frequency and Volume of Hydraulic Fracturing Chemical Use
This section highlights the different chemicals used in hydraulic fracturing fluids and discusses the
frequency and volume of use. Using the EPA FracFocus 1.0 project database (Text Box 5-1], we
focus our analysis on the individual chemicals that are used as ingredients in additive formulations,
rather than on the complete mixture of chemicals that may be present in a hydraulic fracturing
fluid. Operators can report information about well location, date of operations, and water and
chemical use to the FracFocus registry. Chemicals are reported in FracFocus by using the chemical
5-17
-------
Chapter 5 - Chemical Mixing
name and the Chemical Abstract Services Registration Number (CASRN), which is a unique number
identifier for every chemical.1 The information on specific chemicals, particularly those most
commonly used, can be used to assess potential impacts on drinking water resources. The volume
of chemicals stored on-site provides information on the potential volume of a chemical spill.
Text Box 5-1. The FracFocus Registry and EPA FracFocus Report.
The Ground Water Protection Council (GWPC) and the Interstate Oil and Gas Compact Commission (IOGCC)
developed a national hydraulic fracturing chemical registry, FracFocus fwww.fracfocus.orgl Well operators
can use the registry to disclose information about chemicals and water they use during hydraulic fracturing.
As part of the EPA's Study of the Potential Impacts of Hydraulic Fracturing for Oil and Gas on Drinking Water
Resources, the EPA published the report titled Analysis of Hydraulic Fracturing Fluid Data from the FracFocus
Chemical Disclosure Record Registry 1.0 (U.S. EPA. 2015a). For this report, the EPA accessed data from
FracFocus 1.0 from January 1, 2011 to February 28, 2013, which included more than 39,000 disclosures
(records of well data) in 20 states that had been submitted by operators prior to March 1,2013.
Accompanying the U.S. EPA f2015a1 report is the published EPA FracFocus 1.0 project database, which. It
supported analyses of FracFocus chemical and water use data (U.S. EPA. 2015c). and a report describing the
details of data management for development of the project database (U.S. EPA. 2015b).
Submission to FracFocus was initially voluntary and varied from state to state. During the timeframe covered
in the EPA FracFocus 1.0 report (January 2011 to February 2013), six of the 20 states with data submitted to
FracFocus and included in the EPA FracFocus 1.0 project database began requiring operators to disclose
chemicals used in hydraulic fracturing fluids to FracFocus (Colorado, North Dakota, Oklahoma, Pennsylvania,
Texas, and Utah). Three other states started requiring disclosure to either FracFocus or the state (Louisiana,
Montana, and Ohio), and five states required or began requiring disclosure to the state (Arkansas, Michigan,
New Mexico, West Virginia, and Wyoming). Alabama, Alaska, California, Kansas, Mississippi, and Virginia did
not have reporting requirements during the period of the EPA's study.
The EPA's analysis may or may not be nationally representative. Disclosures from the five states reporting the
most disclosures to FracFocus (Texas, Colorado, Pennsylvania, North Dakota, and Oklahoma) comprise over
78% of the disclosures in the database; nearly half (47%) of the disclosures are from Texas. Thus, data from
these states are most heavily represented in the EPA's analyses.
A disclosure reports the total water volume (in gallons) and the chemicals used in the fluid (as maximum
ingredient concentration by mass both in the additive and in the hydraulic fracturing fluid). The actual mass
of the chemicals used in the fluid are not reported. The fluid composition reported in the disclosure does not
necessarily reflect the actual composition of the fluid at any time. Rather, the disclosure represents what the
total composition of the fluid would be if all chemicals used were mixed together at their maximum reported
concentration.
The EPA summarized information on the locations of the wells in the disclosures, water volumes used, and
the frequency of use and maximum ingredient concentrations of the chemicals in the additives and the
hydraulic fracturing fluid. Additional information can be found in the EPA FracFocus 1.0 report (U.S. EPA.
2015a) and in the EPA FracFocus 1.0 project database (U.S. EPA. 2015c).
1A CASRN and chemical name combination identify a chemical substance, which can be a single chemical (e.g.,
hydrochloric acid, CASRN 7647-01-0] or a mixture of chemicals (e.g., hydrotreated light petroleum distillates (CASRN
64742-47-8], a complex mixtures of C9 to C16 hydrocarbons]. For simplicity, we refer to both pure chemicals and
chemical substances that are mixtures, which have a single CASRN, as "chemicals."
5-18
-------
Chapter 5- Chemical Mixing
The EPA compiled a list of 1,084 chemicals with unique CASRNs reported as used in the hydraulic
fracturing process between 2005 and 2013 (full list, methodology, and details on sources in
Appendix H).1 These chemicals fall into different chemical classes and include 455 organic
chemicals, 258 inorganic chemicals, and 361 organic mixtures or polymers. The chemical classes of
commonly used hydraulic fracturing chemicals include but are not limited to:
• Acids (e.g., hydrochloric acid, peroxydisulfuric acid, acetic acid, citric acid);
• Alcohols (e.g., methanol, isopropanol, ethylene glycol, propargyl alcohol, ethanol);
• Aromatic hydrocarbons (e.g., benzene, naphthalene, heavy aromatic petroleum solvent
naphtha);
• Bases (e.g., sodium hydroxide, potassium hydroxide);
• Hydrocarbon mixtures (e.g., petroleum distillates);
• Polysaccharides (e.g., guar gum);
• Surfactants (e.g., poly(oxy-l,2-ethanediyl)-nonylphenyl-hydroxy, 2-butoxyethanol); and
• Salts (e.g., sodium chlorite, dipotassium carbonate).
Further details on these chemicals and their associated hazards are presented in Chapter 9.
All of the sources of information used to compile the list of chemicals found in hydraulic fracturing
fluids (Appendix H) relied on reported use of those chemicals. In some cases, analysis of produced
water samples by advanced analytical methods could provide information on suspected hydraulic
fracturing additives, but other sources for the chemicals need careful consideration (Hoelzer etal..
20161. These sources include chemicals originating from components of the well, lab
contamination, or subsurface reaction. We limit our discussion of hydraulic fracturing fluid
chemicals to those directly reported as used.
An additional complication in providing an assessment on the use of chemicals in hydraulic
fracturing is that companies can withhold reporting chemicals to the FracFocus registry by claiming
that a chemical is Confidential Business Information (CBI). The use of CBI is to protect proprietary
information, such as trade secrets. Details on CBI are provided in Text Box 5-2.
1 The EPA used eight different sources to identify chemicals used in hydraulic fracturing fluids. This included the EPA
FracFocus report fU.S. EPA. 2015al and seven other sources fU.S. EPA. 2013a: Colborn etal.. 2011: House of
Representatives. 2011: NYSDEC. 2011: PA DEP. 2010a: U.S. EPA. 2004a: Material Safety Data Sheets 1.
5-19
-------
Chapter 5 - Chemical Mixing
Text Box 5-2. Confidential Business Information (CBI).
This assessment relies in large part upon information provided to the EPA or to other organizations. The
submitters (e.g., businesses that operate wells or perform hydraulic fracturing services) may view some of
the information as confidential business information (CBI) and accordingly asserted CBI claims to protect it.
Information deemed to be CBI may include trade secrets or other proprietary business information entitled to
confidential treatment under Exemption 4 of the Freedom of Information Act (FOIA) and other applicable
laws. The FOIA and EPA's CBI regulations may allow for information claimed as CBI provided to the EPA to be
withheld from the public, including in this document. In practical terms, when a well operator claims CBI for a
specific chemical, they do not report the name or CASRN for that chemical in the disclosure submitted to the
FracFocus registry (see Text Box 5-1 for information on FracFocus).
The EPA evaluated data from FracFocus, a national hydraulic fracturing chemical registry used and relied
upon by some states, industry groups, and non-governmental organizations, as described in Text Box 5-1. A
company submitting a disclosure to FracFocus may choose to not report the identity of a chemical it
considers CBI. More than 70% of disclosures contained at least one chemical claimed as CBI and 11% of all
chemicals were claimed as CBI. Of the disclosures containing CBI chemicals, there were an average of five CBI
chemicals per disclosure (U.S. EPA. 2015a). Rates of withholding chemical information (designating a
chemical as CBI) have increased from 11% in the 2011 to early 2013 time period of the EPA report, to 16.5%
across the 2011 to early 2015 time period in another study using FracFocus data, with 92% of FracFocus 2.0
disclosures including at least one chemical claimed as CBI (Konschnik and Davalu. 2016). When a chemical is
claimed as CBI, there is no public means of accessing information on these chemicals (e.g., CASRN, name).
Sometimes a CBI entry will provide the chemical family (Appendix H).
Consistent with the EPA's Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water
Resources (U.S. EPA. 2011d). data were submitted by nine service companies to the EPA regarding chemicals
used in hydraulic fracturing from 2005 to 2009. These data were separate from the EPA FracFocus 1.0 project
database. The data were submitted directly to the EPA, with the actual names and CASRNs of any chemicals
the company considered CBI. This included a total of 381 CBI chemicals, with a mean of 42 CBI chemicals per
company and a range of 7 to 213 fU.S. EPA. 2013al Of these 381 chemicals, some companies only provided a
generic chemical name and no CASRN, some provided neither a chemical name or CASRN, while others
provided a CASRN and a specific chemical name. This resulted in 80 CASRNs/chemical names on this CBI list.
Table H-3 lists generic chemical names, which may have been designed to mask CBI chemical names given to
the EPA. The EPA does not know if the 381 chemicals represent 381 unique chemicals or if there are
duplicates on this list.
The prevalence of CBI claims in the EPA FracFocus 1.0 project database limits completeness of the data set
and introduces uncertainty. Ideally, all data would be available on all chemicals to do a full assessment.
5.4.1 National Frequency of Use of Hydraulic Fracturing Chemicals
A total of 692 chemicals were identified in the EPA FracFocus 1.0 project database that were
reported as used in hydraulic fracturing from January 1, 2011, to February 28, 2013. This
information comes from a total of 35,957 disclosures with chemical data in the database (U.S. EPA.
2015a).1
1 Chemicals may be pure chemicals (e.g., methanol] or chemical mixtures (e.g., hydrotreated light petroleum distillates],
and they each have a single CASRN. Of these 692 chemicals, 598 had valid fluid and additive concentrations (34,675
disclosures). Sixteen chemicals were removed, because they were minerals listed as being used as proppants. This left a
total of 582 chemicals (34,344 disclosures).
5-20
-------
Chapter 5- Chemical Mixing
Table 5-3 presents the 35 chemicals (5% of all chemicals identified in the EPA's study) that were
reported as ingredients in additives in at least 10% of the EPA FracFocus 1.0 project database
disclosures for all states reporting to FracFocus 1.0 during this time fU.S. EPA. 2015cl. This table
also includes the top four additives in which the given chemical was reported as an ingredient
Table 5-3. Chemicals identified in the EPA FracFocus 1.0 project database in 10% or more
disclosures, with the percent of disclosures for which each chemical is reported as an
ingredient in an additive and the top four reported additives for which the chemical is used.
If a chemical is reported to be used in less than four additives, the table presents all additives (U.S. EPA, 2015c).
No.
Chemical name3
CASRN
Percent of
disclosures'3
Additives in which chemical is used
(four most common, EPA FracFocus 1.0
project database)0
1
Methanol
67-56-1
72%
Corrosion inhibitors, surfactants, non-
emulsifiers, scale control
2
Hydrotreated light
petroleum distillatesd
64742-47-8
65%
Friction reducers, gelling agents and gel
stabilizers, crosslinkers and related additives,
viscosifiers
3
Hydrochloric acid
7647-01-0
65%
Acids, solvents, scale control, clean
perforations
4
Water6
7732-18-5
48%
Acids, biocides, clay control, scale control
5
Isopropanol
67-63-0
47%
Corrosion inhibitors, non-emulsifiers,
surfactants, biocides
6
Ethylene glycol
107-21-1
46%
Crosslinkers and related additives, scale
control, corrosion inhibitors, friction reducers
7
Peroxydisulfuric acid,
diammonium salt
7727-54-0
44%
Breakers and breaker catalysts, oxidizer,
stabilizers, clean perforations
8
Sodium hydroxide
1310-73-2
39%
Crosslinkers and related additives, biocides, pH
control, scale control
9
Guar gum
9000-30-0
37%
Gelling agents and gel stabilizers, viscosifiers,
clean perforations, breakers and breaker
catalysts
10
Quartz6
14808-60-7
36%
Breakers and breaker catalysts, gelling agents
and gel stabilizers, scale control, crosslinkers
and related additives
11
Glutaraldehyde
111-30-8
34%
Biocides, surfactants, crosslinkers and related
additives, sealers
12
Propargyl alcohol
107-19-7
33%
Corrosion inhibitors, inhibitors, acid inhibitors,
base fluid
13
Potassium hydroxide
1310-58-3
29%
Crosslinkers and related additives, pH control,
friction reducers, gelling agents and gel
stabilizers
5-21
-------
Chapter 5 - Chemical Mixing
No.
Chemical name3
CASRN
Percent of
disclosures'3
Additives in which chemical is used
(four most common, EPA FracFocus 1.0
project database)0
14
Ethanol
64-17-5
29%
Surfactants, biocides, corrosion inhibitors, fluid
foaming agents and energizers
15
Acetic acid
64-19-7
24%
pH control, iron control agents, acids, gelling
agents and stabilizers
16
Citric acid
77-92-9
24%
Iron control agents, scale control, gelling
agents and gel stabilizers, pH control
17
2-Butoxyethanol
111-76-2
21%
Surfactants, corrosion inhibitors, non-
emulsifiers, fluid foaming agents and
energizers
18
Sodium chloride
7647-14-5
21%
Breakers/breaker catalysts, friction reducers,
scale control, clay control
19
Solvent naphtha,
petroleum, heavy arom.f
64742-94-5
21%
Surfactants, non-emulsifiers, inhibitors,
corrosion inhibitors
20
Naphthalene
91-20-3
19%
Surfactants, non-emulsifiers, corrosion
inhibitors, inhibitors
21
2,2-Dibromo-3-
nitrilopropionamide
10222-01-2
16%
Biocides, clean perforations, breakers and
breaker catalysts, non-emulsifiers
22
Phenolic resin
9003-35-4
14%
Proppants, biocides, clean perforations, base
fluid
23
Choline chloride
67-48-1
14%
Clay control, clean perforations, base fluid,
biocides
24
Methenamine
100-97-0
14%
Proppants, crosslinkers and related additives,
biocides, base fluid
25
Carbonic acid,
dipotassium salt
584-08-7
13%
pH control, proppants, acids, surfactants
26
1,2,4-Trimethylbenzene
95-63-6
13%
Surfactants, non-emulsifiers, corrosion
inhibitors, inhibitors
27
Quaternary ammonium
compounds, benzyl-C12-
16-alkyldimethyl,
chlorides8
68424-85-1
12%
Biocides, non-emulsifiers, corrosion inhibitors,
scale control
28
Poly(oxy-l,2-ethanediyl)-
nonylphenyl-hydroxy
(mixture)h
127087-87-0
12%
Surfactants, friction reducers, non-emulsifiers,
inhibitors
29
Formic acid
64-18-6
12%
Corrosion inhibitors, acids, inhibitors, pH
control
5-22
-------
Chapter 5- Chemical Mixing
No.
Chemical name3
CASRN
Percent of
disclosures'3
Additives in which chemical is used
(four most common, EPA FracFocus 1.0
project database)0
30
Sodium chlorite
7758-19-2
11%
Breakers/breaker catalysts, biocides, oxidizer,
proppants
31
Nonyl phenol ethoxylate
9016-45-9
11%
Non-emulsifiers, resin curing agents,
activators, friction reducers
32
Tetrakis(hydroxymethyl)
phosphonium sulfate
55566-30-8
11%
biocides, scale control, clay control
33
Polyethylene glycol
25322-68-3
11%
Biocides, non-emulsifiers, surfactants, clay
control
34
Ammonium chloride
12125-02-9
10%
Friction reducers, crosslinkers and related
additives, scale control, clay control
35
Sodium persulfate
7775-27-1
10%
Breakers and breaker catalysts, oxidizer, pH
control
a Chemical refers to chemical substances with a single CASRN; these may be pure chemicals (e.g., methanol) or chemical
mixtures (e.g., hydrotreated light petroleum distillates). Chemical names are sometimes different between FracFocus 1.0 and
Appendix H, though they will have the same CASRN.
b Analysis considered 34,675 disclosures and 676,376 ingredient records that met selected quality assurance criteria, including:
completely parsed; unique combination of fracture date and API well number; fracture date between January 1, 2011, and
February 28, 2013; valid CASRN; and valid concentrations. Disclosures that did not meet quality assurance criteria (3,855) or
other, query-specific criteria were excluded from analysis.
c Analysis considered 32,885 disclosures and 615,436 ingredient records that met selected quality assurance criteria, including:
completely parsed; unique combination of fracture date and API well number; fracture date between January 1, 2011, and
February 28, 2013; valid CASRN; valid concentrations; and valid purpose. Disclosures that did not meet quality assurance
criteria (5,645) or other, query-specific criteria were excluded from analysis.
d Hydrotreated light petroleum distillates (CASRN 64742-47-8) is a mixture of hydrocarbons, in the C9 to C16 range.
0 Quartz (CASRN 14808-60-7), the proppant most commonly reported, and water were also reported as an ingredient in other
additives (U.S. EPA, 2015a).
f Heavy aromatic solvent naphtha (petroleum) (CASRN 64742-94-5) is mixture of aromatic hydrocarbons in the C9 to C16 range.
5 Quaternary ammonium compounds, benzyl-C12-16-alkyldimethyl, chlorides (CASRN 68424-85-1) is a mixture of benzalkonium
chloride with carbon chains between 12 and 16.
h Poly(oxy-l,2-ethanediyl)-nonylphenyl-hydroxy (mixture) (CASRN 127087-87-0) is mixture with varying length ethoxy links.
There is no single chemical used in all hydraulic fracturing fluids across the United States. Methanol
is the most commonly used chemical, reported at 72.1% of wells in the EPA FracFocus 1.0 project
database and is associated with 33 types of additives, including corrosion inhibitors, surfactants,
non-emulsifiers, and scale control (U.S. EPA. 2015c).1 Table 5-3 also shows the variability in
different chemicals included in the EPA FracFocus 1.0 project database. The percentage of
disclosures reporting a given chemical suggests the likelihood of that chemical's use at a site. Only
three chemicals (methanol, hydrotreated light petroleum distillates, and hydrochloric acid) were
used at more than half of the sites nationwide, and only 12 were used at more than one-third.
1 The number of additives may be an overestimate due to parsing issues. The true number of additives may be smaller.
5-23
-------
Chapter 5 - Chemical Mixing
In addition to providing information on frequency of use, the EPA FracFocus 1.0 project database
provides the maximum concentration by mass of a given chemical in an additive. For example,
methanol is the most frequently reported chemical. The median value for the maximum mass
concentration reported for an additive in FracFocus disclosures is 30%, with a range of 0.44% to
100% (5th to 95th percentile).1 Thus, methanol is generally used as part of a mixture of chemicals in
the hydraulic fracturing fluid (typically at a concentration around 30% by mass). Other times,
methanol is used as an additive in its pure form (concentration 100%). Therefore, methanol will
sometimes be stored on-site in a mixture of chemicals and other times as pure methanol. This wide
range of possible concentrations of methanol further complicates assessing the potential impact of
spills, as the properties of the fluid will depend on the different chemicals present and on their
concentrations. For all chemicals, spills of a highly concentrated chemical can have different
potential impacts than spills of dilute mixtures. For more discussion on fluid and additive chemical
composition, see Section 5.4.5.
A more recent study of FracFocus 2.0 data evaluated disclosures dating from March 9, 2011 to April
13, 2015 (96,449 disclosures) and reported 981 unique chemicals used in hydraulic fracturing
fDavalu and Konschnik. 2016: Konschnik and Davalu. 20161. The earlier, EPA study (covering the
2011 to early 2013 time period) found 692 chemicals (U.S. EPA. 2015a). Konschnik and Davalu
(2016) identified 263 new CASRNs in addition to the 1,084 identified by the EPA (Appendix H),
increasing the number of chemicals by approximately 24%. Of the new CASRNs, the only chemical
reported in more than 1% of all disclosures was Alcohols, C9-ll-iso-,C10-rich, ethoxylated
propoxylated (CASRN 154518-36-2).
The 20 most common chemicals reported in Konschnik and Davalu (2016) are similar to those
listed in Table 5-3. There are three chemicals reported on their 20 most common list that are not
included in Table 5-3. These chemicals are: sorbitan, mono-(9Z)-9-octadecenoate (CASRN 1338-43-
8, reported in 29.6% disclosures (Konschnik and Davalu. 2016) vs. 4% (U.S. EPA. 2015c).
ethoxylated C12-16 alcohols (CASRN 68551-12-2, 27.9% vs. 4%), and thiourea polymer (CASRN
68527-49-1, 24.8% vs. 8%). Ammonium chloride was on each list, but disclosures increased from
10% to 30.5%. Four chemicals in Table 5-3 were not on their 20 most frequently used list: solvent
naphtha, petroleum, heavy arom. (CASRN 64742-94-5), naphthalene (CASRN 91-20-3), 2,2-
Dibromo-3-nitrilopropionamide (CASRN 10222-01-2), and phenolic resin (CASRN 9003-35-4).
5.4.2 Nationwide Oil versus Gas
Analyses based on the EPA FracFocus 1.0 project database also can elucidate the differences
between the chemicals used during hydraulic fracturing for oil production and those used for gas
production, providing a better understanding of potential spill impacts from each. Appendix Tables
C-l and C-2 presentthe chemicals reported in at least 10% of all gas (34 chemicals) and oil (39
chemicals) disclosures nationwide.
1 For more information on how chemicals are reported to FracFocus see www.fracfocus.org and U.S. EPA f2015al
5-24
-------
Chapter 5- Chemical Mixing
Many of the same chemicals are used for oil and gas, but some chemicals are used more frequently
in oil production and others more frequently in gas.1 For example, hydrochloric acid is the most
commonly reported chemical for gas wells (73% of disclosures); it is the fifth most frequently
reported chemical for oil wells (58% of disclosures). However, both oil and gas operators each
reports using methanol in 72% of disclosures. Methanol is the most common chemical used in
hydraulic fracturing fluids at oil wells and the second most common chemical in hydraulic
fracturing fluids at gas wells.
5.4.3 State-by-State Frequency of Use of Hydraulic Fracturing Chemicals
The composition of hydraulic fracturing fluids varies from site to site. Since the impacts of hydraulic
fracturing occur locally, the potential impact depends on the chemicals used locally. We
investigated geographic variation of chemical use based on the frequency of chemicals reported to
FracFocus and included in the EPA FracFocus 1.0 project database by state (U.S. EPA. 2015c).
Appendix Table C-3 presents and ranks chemicals reported most frequently for each state (U.S. EPA.
2015c). The list of the 20 most frequently reported chemicals used in each state together include 94
unique chemicals. A total of 94 chemicals indicates some level of similarity in chemical usage among
states.2
Methanol is reported in 19 of the 20 states (95%). Alaska is the only state in which methanol is not
reported (based on the state's 20 disclosures). The percentage of disclosures reporting use of
methanol ranges from 38% (Wyoming) to 100% (Alabama, Arkansas).
Ten chemicals (excluding water) are among the 20 most frequently reported in 14 of the 20 states.
These chemicals are: methanol; hydrotreated light petroleum distillates; ethylene glycol;
isopropanol; quartz; sodium hydroxide; ethanol; guar gum; hydrochloric acid; and peroxydisulfuric
acid, diammonium salt3 These 10 chemicals are also the most frequently reported chemicals
nationwide.
This state analysis showed that methanol is used across the contiguous U.S. (not Alaska). There are
9 other chemicals that are frequently used across the United States. Beyond those, however, there
are a number of different chemicals that are used in one state more commonly than others, and
many chemicals may not be used at all in other states.
1 This separation was done solely based on whether it was an oil or gas disclosure. The analysis did not separate out
reservoir factors, such as temperature, pressure, or permeability, which may be important factors for which chemicals are
used. There is no nationwide criterion to distinguish oil wells from gas wells. Production wells often produce some of
both. A well identified as gas-producing in one place might be identified as oil-producing in another. This could affect the
distribution of chemical use among these wells.
2 The range of possible number of chemicals is from 20 to 400. If every state used the same 20 chemicals, there would be
20 different chemicals. If all 20 states each used 20 different chemicals, then there would be a total of 400 chemicals used.
3 Quartz was the most commonly reported proppant and also reported as an ingredient in other additives flJ.S. EPA.
2015a],
5-25
-------
Chapter 5 - Chemical Mixing
5.4.4 Volume of Chemical Use
Understanding the volume of chemicals used at each site is important for understanding potential
impacts of chemicals as well as potential severity of impacts on drinking water resources. The
chemical volume governs how much will be stored on-site, the types of containers required, the
total amount that could spill, and how much could end up in a drinking water resource. While the
on-site hydraulic fracturing service company has precise knowledge of the composition and volume
of chemicals stored on-site, this information is not generally publicly available. We conducted a
comprehensive review of publicly available sources and found two sources fOSHA. 2014a. b;
Sjolander et al.. 2 011) that identify specific chemicals used at a hydraulic fracturing site and
provide information on volumes. These are presented in Table 5-4. The volume of chemicals totaled
7,500 gal (28,000 L) and 14,700 gal (55,600 L) for the two sources, with a mean volume for an
individual chemical of 1,900 gal (7,200 L) and 1,225 gal (4,637 L), respectively. The range of
volumes for each chemical used is 30 to 3,690 gal (114 to 14,000 L).
Table 5-4. Example list of chemicals and chemical volumes used in hydraulic fracturing.
Volumes are for wells with an unknown number of stages and at least one perforation zone. Every well and fluid
formulation is unique. Blank cells are data not reported.
Ingredient
Examples
Siolander et al. (2011)a
OSHA (2014a. 2014b)b
Volume (gal)
or mass (lb)
Percent
overall0
Volume
(gal)
Percent by
volume
Water
4,000,000 gal
94.62
2,700,000 gal
90
Proppant
Sand
~ 1,500,000 lbd
5.26
285,300 gal
9.51
Acid
Hydrochloric acid
or muriatic acid
1,338 gal
0.03
3,690 gal
0.123
Friction reducer
Polyacrylamide,
mineral oil
2,040 gal
0.05
2,640 gal
0.088
Surfactant
Isopropanol
2,550 gal
0.085
Potassium
chloride
1,800
0.06
Gelling agent
Guar gum or
hydroxymethyl
cellulose
_e
_e
1,680
0.056
Scale inhibitor
Ethylene glycol,
alcohol, and
sodium
hydroxide
2,040 gal
0.05
1,290
0.043
pH buffer
Carbonate
330
0.011
Preservative
Ammonium
persulfate
300
0.01
5-26
-------
Chapter 5- Chemical Mixing
Ingredient
Examples
Siolander et al. (2011)a
OSHA (2014a. 2014b)b
Volume (gal)
or mass (lb)
Percent
overall0
Volume
(gal)
Percent by
volume
Crosslinker
Borate salts
_e
_e
210
0.007
Iron control
Citric acid
_e
_e
120
0.004
Corrosion
inhibitor
n,n-Dimethyl
formamide
_e
_e
60
0.002
Biocide /
antimicrobial
agent
Glutaraldehyde,
ethanol,
methanol
2,040 gal
0.05
30
0.001
Gel-breaker
Ammonium
persulfate
_e
_e
Total volume of all chemicals
7,458 gal
0.18
14,700
0.49
Individual chemical volume: mean
(full range)
1,864.5 gal
(1,338-2,040 gal)
1,225
(30 - 3,690)
a Adapted from Penn State "Water Facts" publication entitled Introduction to Hydrofracturing (Siolander et al., 2011).
Composite from two companies: Range Resources, LLC, and Chesapeake Energy, which released in July 2010 the chemistry and
volume of materials typically used in their well completions and stimulations.
b Adapted from a table generated by the Occupational Safety and Health Administration (OSHA) for use in a training module
(OSHA. 2014a. b).
c As presented in Siolander et al. (2011); does not explicitly state percent by mass or by volume.
d Siolander et al. (2011) presents proppant in pounds instead of gallons.
6 Listed as an ingredient, but no information on volume or percentage.
Because of the limited information on chemical volumes publicly available, we estimated chemical
volumes used across the nation based on the information provided in the EPA FracFocus 1.0 project
database. Figure 5-5 plots median estimated chemical volumes, ranked from high to low, with the
range of 5th to 95th percentiles. Estimated volumes used are presented for the 74 chemicals that
were reported in at least 100 disclosures in the EPA FracFocus 1.0 project database and for which
density data were available. The estimated median volumes vary widely among the different
chemicals, covering a range of near zero to 27,000 gal (98,000 L). The mean of the estimated
median volumes was 650 gal (2,500 L), and the mean of the estimated median mass was 3,200 lb
(1,500 kg) (U.S. EPA. 2015c). Mass, volume, and density data are presented in Appendix C along
with the estimation methodology and assumptions used.
With the median chemical volume, we can estimate total chemical volume for all chemicals used.
Based on the above mean of median chemical volumes of 650 gal (2,500 L) per chemical, and given
that the median number of chemicals used at a site is 14 (U.S. EPA. 2015al. an estimated 9,100 gal
(34,000 L) of chemicals may be used per well. Given that the number of chemicals per well ranges
from 4 to 28 fU.S. EPA. 2015al. the total volume of chemicals per well may range from 2,600 to
18,000 gal (9,800 to 69,000 L).
5-27
-------
Chapter 5 - Chemical Mixing
Another way to estimate total volume of chemicals per well is to use the estimated median volume
of 1.5 million gal (5.7 million L) of fluid used to fracture a well fU.S. EPA. 2015al (Chapter 4} and
assume that up to 2% of that volume consists of chemicals added to base fluid f Carter etal.. 2013:
Knappe and Fireline. 2012). resulting in up to 30,000 gal (114,000 L] of chemicals used per well.
Using the estimated volume per chemical of 650 gal (2,500 L), we can also estimate volume per
additive and extrapolate to estimate on-site chemical storage. If we assume three to five chemicals
per additive, then total volume per additive stored on-site would be approximately 1,900 to 3,200
gal (7,400 to 12,000 L). On-site containers generally store 20% to 100% more additive volume than
ultimately used (Houston etal.. 2009: Malone and Ely. 20071. This would suggest that 2,300 to
6,500 gal (8,800 to 25,000 L) per additive are stored on site.
100000 —
10000 —
1000 —
100 —
U) 10 —
Median of median volumes: 21 gal
Mean of median volumes: 650 gal
Maximum of median volumes: 27,000 gal
0.1 —
0.01 -=—
0.001 —
0.0001 —
Figure 5-5. Estimated median volumes for 74 chemicals reported in at least 100 disclosures in
the FracFocus 1.0 project database for use in hydraulic fracturing from January 1, 2011 to
February 28, 2013.
Chemicals are plotted in order of largest to smallest median volume. Shaded area represents the zone of 5% and
95% confidence limits. Derived from U.S. EPA (2015c).
5.4.5 Chemical Composition of Hydraulic Fracturing Fluids and Additives
As the hydraulic fracturing process proceeds, the composition of the fluid injected changes over
time. The overall composition of additives and hydraulic fracturing fluid may be reported by well
5-28
-------
Chapter 5- Chemical Mixing
operators to the FracFocus national registry, depending on the local disclosure requirements and
operator preference. For each chemical that is injected into a well (excluding CBI chemicals), the
maximum concentration in the resulting overall fluid and in each additive is given as maximum
percent by mass. Based on this information, we calculated the median chemical composition
reported in at least 10% of the disclosures in the EPA FracFocus 1.0 project database (Table 5-3)
and a range based on the 5th and 95thpercentile. Table 5-5 shows that some chemicals may be used
in their pure form (100% of mass in a given additive). These chemicals include: methanol,
hydrochloric acid, water, isopropanol, guar gum, citric acid, 2,2-Dibromo-3-nitrilopropionamide,
tetrakis(hydroxymethyl) phosphonium sulfate, and sodium persulfate.
Chemicals may be stored in their concentrated, pure form, resulting in the potential for spills of
concentrated volumes of these chemicals, which may increase the severity of impacts if they reach a
drinking water resource. Once chemicals are mixed with the base fluid to form the hydraulic
fracturing fluid, the chemical is diluted to much lower concentrations, which has the potential for a
less severe impact. However, a larger volume of spill could occur with smaller concentrations. The
larger volume may increase the potential for a spill reaching a drinking water resource, albeit at a
lower concentration. There is the further complication of the hazard of the associated chemicals,
since a smaller mass of a more hazardous chemical may be of more concern than a larger mass of a
less hazardous chemical (as discussed in Chapter 9). It is therefore impossible to make a general
statement without more detail on the spill characteristics, including the hazard, concentration, and
volume.
Appendix Table C-6 provides mean, median, 5th and 95th percentile mass (kg) estimates for all
reported chemicals in 100 or more disclosures in the EPA FracFocus 1.0 project database where
density information was available.
Table 5-5. Fluid and additive composition by maximum mass percent.
Median, 5th and 95th percentile maximum concentration in hydraulic fracturing fluid and in additive (percent by
mass) for the chemicals identified in the EPA FracFocus 1.0 project database in 10% or more disclosures. See Table
5-3 for percentage of disclosures and the common additives for which these chemicals are used. Analysis
considered 34,675 disclosures and 676,376 ingredient records that met selected quality assurance criteria,
including: completely parsed; unique combination of fracture date and API well number; fracture date between
January 1, 2011, and February 28, 2013; valid CASRN; and valid concentrations. Disclosures that did not meet
quality assurance criteria (3,855) or other, query-specific criteria were excluded from analysis.
EPA-standardized
chemical name
CASRN
Maximum concentration in
hydraulic fracturing fluid
(percent by mass)
Maximum concentration in
additive (percent by mass)
Median
5th
Percentile
95th
Percentile
Median
5th
Percentile
95th
Percentile
Methanol
67-56-1
0.0092
0.00011
0.12
30
0.44
100
Distillates, petroleum,
hydrotreated light
64742-47-8
0.025
0.0013
0.35
30
0.70
70
Hydrochloric acid
7647-01-0
0.15
0.0083
1.3
15
2.8
60
5-29
-------
Chapter 5 - Chemical Mixing
EPA-standardized
chemical name
CASRN
Maximum concentration in
hydraulic fracturing fluid
(percent by mass)
Maximum concentration in
additive (percent by mass)
Median
5th
Percentile
95th
Percentile
Median
5th
Percentile
95th
Percentile
Water
7732-18-5
0.53
0.00065
82
65
5.0
100
Isopropanol
67-63-0
0.0038
0.000020
0.15
20
0.30
100
Ethylene glycol
107-21-1
0.016
0.00027
0.11
30
0.59
60
Peroxydisulfuric acid,
diammonium salt
7727-54-0
0.0069
0.00010
0.064
100
0.11
100
Sodium hydroxide
1310-73-2
0.0092
0.000040
0.077
10
0.085
52
Guar gum
9000-30-0
0.16
0.0019
0.42
50
1.6
100
Quartz
14808-60-7
0.0033
0.000030
12
2.0
0.10
97
Glutaraldehyde
111-30-8
0.0072
0.00039
0.023
27
0.040
60
Propargyl alcohol
107-19-7
0.00015
0.000010
0.0028
8.0
0.0032
30
Potassium hydroxide
1310-58-3
0.0070
0
0.053
15
0.14
50
Ethanol
64-17-5
0.0034
0.000060
0.098
30
1.0
60
Acetic acid
64-19-7
0.0033
0
0.037
50
1.0
90
Citric acid
77-92-9
0.0027
0.000060
0.017
60
7.0
100
2-Butoxyethanol
111-76-2
0.0047
0
0.14
10
0.29
60
Sodium chloride
7647-14-5
0.0083
0
0.14
30
0.020
50
Solvent naphtha,
petroleum, heavy arom.
64742-94-5
0.0051
0.000020
0.035
10
0.00052
30
Naphthalene
91-20-3
0.0014
0
0.0055
5.0
0.0023
5.0
2,2-Dibromo-3-
nitrilopropionamide
10222-01-2
0.0018
0.000010
0.022
98
10
100
Phenolic resin
9003-35-4
0.12
0.0046
1.1
5.0
0.80
20
Choline chloride
67-48-1
0.062
0.00068
0.14
75
0.75
80
Methenamine
100-97-0
0.010
0
0.21
1.0
0
2.0
Carbonic acid,
dipotassium salt
584-08-7
0.039
0
0.15
60
30
60
1,2,4-Trimethylbenzene
95-63-6
0.00067
0
0.0068
1.0
0.010
20
5-30
-------
Chapter 5- Chemical Mixing
EPA-standardized
chemical name
CASRN
Maximum concentration in
hydraulic fracturing fluid
(percent by mass)
Maximum concentration in
additive (percent by mass)
Median
5th
Percentile
95th
Percentile
Median
5th
Percentile
95th
Percentile
Quaternary ammonium
compounds, benzyl-
C12-16-alkyldimethyl,
chlorides
68424-85-1
0.0019
0
0.0041
7.0
3.0
10
Poly(oxy-l,2-
ethanediyl)-
nonylphenyl-hydroxy
(mixture)
127087-87-0
0.0025
0.000010
0.0089
5.0
5.0
10
Formic acid
64-18-6
0.0021
0
0.030
60
0.11
98
Sodium chlorite
7758-19-2
0.0040
0.00018
0.037
10
5.0
30
Nonyl phenol
ethoxylate
9016-45-9
0.0088
0.000030
0.085
10
5.0
54
Tetrakis(hydroxymethyl)
phosphonium sulfate
55566-30-8
0.011
0.00025
0.065
60
0.029
100
Polyethylene glycol
25322-68-3
0.0035
0.000010
0.038
20
0.0071
70
Ammonium chloride
12125-02-9
0.0025
0.00029
0.022
10
1.5
30
Sodium persulfate
7775-27-1
0.0017
0.000020
0.022
100
100
100
5.5 Chemical Management and Spill Potential
This section provides a description of the primary equipment used in the chemical mixing and well
injection processes, along with a discussion of the spill vulnerabilities specific to each piece of
equipment Equipment breakdown or failure can trigger a spill itself, and it can also lead to a
suspension of activity and the disconnection and reconnection of various pipes, hoses, and
containers. Each manipulation of equipment poses additional potential for a spill. The EPA found
that 31% of chemical spills on or near the well pad related to hydraulic fracturing resulted from
equipment failure (U.S. EPA. 2015m). When possible, we describe documented spills, associated
with or attributed to specific pieces of equipment, in text boxes in the relevant subsections.
Equipment used in hydraulic fracturing operations typically consists of chemical storage trucks, oil
storage tanks/tanker trucks; a slurry blender; one or more high-pressure, high-volume fracturing
pumps; the main manifold; surface lines and hoses; and a central control unit (Table 5-6). There are
many potential sources for leaks and spills in this interconnected system. Furthermore, hydraulic
fracturing operations are mobile and must be assembled at each well site, and each assembly and
disassembly presents a potential for spills.
5-31
-------
Chapter 5 - Chemical Mixing
Equipment varies in age and technological advancement depending upon service company
standards and costs associated with purchase and maintenance. Older equipment may have
experienced wear and tear, which may be a factor in spills caused by equipment failure. New
equipment may be more automated, potentially reducing opportunities for human error.
Information detailing the extent of technological and age differences in fracturing equipment across
sites and operators is limited.
Table 5-6. Examples of typical hydraulic fracturing equipment and its function.
Equipment
Function
Acid transport truck
Transports acids to job sites; the truck has separate compartments for
multiple acids or additives.
Chemical storage truck
Transports chemicals to the site in separate containment units or totes.
Chemicals are typically stored on and pumped from the chemical storage
truck.
Base fluid tanks
Stores the required volume of base fluid to be used in the hydraulic
fracturing process.
Proppant storage units
Holds proppant and feeds it to the blender via a large conveyor belt.
Blender
Takes fluid (e.g., water) from the fracturing tanks and proppant (e.g., sand)
from the proppant storage unit and combines them with additives before
transferring the mixture to the fracturing pumps
High-pressure fracturing pumps
Pressurizes mixed fluids received from the blender and injected into the well.
Manifold trailer with hoses and
pipes
Serves as a transfer station for all fluids. Includes a trailer with a system of
hoses and pipes connecting the blender, the high-pressure pumps, and the
fracturing wellhead.
Fracturing wellhead or frac head
Allows fracture equipment to be attached to the well; located at the
wellhead.
Central control unit or frac van
Monitors the hydraulic fracturing job using pressure and rate data supplied
from around the job site.
While the primary equipment and layout are generally the same across well sites, the type, size, and
number of pieces of equipment may vary depending on a number of factors (Malone and Ely. 2007):
• Size and type of the fracture treatment;
o Length of well and number of stages;
o Number of wells drilled per well pad;
o Geographic location;
o Depth below surface;
o Length of the fractures;
5-32
-------
Chapter 5 - Chemical Mixing
• Volumes and types of additives, proppants, and fluids used; and
• Operating procedures of the well operator and service company (e.g., some companies
require backup systems in case of mechanical failure, while others do not).
Figure 5-6 provides a schematic diagram of a typical layout of hydraulic fracturing equipment
Water Tanks
I CO I
/ /
III II
High Pressure Pumps
/
Manifold
/ /
/\/ /\/
High Pressure Pumps
/
/ ' / '
/
/\/
/
Flowback
Tanks
*
J
Frac
Head
/
low pressure lines ~ high pressure lines
Figure 5-6. Typical hydraulic fracturing equipment layout.
This illustration shows how the various components of a typical hydraulic fracturing site fit together. The numbers
of pumps and tanks vary from site to site. Some sites do not use a hydration unit as the gel is batch mixed prior to
the treatment (Olson, 2011; BJ Services Company, 2009).
5.5.1 Storage
This section provides an overview of publicly available information on storage and containment of
chemicals used in the hydraulic fracturing process. Most public sources provide general
information on the types and sizes of containment units. While operators maintain a precise
inventory of volumes of chemicals stored and used for each site, this information is typically not
made public.
The volumes of each chemical used are based on the size and site-specific characteristics of each
fracture treatment Sites often store an excess of the design volume of chemicals for contingency
purposes, typically 20% to 100% beyond what is necessary (Houston et al.. 2009: Malone and Ely.
20071. See Text Box 5-3 for documented spills from storage units.
5-33
-------
Chapter 5 - Chemical Mixing
Text Box 5-3. Spills from Storage Units.
Of the 151 spills of chemicals, additives, or fracturing fluid discussed and evaluated in fU.S. EPA. 2015ml (see
Text Box 5-10 for more information), 54 spills were from storage units. Storage units include totes or tanks
used for storing individual chemicals or additives and larger tanks containing hydraulic fracturing fluid.
These spills resulted from equipment failure, failure of storage integrity, or human error. Sixteen of
these spills were due to failure of container integrity, which includes holes and cracks in containers,
demonstrating the importance of properly constructed and maintained storage units. The remaining spills
from storage containers resulted from human error or equipment malfunctions or had an unknown cause.
5.5.1.1 Hydraulic Fracturing Base Fluid Storage
Base fluids used in hydraulic fracturing are typically stored on-site in large volume tanks. Non-
water-based fluids may be stored in specialized containment units designed to prevent or minimize
releases. For example, nitrogen and carbon dioxide must be stored in compressed gas or cryogenic
liquid cylinders, as required by U.S. Department of Transportation (DOT) and OSHA regulations.
Due to the large volume of base fluid storage tanks (about 21,000 gal or 80,000 L) (Halliburton.
1988). uncontrolled spills could damage other storage units and equipment, which could result in
additional spills. Fresh water used as a base fluid is generally not a source of concern for spills.
Reused wastewater, brine, and non-aqueous base fluids have the potential to adversely impact
drinking water resources in the event of a spill. Chapter 7 discusses reusing hydraulic fracturing
wastewater as a base fluid and the spill/release potential on-site from pits and impoundments.
5.5.1.2 Additive Storage
Additives are typically stored on-site in the containers in which they were transported and
delivered. The additive trailer typically consists of a flatbed truck or van enclosure that holds a
number of chemical totes, described below, and is equipped with metering pumps that feed
chemicals to the blender. Depending on the size and type of the fracturing operation, there may be
one or more additive trailers per site (NYSDEC. 2015: ALL Consulting. 2012). While additives
constitute a relatively small portion of fluids used in a hydraulic fracturing fluid, additive volumes
can range from the tens to tens of thousands of gallons.
The storage totes generally remain on the transportation trailers, but they also may be unloaded
from the trailers and transferred to alternative storage areas before use. Our investigation did not
find much information on how often, when, or why these transfers occur. Additional transfers and
movement can increase the likelihood of a spill. See Text Box 5-4 for a documented spill from an
additive storage unit.
Text Box 5-4. Spill from Additive (Crosslinker) Storage Tote.
On Sept 19, 2009, during a tote transfer in Pennsylvania, a tote of crosslinker fell off a forklift spilling
approximately 15-20 gal (60 - 80 L) onto the well pad. The area was scraped clean with a backhoe, and the
waste was placed in a lined containment area (PA DEP. 2012. ID# 1845178).
5-34
-------
Chapter 5 - Chemical Mixing
The most commonly used chemical totes are 200 - 375 gal (760 - 1,420 L) capacity polyethylene
containers that may be reinforced with steel or aluminum mesh fNYSDEC. 20151. Metal containers
may also be used. The totes are typically equipped with bottom release ports, which enable direct
feed of the additives to the blending equipment fNYSDEC. 20151. Spills may occur if lines are
improperly connected to these ports or if the connection equipment is faulty.
Figure 5-7. Metal and high-density polyethylene (HDPE) additive units.
The image on the left depicts metal totes (industry source). The image on the right depicts plastic totes. Source:
NYSDEC (2011).
Certain additives require specialized containment units with added spill prevention measures. For
example, additives containing methanol may be subject to federal safety standards, and industry
has developed guidance on methanol's safe storage and handling (Methanol Institute. 2013).
Dry additives are typically transported and stored on flatbed trucks in 50 or 55 lb (23 or 25 kg)
bags, which are set on pallets containing 40 bags each fNYSDEC. 2015: UWS. 2008: Halliburton.
19881. Proppants are stored on-site in large tanks or bins with typical capacities of 350,000 to
450,000 lb (150,000 to 200,000 kg] fALL Consulting. 2012: Bf Services Company. 2009:
Halliburton. 19881.
5.5.1.3 Acid Storage
Acids are generally stored on-site in the containment units in which they are transported and
delivered. A typical acid transport truck holds up to 5,000 gal (19,000 L] of acid and can have
multiple compartments to hold different kinds of acid (Arthur et al.. 2009bl. Acids such as
hydrochloric acid and formic acid are corrosive and can be extremely hazardous in concentrated
form. Therefore, acid transport trailers and fracture tanks must be lined with chemical-resistant
coating designed to prevent leakage and must meet applicable DOT regulatory standards (pursuant
to 40 Code of Federal Regulations (CFR] 173} designed to prevent or minimize spills.
Acid fracture treatments may use thousands of gallons of acid per treatment fSpellman. 20121.
Cxiven the large volumes used, failure of containment vessels during storage or failure of
connections and hoses during pumping could result in high-volume acid spills. Details of a
documented acid spill are presented in Text Box 5-5,
5-35
-------
Chapter 5 - Chemical Mixing
Text Box 5-5. Spill of Acid from Storage Container.
In July 2014, in Oklahoma, 20,000 gal (76,000 L] of hydrochloric acid spilled from a storage container when a
flange malfunctioned. The acid spilled into a nearby alfalfa field, where it was contained with an emergency
berm (Phillips. 2014: Wertz. 20141 There is no information on how much leached into soils or if the spill
reached drinking water resources.
5.5.1.4 Gel Storage
Gels can be added to hydraulic fracturing fluid using either batch or continuous (also called "on-the-
fly") mixing systems. Gelling agents and gel slurries are stored differently on-site and can pose
different potential spill scenarios depending on whether the site is using batch or continuous
mixing processes (BT Services Company. 20091.
In a typical batch mixing process, powdered gelling agents and related additives (e.g., buffers,
surfactants, biocides) are mixed on-site with base fluid water and proppant in large tanks, typically
20,000 gal (80,000 L) (BT Services Company. 2009: Halliburton. 1988). The number of gel slurry
tanks used varies based on site-specific conditions and the size of the fracture job. These tanks can
be subject to leaks or overflow during the batch mixing process and during storage prior to
injection. One of the disadvantages of batch mixing is the need for multiple suction hoses to draw
pre-gelled fluids from storage tanks into the blender, if used, which can increase the potential for
spills. Yeager and Bailey (2013) state that a drawback of batch mixing is the "fluid spillage and
location mess encountered when pre-mixing tanks," suggesting that small spills are not uncommon
during batch mixing. Details of a documented gel slurry spill are presented in Text Box 5-6. Details
of a documented gel slurry spill are presented in Text Box 5-6.
Text Box 5-6. Spill of Gel Slurry during Mixing.
On April 9, 2010, in Louisiana, a company was mixing a gel slurry for an upcoming fracture job. The tank had
developed a crack, which allowed approximately 10,000 gal (38,000 L] of water mixed with 60 gal (230 L] of
gel to leak out. The mixture did not reach a water receptor, and absorbents were used to clean up the gel
fLDEO.20131
In continuous mixing operations, powdered gels are typically replaced with liquid gel concentrates
(Allen. 2013: BT Services Company. 20091. Operators prepare dilute gelling agents as needed using
specialized hydration units (BT Services Company. 2009). Liquid gel concentrates may be stored on-
site in single-purpose tanker trucks (Harms and Yeager. 1987) but are more often stored in
specialized mixing and hydration units (Avala etal.. 2006). Continuous mixing requires less
preparation than batch mixing but typically requires more equipment (BT Services Company. 2009:
Browne and Lukocs. 19991. This can increase the possibility for spills resulting from equipment
malfunctions or human error.
5-36
-------
Chapter 5 - Chemical Mixing
5.5.2 Hoses and Lines
High- and low-pressure hoses and lines are used to transfer hydraulic fracturing fluids from storage
units to specialized mixing and pumping equipment and ultimately to the wellhead. A discussion of
the different types of hoses and lines and possible points of failure is provided below. Figure 5-8
shows an example of hoses and lines at a hydraulic fracturing site.
Figure 5-8. Hoses and lines at a site in Arkansas.
Photo credit: Christopher Knightes (U.S. EPA).
Suction pumps and hoses move large volumes of base fluid to the blender. Incomplete or damaged
seals in inlet or outlet connections can cause fluid leaks at the connection points. Improperly fitted
seals also severely limit or eliminate suction lift, which can impair the suction pump and increase
spill potential. Suction hoses themselves are susceptible to leaks due to wear and tear. Equipment
providers recommend hoses be closely inspected to ensure they are in good operating condition
prior to use fUpstream Pumping. 2015: B1 Services Company. 2009: Malone and Ely. 20071.
Discharge hoses transfer additives from containment vessels or totes to the blender. Given the
potential for concentrated chemicals to spill during transfer from storage totes to the blender, it is
particularly important that these hoses are in good condition and that connector seals or washers
5-37
-------
Chapter 5 - Chemical Mixing
fit properly and are undamaged. Discharge hoses are also used to carry fracturing fluid pumped
from the blender via the low-pressure manifold to the high-pressure pumps. Proppant-heavy fluids
are pumped through discharge hoses at relatively low rates. If a sufficient flow rate is not
maintained, proppant may settle out, damaging pumps and creating a potential for spills and leaks
(Upstream Pumping. 2015: BT Services Company. 2009: Malone and Ely. 20071.
High-pressure flow lines convey pressurized fluids from the high-pressure pumps into the high-
pressure manifold and from the manifold into the wellbore. High-pressure flow lines are subject to
erosion caused by the high-velocity movement of abrasive, proppant-laden fluid. Curved sections of
flow lines (e.g., swivel joints) where abrasive fluids are forced to turn corners are particularly
subject to erosion and are more likely to develop stress cracks or other defects that can result in a
leak or spill. Safety restraints are typically used to prevent movement of flow lines such as in the
event of failure and to help control spills. High-pressure flow lines are pressure-tested to detect
fatigue or stress cracks prior to the fracturing treatment (OSHA. 2015: BT Services Company. 2009:
Arthur etal.. 2008: Malone and Ely. 2007: Halliburton. 19881.
Nineteen spills of chemicals or fracturing fluids associated with leaks from hoses or lines had a total
spill volume of 12,756 gal (48,287 L), with a median volume of 420 gal (1,600 L) (U.S. EPA. 2015ml.
5.5.3 Blender
The blender is the central piece of equipment used to create the fracturing fluid for injection. It
moves, meters, and mixes precise amounts of the base fluid, additives, and proppant and pumps the
mixed slurry to high-pressure pumping equipment (BT Services Company. 2009: Malone and Ely.
2007: Halliburton. 19881 (Figure 5-6). A typical blender consists of a centrifugal suction pump for
pulling base fluid, one or more chemical metering pumps to apportion the additives, one or more
proportioners to measure and feed proppant, and a central agitator tank where fluid components
are mixed together.
The blending process is monitored to ensure that a uniform mixture is maintained regardless of
injection rates and volumes. Excessive or reduced rates of flow during treatment can cause the
blender to malfunction or to shut down, which can result in spills (Malone and Ely. 2007:
Halliburton. 19881. For aqueous hydraulic fracturing fluid blends, spills that occur downstream of
the blender will be a dilute mixture (less than or equal to 2%) of chemicals. Details of a spill from a
blender are presented in Text Box 5-7.
Text Box 5-7. Spill of Hydraulic Fracturing Fluid from Blender.
In May 2006, a blender malfunctioned during a fracture job in Oklahoma. Approximately 294 gal (1,110 L) of
fluid spilled into a nearby wheat field. The fluid consisted of hydrochloric acid, clay stabilizer, diesel, and
friction reducer. Contaminated soil was removed by the operator (OCC. 2013. ID#1370001.
5-38
-------
Chapter 5- Chemical Mixing
5.5.4 Manifold
A trailer-mounted manifold and pump system functions as a central transfer station for all fluids
used in the hydraulic fracturing operation. The manifold is a collection of low- and high-pressure
pipes equipped with multiple fittings for connector hoses. Fluids are pumped from the blender
through the low-pressure manifold hoses, which distribute fluids to high-pressure pump trucks.
Pressurized slurry is sent from the pump trucks through high-pressure manifold lines and into
additional high pressure lines that lead to the wellhead fMalone and Ely. 20071.
Manifold and pump system components require varying amounts of manual assembly and undergo
varying amounts of pre-testing (Malone and Ely. 20071. Improperly tested parts may be more likely
to break or lose functionality, leading to a spill. In manifolds requiring more manual assembly, there
may be more opportunities for human error.
5.5.5 High-Pressure Fracturing Pumps
High-pressure fracturing pumps take the fracturing fluid mixture from the blender, pressurize it,
and propel it down the well. Typically, multiple high-pressure, high-volume fracturing pumps are
needed for hydraulic fracturing (Upstream Pumping. 2015). Such pumps come in a variety of sizes.
Bigger pumps move greater volumes of fluid at higher pressures; therefore, spills from these pumps
can be larger. Smaller pumps can require more operators and more maintenance fBT Services
Company. 20091. and therefore have the potential for an increased frequency of spills.
The "fluid ends" of hydraulic fracturing pumps are the pump components through which fluids are
moved and pressurized. Pump fluid ends must withstand high pressure and move a large volume of
abrasive fluid high in solids content These pumps have multiple parts (e.g., seals, valves, seats and
springs, plungers, stay rods, studs) that can wear out under the stress of high-pressure pumping
(Upstream Pumping. 2015). Given the sustained pressures involved, careful maintenance of fluid
ends is necessary to prevent equipment failure (Upstream Pumping. 2015: API. 2011). Details of a
documented spill from a fracture pump are presented in Text Box 5-8.
Text Box 5-8. Spill of Fluid from Fracture Pump.
On December 19,2011, in Arkansas, a fluid end on a fracture pump developed a leak, spilling approximately
840 gal (3,2 00 L) of fracturing fluid. A vacuum truck was used to recover the spilled fluid, and all affected
soils were neutralized and taken to a landfill at the end of the job, after removal of the equipment (Arkansas
DEO. 2012. ID#0630121.
5.5.6 Surface Wellhead for Fracture Stimulation
A wellhead assembly, often referred to as a frac head or frac stack, is temporarily installed on the
wellhead during the fracture treatment The frac head assembly allows high volumes of high-
pressure proppant-laden fluid to be injected into the formation (OSHA. 2015: Halliburton. 2014:
Stinger Wellhead Protection. 2010). The temporary frac head is equipped with specialized isolation
tools so that the wellhead is protected from the effects of pressure and abrasion.
5-39
-------
Chapter 5 - Chemical Mixing
Figure 5-9. Multiple fracture heads.
Source: DOE/NETL.
As with all components of hydraulic fracturing operations, repeated and prolonged stress from
highly pressurized, abrasive fluids may lead to equipment damage. The presence of minute holes or
cracks in the frac head may result in leaks when pressurized fluids are pumped. In addition, surface
blowouts or uncontrolled fluid releases may occur at the frac head because of valve failure or
failure of other components of the assembly.1 Details of a documented frac head failure are
presented in Text Box 5-9.
Text Box 5-9. Spill from Frac Head Failure.
On March 2, 2011, in Colorado, a frac head failed during fracturing operations. Approximately 8,400 gal
(32,000 L) of slickwater fracturing fluid leaked. The majority of the spill was contained on-site, though a small
amount ran off into a nearby cornrow. There were 5,460 gal (20,700 L) of the fluid recovered, and saturated
soils were scraped and stockpiled on the well pad. There was a net loss of 2,940 gal (11,100 L) (COGCC. 2012.
ID#25245861
1A well blowout is when there is uncontrolled flow of fluids out of a well.
5-40
-------
Chapter 5- Chemical Mixing
5.6 Overview of Chemical Spills Data
Spills of hydraulic fracturing fluids have occurred across the country and have affected the quality
of drinking water resources fU.S. EPA. 2015m: Brantley et al.. 2014: COGCC. 2014: Gradient. 20131.1
Spills may infiltrate drinking water resources by reaching surface water or by leaching into the
groundwater. Potential impacts depend upon a variety of factors including the chemical spilled,
environmental conditions, and actions taken in response to the spill.
5.6.1 EPA Analysis of Spills Associated with Hydraulic Fracturing
The EPA used data gathered from state and industry sources to characterize hydraulic fracturing-
related spills between January 2006 and April 2012 (2015m) (see Text Box 5-10 for additional
information). In this study, the sources had data on over 36,000 spills. Of these spills, the EPA
identified 457 spills that occurred on or near the well pad and definitively related to hydraulic
fracturing. Of these 457 spills, 151 were related to the chemical mixing process - spills that
consisted of chemicals, additives, or fracturing fluids. Information in the spill reports included: spill
causes (e.g., human error, equipment failure), sources (e.g., storage tank, hose or line), volumes, and
environmental receptors. Spill reports contain little information on chemical-specific spill
composition. Spilled fluids were often described by their additive type (e.g., acids, biocides, friction
reducers, cross-linkers, gels,) or as a blended hydraulic fracturing fluid. Specific chemicals
mentioned in spill reports included hydrochloric acid and potassium chloride.
Text Box 5-10. EPA Review of State and Industry Spill Data: Characterization of Hydraulic
Fracturing-Related Spills.
As part of the EPA's Study of the Potential Impacts of Hydraulic Fracturing for Oil and Gas on Drinking Water
Resources, the EPA published the report titled Review of State and Industry Spill Data: Characterization of
Hydraulic Fracturing-Related Spills fU.S. EPA. 2015m). In this document, hereafter referred to as the EPA
spills report, the EPA used data gathered from state and industry sources to characterize hydraulic
fracturing-related spills with respect to volumes spilled, materials spilled, sources, causes, environmental
receptors, containment, and responses. For the purposes of the study, hydraulic fracturing-related spills were
defined as those occurring on or near the well pad before or during the injection of hydraulic fracturing fluids
or during the post-injection recovery of fluids. Because the main focus of this study is to identify hydraulic
fracturing-related spills on the well pad that may reach surface or groundwater resources, the following
topics were not included in the scope of this project: transportation-related spills, drilling mud spills, and
spills associated with disposal through underground injection control wells.
Data on spills that occurred between January 2006 and April 2012 were obtained from nine state agencies
with online spill databases or other data sources, nine hydraulic fracturing service companies, and nine oil
and gas production well operators. The data sources used in this study contained over 36,000 spills. The EPA
searched each spill report for keywords related to hydraulic fracturing (e.g., frac, glycol, flowback). Spill
records from approximately 12,000 spills (33 percent of the total number of spills reviewed) contained
insufficient information to determine whether the event was related to hydraulic fracturing.
(Text Box 5-10 is continued on the following page.)
1 In this assessment, a spill is considered to be any release of fluids. Spills can result from accidents, fluid management
practices, or illegal dumping.
5-41
-------
Chapter 5 - Chemical Mixing
Text Box 5-10 (continued). EPA Review of State and Industry Spill Data: Characterization of
Hydraulic Fracturing-Related Spills.
Of the spills with sufficient information, the EPA identified approximately 24,000 spills (66%) as not related
to hydraulic fracturing on or near the well pad. The remaining 457 spills (approximately 1%) occurred on or
near the well pad and were definitively related to hydraulic fracturing. These 457 spills occurred in 11
different states over six years (January 2006 to April 2012). Of these 457 spills, 151 spills were chemical
mixing-related and included spills of chemicals, additives, and hydraulic fracturing fluid, and 225 releases
were of produced water (Chapter 7).
The EPA categorized spills according to the following causes: equipment failure, human error,
failure of container integrity, other (e.g., well communication, weather, vandalism), and unknown.1
Figure 5-10 presents the percent distribution of causes of hydraulic fracturing-related spills and for
spills associated specifically with chemicals or fracturing fluid. The distributions for causes of
hydraulic fracturing- and chemical mixing-related spills are similar.2
Spills in the EPA spills report were also categorized by the following sources: storage, equipment,
well or wellhead, hose or line, and unknown. Figure 5-11 presents the percent distribution for all
hydraulic fracturing- and chemical mixing-related spills associated with each source.
¦ Equipment failure
¦ Failure of
container integrity
¦ Human error
¦ Other
¦ Unknown
Figure 5-10. Percent distribution of the causes of spills.
Percent distribution by spill type for (a) 457 hydraulic fracturing-related spills (all spills) and (b) 151 chemical
mixing-related spills. Data from U.S. EPA (2015m). Legend shows categories in clockwise order, from the top left of
each pie chart.
1 Well communication is when hydraulic fracturing fluids or displaced subsurface fluids move through newly created
fractures into an offset well or its fracture network (See Section 6.3.2.3 for more details],
2 Hydraulic fracturing-related spills are spills that occur at any phase within the hydraulic fracturing water cycle. These
include chemicals, additives, hydraulic fracturing fluids (chemical mixing phase); produced water; and wastewater.
5-42
-------
Chapter 5 - Chemical Mixing
Equipment
Hose or line
Storage
Well or wellhead
Unknown
Note: One spill was associated with two sources,
Equipment and Well or Wellhead, and was counted
in both categories.
Note: Percentages do not sum to 100%
due to rounding.
Figure 5-11. Percent distribution of the sources of spills.
Percent distribution of spill source of (a) 457 hydraulic fracturing-related spills (all spills) and (b) 151 chemical
mixing-related spills. Data from U.S. EPA (2015m . Legend shows categories in clockwise order, from the top left of
each pie chart.
Figure 5-12 presents the distribution of the number of spills for different volumes for hydraulic
fracturing- and chemical mixing-related spills. The spills associated with chemical mixing ranged in
volume from 5 to 19,320 gal (19 to 73,130 L), with a median volume of 420 gal (1,600 L). The
source of largest spills was storage containers, which released approximately 83,000 gal (314,000
L) of spilled fluid (Figure 5-13b). Spills from wells or wellheads are often associated with high spill
volumes. There were no reported chemical mixing-related spills greater than 100,000 gal (380,000
L) (Figure 5-15b).
^160 40% C3) 70 54%
~o
QJ
50
-1 40
(LI
w 80 ¦ o cf
X m 1iZ°i5% 30
40
o
aj
_Q
E
13
H7% 8% 1120 I
¦ 4% 4% ¦ 5% z5 10 6% 4%
| ¦ ¦ ¦ ¦ 0% 03% 0 ¦ ¦ ¦ ¦ ¦
5% -% 2% 0% 0%
^ a* # & ^ ^ ^ ^ ^
Reported Volume per Spill (gallons) Reported Volume per Spill (gallons)
Figure 5-12. Distribution of the number of spills for different ranges of spill volumes.
Number of spills due to Hydraulic Fracturing related activities and distribution of spill volumes for (a) 457 hydraulic
fracturing-related spills (all spills) and (b) 151 chemical mixing-related spills. A value of 0% means that there were
no spills in that category. Data from U.S. EPA (2015m).
5-43
-------
Chapter 5 - Chemical Mixing
Figure 5-13 presents the total volume of spills for different sources for all hydraulic fracturing-
related activity and those associated with chemicals and fracturing fluid. The reported total volume
of 125 of 151 chemical or hydraulic fracturing fluid spills was approximately 184,000 gal
(697,000 L). The volume was unknown for 26 of these spills.
(a) (b)
1800 90
~v 1600 HK 80
I 1400 ¦ I ™
1? 1200 S 60
£ 1000 ¦ £ 50
S 800 « 40
J 600 J 30
<0 400 13 20
200 10
m
c 0 c 0
5 Equipment Hose or Storage Well or Unknown « Equipment Hose or Storage Well or Unknown
_g line wellhead o line wellhead
Source of Spill Source of Spill
Figure 5-13. Total volume of fluids spilled from different sources.
Total volume of fluids spilled for (a) 457 hydraulic fracturing-related spills (all spills) and (b) 151 chemical mixing-
related spills. Data from U.S. EPA (2015m).
Figure 5-14 presents the number of spills that reached environmental receptors, by receptor type,
for all hydraulic fracturing-related activity (Figure 5-14a) and those associated with chemicals and
fracturing fluid (Figure 5-14b). Environmental receptors (i.e., surface water, groundwater, soil)
were identified in 101 of the 151 chemical mixing-related spills, or 67% of all chemical and
fracturing fluid spills in the EPA's analysis (U.S. EPA. 2015ml. Soil was by far the dominant
environmental receptor, with 97 spills reaching soil; reported spill volumes ranged from 5 gal to
8,300 gal (19 L to 31,000 L). Thirteen spill reports indicated that the spilled fluid had reached
surface water; reported spill volumes ranged from 28 gal to 7,350 gal (105 L to 27,800 L). Nine spill
reports identified both soil and surface water as a receptor; spill volumes ranged from 28 gal to
2,856 gal (106 L to 10,800 L). Groundwater was not identified as a receptor from spills of chemicals
or hydraulic fracturing fluid in any of the spill reports. Due to the lack of observations, it is often
unclear if there was impact on groundwater. Movement through the subsurface is generally slow.1
It may take years for a spilled fluid to reach groundwater or to reach a drinking water well. Thus,
even if there is a pre-drilling characterization of groundwater chemistry in private/public wells, the
time period of transport to actually detect a release at these private/public wells for contaminants
that are transported at the rates of groundwater flow (see Section 5.8 for discussion on fate and
transport of spilled chemicals).
1 For example, a groundwater flow rate of 1 foot per day (not uncommon] would mean it could take approximately 1,000
days (~3 years] to travel 1,000 ft (305 m] from the well pad. Likewise, for a groundwater travel rate of 0.1 ft (0.03m] per
day, impact would not be observed for at least 10,000 days (—27 years]. For a travel rate of 10 ft (3 m] per day, the time
for impact would be at least 100 days (—0.3 years].
5-44
-------
Chapter 5 - Chemical Mixing
(a) (b)
350 120
300
l/>
= 250
i/i
t 200
Ja 150
£
^ 100
50
100
1/)
I 80
° 60
01
-Q
E 40
3
Z
20
Surface Water Groundwater Soil Surface Water Groundwater
Environmental Receptor Environmental Receptor
¦ Yes ¦ Unknown ¦ No
Figure 5-14. Number of spills by environmental receptor.
Number of hydraulic fracturing-related spills and chemical mixing-related spills that reported whether an
environmental receptor was reached for (a) 457 hydraulic fracturing-related spills (all spills) and (b) 151 chemical
mixing related spills. "Yes" means that the spill was reported to reach this receptor. "Unknown" refers to hydraulic
fracturing related spill events for which environmental receptors were specified as unknown or not identified
(positively or negatively). "No" means the spill was reported to not meet this receptor. Data from U.S. EPA
(2015m).
Storage units were the predominant sources of spills that reached an environmental receptor. Six
spills from storage containers reached a surface water receptor. Thirty-eight of the spills from
storage units reached a soil receptor. If a spill was confined to a lined well pad, for example, it might
not have reached the soil, but most incident reports did not include whether the well pad was lined
or unlined. Regarding spills of hydraulic fluids and chemicals from storage containers, 16 spills
were due to failure of container integrity, which includes holes and cracks in containers, and
overflowing containers as a result of human error or equipment malfunctions.
5.6.2 Estimated Spill Rate and Other Spill Reports and Data
The rate of reported spills during the hydraulic fracturing water cycle is estimated to range from
0.4 to 12.2 reported spills for every 100 wells, based on spills data from Brantley et al. (2014).
Gradient (20131. Rahm etal. (2015). U.S. EPA (2013a). and North Dakota Department of Health
f2015 (Appendix E) with a median rate of 2.6 reported spills for every 100 wells. (See Appendix
Section C.4 and Appendix Table C-8 for details.) The estimated rates provide an approximate
estimate of the potential frequency of the number of spills at a site. It is uncertain how
representative these rates are of national spill rates or rates in other states. These numbers are not
specific to the chemical mixing stage.
There are an estimated 2.6 reported spills of injected fluids and chemicals per 100 wells
hydraulically fractured in North Dakota, based on an analysis of the North Dakota spills database
for 2015, separate from the EPA spills report. The median spill volume of injection fluid was 1750
gal (6620 L), with a range of 2.9 to 17,600 gal (11 to 66,600 L). The median spill volume of injection
chemical was 44 gal (167 L), with a range of 2.1 to 126 gal (7.9 to 477 L) (see Appendix E for more
information).
5-45
-------
Chapter 5 - Chemical Mixing
A study of spills reported to the Colorado Oil and Gas Conservation Commission identified 125
spills during well stimulation (i.e., a part of the life of an oil and gas well that often, but not always,
includes hydraulic fracturing) between January 2010 and August 2013. Of these spills, 51% were
caused by human error and 46% were due to equipment failure (COGCC. 2014).
Considine etal. f20121 identified spills related to oil and gas development in the Marcellus Shale
that occurred between January 2008 and August 2011 from Notices of Violations issued by the
Pennsylvania Department of Environmental Protection. The authors identified spills greater than
400 gal (1,500 L) and spills less than 400 gal (1,500 L). Among these spills, spilled fluids included
hydrochloric acid, gel friction reducer, and blended hydraulic fracturing fluid. Brantley etal. (2014)
identified fewer than 10 instances of spills of additives and/or hydraulic fracturing fluids greater
than 400 gal (1,500 L) that reached surface waters in Pennsylvania between January 2008 and
September 2013. Reported spill volumes, among these spills, ranged from 3,400 gal to 227,000 gal
(13,000 L to 859,000 L).
Surface spills related to hydraulic fracturing activities are not well documented in the scientific
literature. There is some evidence of spills and impacts on environmental media (e.g., U.S. EPA.
2015i: Brantley etal.. 2014: Gross etal.. 2013: Papoulias and Velasco. 2013). Papoulias and Velasco
(2013) stated that fluid overflowed a retention pit into surface water and likely contributed to the
distress and deaths of threatened blackside dace fish in Kentucky. A variety of chemicals entered
the creek and significantly reduced the stream's pH and increased stream conductivity. Using data
from post-spill sampling reports in Colorado, Gross etal. f20131 identified concentrations of
benzene, toluene, ethylbenzene, and xylene (BTEX) in groundwater samples. They attributed this to
numerous hydraulic fracturing-related spills, although not necessarily specifically related to the
chemical mixing process. This work, however, demonstrate that surface spills impacted
groundwater, with a frequency of < 0.5% of active wells. Drollette etal. f20151 reported that
organic compounds detected in shallow aquifers were consistent with surface spills, and that diesel
range compounds had elevated concentrations compared to gasoline range compounds, further
suggesting evidence of feasible groundwater impact.
5.7 Spill Prevention, Containment, and Mitigation
Spill prevention, containment, and mitigation affect the frequency and severity of the impacts of
spills. Several factors influence spill prevention, containment, and mitigation, including federal,
state, and local regulations and company practices. State regulations governing spill prevention,
containment, and mitigation at hydraulic fracturing facilities vary in scope and stringency fPowell.
2013: GWPC. 20091. Employee training and equipment maintenance are also factors in effective
spill prevention, containment, and mitigation. Analysis of these factors was outside the scope of this
assessment.
The province of New Brunswick, Canada released rules for industry on responsible environmental
management of oil and natural gas activities (GNB. 2013). Hydraulic fracturing service companies
themselves may develop and implement spill prevention and containment procedures. It was
beyond the scope of this assessment to evaluate the efficacy of these practices or the extent to
which they are implemented.
5-46
-------
Chapter 5- Chemical Mixing
Spill containment systems include primary, secondary, and emergency containment systems.
Primary containment systems are the storage units, such as tanks or totes, in which fluids are
intentionally kept Secondary containment systems, such as liners and berms installed during site
set-up, are intended to contain spilled fluids until they can be cleaned up. Emergency containment
systems, such as berms, dikes, and booms, can be implemented temporarily in response to a spill.
The EPA investigated spill containment and mitigation measures in an analysis of spills related to
hydraulic fracturing activities fU.S. EPA. 2015ml. Of the approximately 25% of reports that
included information on containment, the most common types of containment systems referenced
in the hydraulic fracturing-related spill records included berms, booms, dikes, liners, and pits,
though many of the spill reports did not indicate specific containment measures. Some spills were
reported to breach the secondary containment systems. Breaches of berms and dikes were most
commonly reported.
In cases where secondary containment systems were not present or were inadequate, operators
sometimes built emergency containment systems. The most common were berms, dikes, and
booms, but there were also instances where ditches, pits, or absorbent materials were used to
contain the spilled fluid. Absorbent materials were generally used when small volumes (10 - 200
gal or 40-800 L) of additives or chemicals were spilled (U.S. EPA. 2015ml. There was not enough
information to detail the use of emergency containment systems or their effectiveness.
Remediation is the action taken to clean up a spill and its affected environmental media. The most
commonly reported remediation activity, mentioned in approximately half of the hydraulic
fracturing-related spill records evaluated by the EPA, was removal of spilled fluid and/or affected
media, typically soil. Other remediation methods reported in U.S. EPA f2015ml included the use of
absorbent material, vacuum trucks, flushing the affected area with water, and neutralizing the
spilled material. Removal activities were found to occur in various combinations. For example, a
spill of approximately 4,200 gal (16,000 L) of acid was cleaned up by first spreading soda ash to
neutralize the acid and then removing the affected soil fU.S. EPA. 2015ml.
5.8 Fate and Transport of Spilled Chemicals
The fate and transport of chemicals in the environment is complex. Due to the complexities of the
processes and the site-specific and chemical-specific nature of spills, it is difficult to develop a full
assessment of their fate and transport. The potential for hydraulic fracturing chemicals and fluids to
reach drinking water resources is further complicated by the fact that these chemicals are typically
present as mixtures, and unlike many organic contaminant mixtures (e.g., gasoline, diesels, PCBs,
PAHs), hydraulic fracturing fluid chemicals are present as complex mixtures of chemicals covering a
range of chemical classes with varying properties, often in aqueous solutions.
In this section, we provide a general overview of fate and transport of hydraulic fracturing-related
chemicals spilled in the environment to give the reader a general understanding of the potential
pathways and processes with which these chemicals can impact drinking water resources (Figure
5-15). We also include a discussion of the physicochemical properties of the organic chemicals used
5-47
-------
Chapter 5 - Chemical Mixing
in hydraulic fracturing fluids, because these properties directly affect the transport of chemicals in
the environment. This presentation is not meant to be exhaustive,
A chemical spill has the potential to migrate to and have an impact on drinking water resources.
Once spilled, there are different paths that chemicals can travel and different processes they can
undergo. Chemicals can react and transform into other chemicals, volatilize, travel to surface water,
leach into and partition to soils, and/or reach groundwater. The potential path and the severity of
the impact of a spill depend on different factors, including site conditions; the length of the path to a
drinking water resource; the type and characteristics of the drinking water resource (stream, lake,
aquifer}; environmental conditions; climate; weather; chemical properties, constituents, and
concentrations; and the volume of the release. The point in the chemical mixing stage where the
spill occurs affects potential impact. If the spill occurs before chemicals are mixed into the base
fluid, the chemicals will be in a more concentrated form. If the hydraulic fracturing fluid spills, then
the chemicals will be diluted by the base fluid and can feasibly be present in lower concentrations.
There can also be effects on persistence and mobility due to interactions among the chemicals
present. The total mass of spilled chemical can therefore be dependent on what stage in the process
a spill occurs.
Hydraulic fracturing-
related spill or release
chemical
groundwater
plume
Surface Water
Sorption
Schematic of the Fate and Transport Processes
Governing Potential Impacts of Spills
and Releases to Drinking
. j Water Resources
Volatilization
Figure 5-15. Fate and transport schematic for a spilled hydraulic fracturing fluid.
Schematic shows the potential paths and governing processes by which spilled chemicals can lead to potential
impacts on drinking water resources.
5-48
-------
Chapter 5- Chemical Mixing
For inorganic chemicals, the properties and processes governing fate and transport depend on pH,
oxidation state, presence of iron oxides, soil organic matter, cation exchange capacity, and major ion
chemistry fU.S. EPA. 19961.1 Transport of these chemicals into groundwater depends on the nature
of groundwater flow and flow through the unsaturated zone above the water table.2 Potential
transformations of inorganic chemicals differ from those of organic chemicals. Some inorganic
anions (i.e., nitrate, chloride, and bromide) move with their carrier liquid and are affected mostly by
physical transport mechanisms. For many inorganic chemicals, transport is driven by the physical
flow processes (advection and dispersion), sorption, and precipitation. The relative role of each of
these depends on both chemical and environmental characteristics.3 4
Determining the fate and transport of organic chemicals and mixtures is a complex problem,
because of the many processes and different environmental media (air, soil, water). Unlike
inorganic chemicals, organic chemicals degrade, which can affect their movement and potential
impact Schwarzenbach etal. (2002) formalized a general framework for organic chemical
transport, where transport and transformation depend on both the nature of the chemical and the
properties of the environment The fate and transport of organic chemicals in soils has been
presented in the literature (e.g., Bouchard et al.. 2011: Rivettetal.. 2011: Abriola and Pinder. 1985a.
b) and in textbooks (e.g., Domenico and Schwartz. 1997: Schnoor. 1996: Freeze and Cherry. 1979b).
5.8.1 Potential Paths
Chemicals and hydraulic fracturing fluids that are released into the environment may travel along
different potential paths, as detailed in Figure 5-15. Liquids can flow overland to nearby surface
water or infiltrate the subsurface, where they may eventually reach the underlying groundwater or
travel laterally to reach surface water. Movement can occur quickly or be delayed and have a later
or longer-term impact Surface and groundwater gain or lose flow to each other (Chapter 2), and
can transport chemicals in the process. A dry chemical (e.g., gelling agents, biocides, friction
reducers) released to the environment can remain where it is spilled. Any spill that is not removed
could act as a long-term source of contamination. Wind could cause the chemical to disperse and
rain could mobilize soluble chemicals. Dissolved chemicals can infiltrate into soil or flow overland.
Insoluble chemicals and those sorbed to soil particles could be mobilized by rain events via runoff
and erosion.
5.8.1.1 Movement across the Land Surface
In low permeability soils, there may be little infiltration and greater overland flow. Higher
permeability soils will allow fluid to penetrate into the soil layer. In either case, some of the
1 Cation exchange capacity is the total amount of cations (positively charged ions] that a soil can hold. For example, when
metal ions like Ca2+ and Na+ pass through the soil, they adhere and remain attached to the soil.
2 The unsaturated zone is also referred to as the vadose zone. Meaning "dry," the vadose zone is the soil zone above the
water table that is only partially filled by water.
3 Advection is a mechanism for moving chemicals in flowing water, where a chemical moves along with the flow of the
water itself.
4 Sorption is the general term used to describe the partitioning of a chemical between soil and water and depends on the
nature of the solids and the properties of the chemical.
5-49
-------
Chapter 5 - Chemical Mixing
chemicals in the fluid can sorb to the soil particles and the vegetation, and then these chemicals can
be mobilized during precipitation, runoff, or erosion. As precipitation percolates through the soil, it
can dissolve stored chemicals, which can then migrate toward groundwater. The type of release is
also important. If the spill is a slow leak, then the liquid may pond and the affected area will expand
slowly with greater potential for infiltration. If a more rapid release occurs, like a blowout or tank
failure, then momentum can result in greater overland movement and less soil infiltration during
the event, with greater potential to reach a nearby surface water.
5.8.1.2 Movement through the Subsurface
The unsaturated and saturated zones are the two zones of soils below the ground surface.
Movement through the unsaturated zone is driven by the depth of ponding of the spilled fluid,
gravity, and capillary properties of the subsurface.1 In fractured rock or highly permeable soils,
fluids can move quickly through the subsurface. In low permeability soil, the movement of the fluid
may be slower. However, the presence of preferential pathways (e.g., fractures, heterogeneities,
root holes, and burrows) can result in faster movement than the overall permeability would
suggest.
As chemicals pass through the subsurface, some can sorb to soil or remain in the open spaces
between soil particles, effectively slowing their movement Chemicals can be mobilized during
future precipitation events, resulting in infiltration towards groundwater or movement through the
unsaturated zone towards surface water.
Fluids that move through the subsurface into the saturated zone will move in the direction of the
flowing groundwater. Generally, fluids travel farther in systems with high groundwater flow rates
and high recharge (e.g., sandy aquifers in humid climates) than in systems with low flow and low
recharge. Chemicals can sorb to suspended soil particles, complex with naturally occurring
chemicals (e.g., dissolved organic carbon), or associate with colloids and be transported with the
flowing water.2 These mechanisms can mobilize sparingly soluble chemicals that would otherwise
be immobile.
5.8.2 Physicochemical Properties of Organic Hydraulic Fracturing Chemicals
Three physicochemical properties are useful to describe the movement of organic chemicals in the
environment: (1) Kow, the octanol-water partition coefficient, (2) the aqueous solubility, and (3) the
Henry's law constant3 These properties describe whether a chemical will sorb to soil and organic
1 Capillarity occurs because of the forces of attraction of water molecules to themselves (cohesion] and to other solid
substances such as soils (adhesion].
2 Complexation is a reaction between two chemicals that form a new complex, either through covalent bonding or ionic
forces. This often results in one chemical solubilizingthe other.
3 The octanol-water partition coefficient (Kow] represents the ratio of the solubility of a compound in octanol (a nonpolar
solvent] to its solubility in water (a polar solvent] in a mixture of the two. The higher the K0w, the more nonpolar the
compound.
5-50
-------
Chapter 5- Chemical Mixing
matter or stay in water (Kow), how much of a chemical may dissolve in water (aqueous solubility),
and whether a chemical will tend to remain in the water or volatilize (Henry's law constant).1
The Kow measures the relative hydrophobicity (chemicals that prefer to be in oil, log Kow >0) and
hydrophilicity (chemicals that prefer to be in water, log Kow <0) of a chemical. Aqueous solubility is
the maximum amount of a chemical that will dissolve in water in the presence of a pure chemical;
solubility generally serves as an upper bound on possible concentrations. The Henry's law constant
is the ratio of the concentration of a chemical in air (or vapor pressure) to the concentration of that
chemical in water.
Estimates and measured values for physicochemical properties were obtained by using the
Estimation Program Interface (EPI) Suite 4.1, as described in Appendix C.2 Of the 1,084 chemicals
the EPA listed as used in hydraulic fracturing (Appendix H), EPI Suite™ has estimated properties for
455 organic chemicals (42% of all chemicals) with structures that are considered suitably
representative of the substance to compute properties within the constraints of EPI Suite™
software. Only uniquely defined organic desalted structures were submitted for property
calculation. Figure 5-16 presents histograms of all 455 of the organic chemicals, sorted by four
physicochemical parameters: measured log Kow (n = 195), estimated log Kow (n=455), estimated log
of the aqueous solubility (n = 455), and estimated log of Henry's law constant (at 77°F or 25°C,
n = 449). Property estimation methods are limited in their ability to predict physicochemical
properties. Chemicals that are different than the chemicals used to develop the estimation
techniques may have more error associated with their predictions. These figures enable
comparison of physicochemical properties across the organic chemicals for which we have values.
These figures show how the physicochemical properties are distributed and which chemicals have
higher values compared to others with lower values. Limitations in knowing what chemicals are
present (e.g., CBI) further hinders our ability to know the physicochemical properties of these
chemicals and their potential to move through the environment and impact drinking water
resources. These estimates are solely for the organic chemicals for which EPI Suite™ could be used.
This does not provide information on the 258 inorganic chemicals or the 361 organic mixtures or
polymers. This limits our ability to make a full assessment on the physicochemical properties of all
chemicals, yet provides insight into the properties of the organic chemicals used.
1 We present the physicochemical parameter values using logio because of the wide range of values that these parameters
cover.
2 EPI Suite™, version 4.1, http://www.epa.gov/opptintr/exposure/puhs/episuite.htm HJ.S. EPA. 2012c). The EPI
(Estimation Programs Interface] Suite™ is a Windows®-based suite of physicochemical property and environmental fate
estimation programs developed by the EPA Office of Pollution Prevention and Toxics and Syracuse Research Corporation.
EPI Suite™ provides estimates of physicochemical properties for organic chemicals and has a database of measured values
for physicochemical properties when available. EPI Suite™ cannot estimate parameters for inorganic chemicals.
5-51
-------
Chapter 5 - Chemical Mixing
Most mobile
_uJL I ri j:j lu
-5 0 5
Measured log K [L/kg]
^ Most mobile
Least mobile' ^
. (b)
I
J
L
-25 -20 -15 -10 -5 0 5 10 15 20 25
Estimated log K [L/kg]
t 150 -
-25 -20 -15 -10 -5 0 5
Estimated log Solubility [mg/L @ 25C]
Stays in water
Readily escapes to air
-80 -70 -60 -50 -40 -30 -20 -10 0 10
Estimated log Henry's Law Constant [atm m3 mole"1 @ 25C]
Figure 5-16. Histograms of physicochemical properties of organic chemicals used in the
hydraulic fracturing process.
Physicochemical properties as given by EPI Suite™ (a) measured values of log Kow, (b) estimated log Kow, (c)
estimated log Solubility, and (d) estimated log Henry's law constant.
We used EPI Suite™ to determine the physicochemical properties for 19 CBI chemicals used in
hydraulic fracturing fluids. These chemicals were submitted to the EPA by nine service companies
from 2005 to 2009 (see Text Box 5-3 for discussion on CBI).1 The CBI chemical physicochemical
properties are plotted as histograms in Appendix Figure C-l. The values of the physicochemical
properties of known and CBI chemicals are similar, covering similar ranges and centered on similar
values, suggesting that even though these chemicals are not publicly known, their physicochemical
properties are not appreciably different from the known chemicals. This suggests that their fate and
transport would not be appreciably different than the chemicals that are publicly known.
5.8.3 Mobility of Organic Hydraulic Fracturing Chemicals
Figure 5-16 shows the distribution of log Kow, solubility, and Henry's Law constant for organic
chemicals used in hydraulic fracturing fluids. These figures suggest that the organic chemicals used
in hydraulic fracturing cover a wide range of physicochemical properties. For example, many
chemicals are centered around log Kow = 0, which indicates that these chemicals are likely to
associate roughly equally with organic or aqueous phases. Many chemicals have log Kow > 0,
indicating less mobility, which may cause these chemicals to serve as later-term or long-term
sources of impact on drinking water. Solubilities range from fully miscible to sparingly soluble.
Many chemicals have log Henry's law constants less than 0, indicating that most are not highly
volatile. Volatilization may not serve as a dominant loss process for hydraulic fracturing chemicals.
1 Well operators may specify certain ingredients as confidential business information (CBI] and not disclose the chemicals
used to FracFocus. The CASRNs of a range of CBI chemicals were provided to the EPA by nine service companies.
5-52
-------
Chapter 5- Chemical Mixing
The 20 chemicals with the smallest Kow (most mobile) may have greater potential to cause
immediate impacts on drinking water resources (Appendix Table C-10). Most of these chemicals
were infrequently reported in disclosures (<2% of wells) in the EPA FracFocus 1.0 project database
fU.S. EPA. 2015al Choline chloride (14% of wells), used for clay control, and
tetrakis(hydroxymethyl)-phosphonium sulfate (11% of wells), abiocide, were more commonly
reported. The 20 chemicals with the largest Kow (least mobile) may have a greater potential to serve
as long-term sources of contamination (Appendix Table C-ll). The estimated aqueous solubilities
for some of these chemicals are extremely low, with highest solubilities of less than 10 ng/L. Seven
low mobility chemicals were reported in disclosures in the EPA FracFocus 1.0 project database (U.S.
EPA. 2015c). Five were reported infrequently (<1% of wells). Tri-n-butyltetradecylphosphonium
chloride (6% of wells), used as a biocide, and C>10-alpha-alkenes (8% of wells), a mixture of alpha-
olefins with carbon numbers greater than 10 used as a corrosion inhibitor, were more commonly
reported. Sorbitan, tri-(9Z)-9-octadecenoate, a mineral oil co-emulsifier (0.05% of wells) had the
highest estimated log Kow of 22.56.1
Table 5-7 shows the EPI Suite™ estimated physicochemical property values of the 20 chemicals
most frequently reported nationwide in disclosures along with estimated mean and median
volumes based on disclosures in the EPA FracFocus 1.0 project database (U.S. EPA. 2015c). Most
have log Kow < 1, meaning that they are generally hydrophilic and will associate with water. These
chemicals also have very high solubilities, so they will be mobile in the environment, transport with
water, and can occur at high concentrations. These chemicals have the potential for faster impacts
on drinking water resources.
Naphthalene (CASRN 91-20-3) has a measured log Kow = 3.3 with an estimated solubility of 142.1
mg/L, which means it will be less mobile in the environment. Naphthalene will sorb to particles and
move slowly through the environment and has the potential to act as a long-term source of
contamination.2 All of these chemicals have low Henry's law constants, so they tend not to
volatilize. We also include ranges of similar physicochemical properties for two chemicals that are
organic mixtures: distillates, petroleum, hydrotreated light (CASRN 64742-47-8) and solvent
naphtha, petroleum, heavy arom. (CASRN 64742-94-5). Both of these are complex organic mixtures,
and thus EPI Suite™ cannot estimate their properties. However, the Total Petroleum Hydrocarbon
Work Group has provided regressions to relate physicochemical properties to the number of
carbons for aliphatic and aromatic hydrocarbons (Gustafson et al.. 1997). which shows that they
have low solubilities and large log Kow.
1 Sorbitan, tri-(9Z]-9-octadecenoate, CASRN 26266-58-0, is soluble in hydrocarbons and insoluble in water, listed as an
effective coupling agent and co-emulsifier for mineral oil f Santa Cruz Biotechnology. 2015: ChemicalBook. 20101
2 Chemicals can have the potential to be long-term sources of contamination when they move slowly through the
environment. In this discussion, we are not accounting for biodegradation or other transformation processes, which may
reduce the persistence of certain chemicals in the environment. Under certain conditions, for example, naphthalene is
biodegradable, which can reduce or remove it from the environment, and thus may not be a long-term source of
contamination.
5-53
-------
Chapter 5 - Chemical Mixing
Table 5-7. The 20 chemicals reported most frequently nationwide for hydraulic fracturing based on the EPA FracFocus 1.0 project
database, with EPI Suite™ physicochemical parameters where available, and estimated mean and median volumes of those
chemicals where density was available.
Excludes water, sodium chloride, and quartz. NA means that the physicochemical parameter is not provided by EPI Suite™ or the volume could not be
estimated due to missing data. For organic salts, parameters are estimated using the desalted form. Analysis considered 34,675 disclosures and 676,376
ingredient records that met selected quality assurance criteria, including: completely parsed; unique combination of fracture date and API well number;
fracture date between January 1, 2011, and February 28, 2013; valid CASRN; and valid concentrations. Disclosures that did not meet quality assurance criteria
(3,855) or other, query-specific criteria were excluded from analysis.
Rank
Chemical name
CASRN
Number of
wells using
chemical
(% of wells)
Log (unitless)
Water solubility
estimate from
log Kow
(mg/L @ 25°C)
Henry's Law Constant
(atm m3/mole @ 25°C)
Estimated
volume, per
disclosure (gal)
Estimated
Measured
Estimated,
bond method
Estimated,
group method
25
Measured
Mean
Median
1
Methanol
67-56-1
24,753 (72%)
-0.63
-0.77
1.00 x 106
4.27 x 10"6
3.62 x 10"6
4.55 x 10"6
1,218
110
2
Distillates, petroleum,
hydrotreated lighta,b
64742-47-8
22,463 (65%)
log Koc = 4.5
to 6.7
NA
0.00035 to 0.12
55 to 69
cm3/cm3
NA
NA
NA
NA
3
Hydrochloric acid
7647-01-0
22,380 (65%)
NA
NA
NA
NA
NA
NA
28,320
3,110
4
Isopropanol
67-63-0
16,039 (47%)
0.28
0.05
4.024 x 105
7.52 x 10"6
1.14 x 10"5
8.10 x 10"6
2,095
55
5
Ethylene glycol
107-21-1
15,800 (46%)
-1.2
-1.36
1.00 x 106
1.31 x 10"7
5.60 x 10"n
6.00 x 10 s
614
184
6
Peroxydisulfuric acid,
diammonium salt
7727-54-0
14,968 (44%)
NA
NA
NA
NA
NA
NA
NA
NA
7
Sodium hydroxide
1310-73-2
13,265 (39%)
NA
NA
NA
NA
NA
NA
551
38
8
Guar gum
9000-30-0
12,696 (37%)
NA
NA
NA
NA
NA
NA
NA
NA
9
Glutaraldehyde
111-30-8
11,562 (34%)
-0.18
NA
1.672 x 105
1.10 x 10"7
2.39 x 10"s
NA
1,313
122
10
Propargyl alcohol
107-19-7
11,410 (33%)
-0.42
-0.38
9.355 x 105
5.88 x 10"7
NA
1.15 x 10"6
183
2
11
Potassium hydroxide
1310-58-3
10,049 (29%)
NA
NA
NA
NA
NA
NA
NA
NA
12
Ethanol
64-17-5
9,861 (29%)
-0.14
-0.31
7.921 x 105
5.67 x 10"6
4.88 x 10"6
5.00E-06
831
121
13
Acetic acid
64-19-7
8,186(24%)
0.09
-0.17
4.759 x 105
5.48 x 10"7
2.94 x 10"7
1.00 x 10"7
646
47
14
Citric acid
77-92-9
8,142 (24%)
-1.67
-1.64
1.00 x 106
8.33 x 10"18
NA
4.33 x 10"14
163
20
5-54
-------
Chapter 5- Chemical Mixing
Rank
Chemical name
CASRN
Number of
wells using
chemical
(% of wells)
Log (unitless)
Water solubility
estimate from
log Kow
(mg/L @ 25°C)
Henry's Law Constant
(atm m3/mole @ 25°C)
Estimated
volume, per
disclosure (gal)
Estimated
Measured
Estimated,
bond method
Estimated,
group method
25
Measured
Mean
Median
15
2-Butoxyethanol
111-76-2
7,347 (21%)
0.57
0.83
6.447 x 104
9.79 x 10"8
2.08 x 10 s
1.60 x 10"6
385
26
16
Solvent naphtha,
petroleum, heavy
arom.b'c
64742-94-5
7,108(21%)
log Koc = 3.2
to 2.7
NA
5.8 to 65
0.028 to 0.39
cm3/cm3
NA
NA
NA
NA
17
Naphthalene
91-20-3
6,354 (19%)
3.17
3.3
1.421 x 102
5.26 x 10"4
3.7 x 10"4
4.4 x 10"4
72
12
18
2,2-Dibromo-3-
nitrilopropionamide
10222-01-2
5,656 (16%)
1.01
0.82
2.841 x 103
6.16 x 10"14
NA
1.91 x 10"8
183
5
19
Phenolic resin
9003-35-4
4,961 (14%)
NA
NA
NA
NA
NA
NA
NA
NA
20
Choline chloride
67-48-1
4,741 (14%)
-5.16
NA
1.00 x 106
2.03 x 10"16
NA
NA
2,131
290
a Hydrotreated light petroleum distillates (CASRN 64742-47-8) is a mixture of hydrocarbons in the C9 to C16 range.
b Physicochemical parameters are estimated using Gustafson et al. (1997). Parameters are presented as log Koc (soil organic carbon-water partition coefficient), solubility (mg/L),
and Henry's Law Constant (cm3/cm3).
c Heavy aromatic solvent naphtha (petroleum) (CASRN 64742-94-5) is mixture of aromatic hydrocarbons in the C9 to C16 range.
5-55
-------
Chapter 5 - Chemical Mixing
For the top 20 chemicals, many chemicals have high solubilities and negative or almost zero log Kow
(e.g., methanol, isopropanol, ethylene glycol). These chemicals are likely to travel quickly through
the environment and could result in an immediate impact. Three chemicals, with larger log Kow and
smaller solubilities (distillates, petroleum, hydrotreated light; solvent naphtha, petroleum, heavy
arom.; and naphthalene) may result in more severe impacts. These chemicals could associate with
the soil particles, releasing into the groundwater at low concentrations slowly over time, and thus
serve as long-term sources of contamination.
Mobility of a chemical is complex, and these numbers solely represent how a chemical behaves in
an infinitely dilute aqueous solution, a simplifying approximation of the real world. Many factors
can affect the fate and transport of a chemical, such as the transformation process (e.g.,
biodegradation), the presence of other chemicals, and site and environmental conditions. We
discuss these factors in the next sections.
5.8.4 Transformation Processes
Once a chemical is released into the environment, it can transform or degrade. Understanding the
processes governing these reactions in the environment is important to assessing potential impacts.
The transformation of a chemical may reduce its concentration over time. Chemicals may
completely degrade before reaching a drinking water resource. Transformation processes can be
biotic or abiotic and may transform a chemical into a less or more harmful chemical.
One important transformation process is biodegradation. Biodegradation is a biotic process where
microorganisms transform a chemical from its original form into another chemical. For example,
the general biodegradation pathway of methanol is CFhOH-^ CH2O CHOOH CO2 or
methanol formaldehyde formic acid carbon dioxide (Methanol Institute. 2013).1 This
pathway shows how the original chemical transforms through a series of steps until it becomes the
final product, carbon dioxide. Some chemicals are readily biodegraded, while others break down
slowly over time. Biodegradation is a highly site-specific process, requiring nutrients, a carbon
source, water, and an energy source. A highly biodegradable chemical could be persistent if the
conditions for biodegradability are not met. Conversely, a chemical could biodegrade quickly under
the right conditions, affecting its potential to impact a drinking water resource. The relationship
between mobility and biodegradability is complex, and a variety of factors can influence a
particular chemical's movement through the environment.
Abiotic processes, such as oxidation, reduction, photochemical reactions, and hydrolysis, can
transform or break apart chemicals. The typical results are products that are more polar than the
1 In methanol biodegradation, PQQ (pyrroloquinoline quinone] is a redox cofactor that goes from PQQ to PQQH2 removing
two hydrogen from methanol in the first step to form formaldehyde. Water is added to formaldehyde to provide the
second oxygen to form formic acid. Nicotinamide adenine dinucleotide (NAD] is a coenzyme that takes up a hydrogen,
going from NAD to NADH+. This removes the hydrogen in the second and third steps, to result in carbon dioxide.
5-56
-------
Chapter 5- Chemical Mixing
original compounds, and thus have different physicochemical properties (Schwarzenbach et al..
20021.1
5.8.5 Fate and Transport of Chemical Mixtures
Spills during the chemical mixing stage are often present as mixtures of chemicals. Additives are
often mixtures of a few to several chemicals, possibly highly concentrated, and hydraulic fracturing
fluids are often dilute mixtures of several additives. Chemical mixtures can act differently in the
environment than individual chemicals. Individual chemicals can affect the fate and transport of
other chemicals in a mixture primarily by changing their physicochemical properties and
transformation rates.
Chemical mixtures can be more mobile than individual chemicals due to cosolvency, which
increases solubility in the aqueous phase. Methanol and ethanol are examples of cosolvent alcohols
used frequently in hydraulic fracturing fluids ("U.S. EPA. 2015a"). The presence of either greatly
increases BTEX solubility ("Rasa etal., 2013; Corseuil etal., 2011; Heermann and Powers. 19981.2 By
increasing solubility, ethanol can affect the fate and transport of other compounds. For example,
BTEX has been observed to travel farther in the subsurface in the presence of ethanol ("Rasa etal.,
2013: Corseuil etal.. 2011: Corseuil et al.. 2004: Powers etal.. 2001: Heermann and Powers. 19981.
The presence of surfactants lowers fluid surface tension and increases solubility of organic
chemicals. Surfactants can mobilize less soluble/less mobile organic chemicals. Two common
surfactants reported in disclosures in the EPA FracFocus 1.0 project database were 2-
butoxyethanol (CASRN 111-76-2, 21% of disclosures) andpoly(oxy-l,2-ethanediyl)-nonylphenyl-
hydroxy (mixture) (CASRN 127087-87-0, 20% of disclosures). Additionally, surfactants can
mobilize bacteria in the subsurface, which can increase the impact of pathogens on drinking water
resources (Brown and Taffe. 2001).
When chemicals are present as mixtures, one chemical can decrease or enhance the
biodegradability of another through inhibition or co-metabolism. The process of inhibition can slow
biodegradation of each of the chemicals present. For example, the biodegradation of ethanol and
methanol can slow the biodegradation rate of BTEX or other organic chemicals present (Rasa etal.,
2013; Powers et al.. 2001). Co-metabolism can increase the biodegradation rate of other chemicals.
For example, when methane or propane is present with tetrachloroethylene, the enzyme produced
by bacteria to degrade methane also degrades tetrachloroethylene (e.g., Alvarez-Cohen and Speitel,
2001 and references therein). For the purposes of chemicals used in hydraulic fracturing, the
presence of other chemicals in additives and hydraulic fracturing fluids could result in increased or
decreased biodegradation if the chemicals are spilled.
1A polar molecule is a molecule with a slightly positive charge at one part of the molecule and a slightly negative charge
on another. The water molecule, H2O, is an example of a polar molecule, where the molecule is slightly positive around the
hydrogen atoms and negative around the oxygen atom.
2 BTEX is an acronym for benzene, toluene, ethylbenzene, and xylenes. These chemicals are a group of single ringed
aromatic hydrocarbons based on the benzene structure. These compounds are found in petroleum and are of specific
importance because oftheir potential health effects.
5-57
-------
Chapter 5 - Chemical Mixing
5.8.6 Site and Environmental Conditions
Environmental conditions at and around the spill site affect the movement and transformation of
chemicals. This section discusses the following: site conditions (e.g., proximity, land cover, and
slope), soil conditions (e.g., permeability and porosity), and weather and climate.
The proximity of a spill to a drinking water resource, either laterally in the case of a surface water
body or downward for groundwater, affects the potential for impact and its severity. Land cover
will affect how readily a fluid moves over land. For example, more rugged land cover such as forest
can impede flow, and an asphalt road can facilitate flow. A spill that occurs on or near a sloped site
can move overland faster, increasing the potential to reach nearby surface water. Flatter surfaces
result in a greater chance for infiltration to the subsurface, which could increase the potential for
groundwater impact
Soil characteristics that affect the transport and transformation of spill chemicals include soil
texture (e.g., clay, silt, sand), permeability, porosity, and organic content12 Fluids will move more
quickly through permeable soil (e.g., sand) than through less permeable soil (e.g., clay). A soil with a
high porosity provides more volume to hold water and spilled chemicals. Another important factor
for a site is the organic content, of which there are two competing types: soil organic carbon and
dissolved organic carbon. Each type of carbon acts as a strong substance for chemicals to associate
with. Soil organic carbon present in a solid phase, such as dead and decaying leaves and roots, is not
mobile and slows the movement of chemicals through the soil. Dissolved organic carbon (DOC)
moves with the water and can act as a shuttling mechanism to mobilize less soluble chemicals
across the surface and through the subsurface. Chemicals may also associate and move with
particulates and colloids.
Weather and climate conditions affect the fate and transport of a spilled chemical. After a spilled
chemical stops moving, precipitation can remobilize the chemical. The amount, frequency, and
intensity of precipitation will impact the volume, distance, and speed of chemical movement
Precipitation can carry chemicals downward or overland, and it can cause erosion, which can move
sorbed chemicals overland.
5.8.7 Peer-Reviewed Literature on the Fate and Transport of Hydraulic Fracturing Fluid Spills
There has been limited peer-reviewed research investigating the fate and transport of chemicals
spilled at hydraulic fracturing sites. Aminto and Olson T20121 modeled a hypothetical spill of
1,000 gal (3,800 L) of hydraulic fracturing fluid using equilibrium partitioning. The authors
evaluated how 12 chemicals typically used for hydraulic fracturing in the Marcellus Shale would
partition among different phases: air, water, soil, and biota.3 They presented a ranking of
1 Permeability of a soil describes how easily a fluid can move through the soil. Under a constant pressure, a fluid will move
faster in a high permeability soil than the same fluid in a low permeability soil.
2 Porosity of a soil describes the amount of empty space for a given volume of soil. The porosity describes how much air,
water, or hydraulic fluid a given volume of soil can hold.
3 The chemicals they investigated included: sodium hydroxide, ethylene glycol, 4,4-dimethyl oxazolidine, 3,4,4-trimethyl
oxazolodine, 2-amino-2-methyl-l-propanol, formamide, glutaraldehyde, benzalkonium chloride, ethanol, hydrochloric
acid, methanol, and propargyl alcohol.
5-58
-------
Chapter 5- Chemical Mixing
concentrations for each phase. In water, they showed that sodium hydroxide (a pH buffer), 4,4-
dimethyl oxazolidine (a biocide), hydrochloric acid (a perforation clean-up additive), and 3,4,4-
trimethyl oxazolidine (a biocide) had the highest simulated water concentrations; however, these
concentrations depended on the chemicals included in the simulated mixture and the
concentrations of each. Their analysis suggested that after a spill, a large fraction of the spill would
volatilize and leave the soil; however, some constituents would be left behind in the water, soil, and
biota compartments, which could act as long-term contamination sources. Aminto and Olson
f20121 only studied this one scenario. Other scenarios could be constructed with different
chemicals in different concentrations. These scenarios may result in different outcomes and
impacts. Any spill would require site- and spill-specific modeling on a case-by-case basis. For this
reason, we cannot make any general statement about fate and transport of hydraulic fracturing
chemicals and fluids. For this reason, we cannot make any general statement about fate and
transport of hydraulic fracturing chemicals and fluids.
Drollette etal. (2015) suggested a link between surface spills and groundwater contamination,
possibly from hydraulic fracturing activity, because the chemicals detected were hydraulic
fracturing additives. This work demonstrates the pathway for surface spills to impact groundwater
sources. They detected low levels of gasoline related organic chemicals with elevated diesel range
organic chemicals, which suggests that the former were degraded or volatilized, while the latter
were more persistent and penetrated into the subsurface and into groundwater.
5.8.8 Potential and Documented Fate and Transport of Documented Spills
There is limited information on the fate and transport of hydraulic fracturing fluids and chemicals.
This section highlights both potential and documented impacts for three reported spills (U.S. EPA.
2015ml. In each case, we provide the documented and potential paths (surface, subsurface, or
combination) and the associated fate and transport governing processes by which the spill has been
documented or has the potential to have an impact on drinking water resources. The three cases
involve a tank overflow with a reported surface water impact, a human error blender spill with a
reported soil impact, and an equipment failure that had no reported impact. We specifically chose
these three spills to highlight three different cases. One demonstrates a documented impact with a
demonstrated pathway that had an observed effect on a nearby drinking water resource. The
second case shows how a release can impact an environmental receptor with a pathway for
potential impact on a drinking water resource, but there was no observed impact The third
example is a spill that was contained and cleaned up resulting in likely no impact None of these
chemical releases have any documented pre- or post-sampling. No information on the specific
chemicals spilled or the concentrations or total mass of any chemical is provided. We cannot
provide any quantitative assessment from these observed cases.
In the first documented spill, shown in Figure 5-17, a tank overflowed twice, releasing a total of
7,350 gal (980 ft3, 28 m3, or 27,800 L) of friction reducer and gel fPADEP. 2012. ID#18301631.1 The
spill traveled across the land surface, crossed a road, and then continued to a nearby stream. The
1 We provide the total volume of the spill in gallons as well as cubic length (cubic feet and cubic meters], because it may
be a little harder to visualize how far a volume of 7,300 gal (28,000 L] might travel.
5-59
-------
Chapter 5 - Chemical Mixing
spill affected wetlands and a stream, where fish were reported to have been killed. The fish kill
indicates an observable impact. This represents a good example for how environmental conditions
can affect the severity and timing of impact, due to the slope of the lands surface, the permeability
of the soil, and the proximity to surface water. We are not aware of any measurements performed
for soils, groundwater, surface water, sediments, or fish tissue. Based on the publicly available
information, we do not know what chemicals were in the friction reducer and gel, which limits
further assessment
Spilled Hydraulic
Fracturing Fluid
Transformation
Sorption
chemicals potentially penetrating soil
•¦¦ layer above ground water J
Ground
Water
potential chemical
ground water plume
Dissolution
Surface
Tank overflow spill with documented impact to surface water:
. 2 spills, 7,350 gal fluid: water, friction reducer, gel
w Documented Impact:
A Fluid reached wetlands and creek
Fish stressed and killed
Tank
Overflowed
Volatilization
Flowed
Over Land
Across Road
Immediate
Impact
HHHMIM
Figure 5-17. Fate and Transport Spill Example: Case 1.
Spills information from PA PEP (2012, ID#1830163).
For this first spill, the documented path was overland flow from the tank to the stream with a
documented, immediate impact There are also other potential paths for potential impacts on
drinking water resources. The spilled chemicals could have penetrated into the soils or sorbed to
soils and vegetation as the fluid moved across the ground towards the stream. Chemicals could then
be mobilized during later precipitation, runoff, or erosion events. Chemicals that infiltrated the
subsurface could serve as long-term sources, travel laterally across the unsaturated zone, or
continue downwards to groundwater. Some chemicals could be lost to transformation processes.
The absence of reported soil or groundwater sampling data prevents the ability to know if these
potential paths occurred or not
The second documented spill, shown in Figure 5-18, occurred when a cap was left off the blender,
and 504 gal (70 ft3 or 2 m3) of biocide and hydraulic fracturing fluid were released fCOGCC. 2012.
5-60
-------
Chapter 5 - Chemical Mixing
ID#26089001. In addition, 294 gal (39 ft3 or 1.1 m3) were retained by a dike with a lined secondary
containment measure, demonstrating the partial effectiveness of this containment mechanism. The
remaining 210 gal (28 ft3 or 0.8 m3) of fluid (biocide and water) ran off-site. Of this, 126 gal were
vacuumed, leaving 84 gal. There was no documented impact on surface or groundwater. However,
potential impacts potentially could have occurred.
Transformation
Unsaturated
Soil
Sorption
chemicals potentially penetrating soil
layer above ground water -y 'i
Ground
Water
potential chemical
ground water plume
Dissolution
Cap left off of blender; biocide and water spill
504 gal spilled: 294 gal caught by dike, 210 gal ran off-site
Blender
Volatilizatio,
Figure 5-18. Fate and Transport Spill Example: Case 2.
Spills information from COGCC (2012, ID#2608900).
In this second case, the uncontained 84 gal could have infiltrated the subsurface, creating a
potential path to groundwater. Highly mobile chemicals could have penetrated the soil more
quickly than less mobile chemicals, which would have sorbed to soil particles. As the chemicals
penetrated into the soil, some could have moved laterally in the unsaturated zone, or traveled
downward to the groundwater table and moved with direction of groundwater flow. These
chemicals could have served as a long-term contamination source. The chemicals also could have
transformed into other chemicals with different physicochemical properties, and any volatile
chemicals could have moved to the air as a loss process. As in the first case, there was no reported
sampling of soil or groundwater, so there is no way to know if chemicals did or did not follow any of
these pathways. We do not have any more information on the types of chemicals present or the
concentrations with which they were present, which limits further assessment
In the third documented spill, shown in Figure 5-19, 630 gal (84 ft3 or 2.4 m3) of crosslinker spilled
onto the well pad when a hose wore off at the cuff (COGCC. 2012. ID#13958271. The spill was
5-61
-------
Chapter 5 - Chemical Mixing
contained in the berm and an on-site vacuum truck was used to clean up the spill. No impact on soil
or water was reported.
Equipment failure:
Hose worn off
nsaturated
Soil
Hose worn at cuff of blender; crosslinker spill
630 gal spilled onto pad. Prevented infiltration
of fluid into soil.
Volatilization
Onsite vac-truck
remediated spill
Ground
Water
Figure 5-19 Fate and Transport Spill Example: Case 3.
The pad may or may not have had a liner. Spills information from COGCC (2012, ID#1395827).
For this third case, we do not have any information on whether the well pad was lined or not. If the
site had a liner, the spill could have been fully contained and cleaned up. Without a liner or if the
integrity of the liner was compromised (e.g., had a tear], any residual chemical that was not
effectively cleaned up could have remained in the soil. This would create potential paths similar to
those above in the second case, where the chemicals could have sorbed to the soils and penetrated
into the subsurface and possibly reach groundwater. There was no reported sampling of soil or
groundwater to determine whether or not chemicals migrated into the soil, and we do know the
types of chemicals or the concentrations of the released chemicals.
5.8.9 Challenges with Unmonitored and Undetected Chemicals
One of the challenges confronting a thorough assessment of the fate and transport of spilled
hydraulic fracturing chemicals lies in the lack of documented observations. It is difficult to prove
absence of impact, and absence of observations does not necessarily imply lack of impact. Also, we
know there are over 1,000 different chemicals reported used in hydraulic fracturing (Section 5.4),
and this number is increasing. For many chemicals, there is not an analytical technique available to
detect them in samples taken to a laboratory. Due to the lack of information on the chemicals used
5-62
-------
Chapter 5- Chemical Mixing
on site (some of which are claimed as CBI), one would not know what chemicals to include in the
lab analysis. Hydraulic fracturing chemicals are typically present as complex mixtures, which also
complicates sample analysis. Chemicals can transform upon release, which can result in different
chemicals in the environment than those originally released. Even if chemicals are detected on-site,
it can be difficult to demonstrate a direct linkage to hydraulic fracturing operations, since many of
the chemicals used in hydraulic fracturing are also used for other purposes (such as gasoline or
diesel from vehicles). Since there are currently no requirements for a detection-monitoring
network to assess the occurrence and extent of chemical releases from the well pad, it is not
possible to conclusively assess the frequency and impact of fluid releases during the chemical
mixing process.
5.9 Trends in the Use of Hydraulic Fracturing Chemicals
Hydraulic fracturing science and engineering continues to advance. A part of this research includes
using different chemicals. This section provides an overview of the changes in chemical use, with an
emphasis on efforts to reduce potential impacts from surface spills by using fewer and safer
chemicals. Reasons for changing the types of chemicals used can include: improving the fracturing
process, using greener/safer chemicals, and reducing overall cost
Representatives from oil and gas companies, chemical companies, and non-profits are working on
strategies to reduce the number and volume of chemicals used and to identify safer chemicals
fWaldron. 20141. Southwestern Energy Company, for example, is developing an internal chemical
ranking tool fSWN. 20141. and Baker Hughes is working on a hazard ranking system designed for
wide-scale external use (Baker Hughes. 2014: Brannon etal.. 2012: Daulton etal.. 2012: Brannon et
al.. 20111. Environmental groups, such as the Environmental Defense Fund, are also developing
hazard rating systems (Penttila etal.. 20131. Typical criteria used to rank chemicals include
mobility, persistence, biodegradation, bioaccumulation, toxicity, and hazard characteristics. In this
assessment, toxicity and a methodology to rank chemical hazards of hydraulic fracturing chemicals
is discussed in Chapter 9.
Given that human error is the cause of 25% of chemical mixing related spills and spill prevention
can never be 100% effective, changes to the types of chemicals used could reduce the frequency or
the severity of potential impacts. Using chemicals with specific physicochemical properties that
affect the fate and transport of chemicals could reduce their potential impacts. Less mobile
chemicals could make cleanup of spills easier. For example, using dry chemicals that are hydrated
on-site could minimize impacts if there were a container failure. Using chemicals with lower
persistence and higher biodegradability, if spill prevention and cleanup are not fully effective,
would lessen the severity of potential impact Use of less hazardous chemicals could lessen impact
in cases where a spill reaches a drinking water resource.
The EPA has not conducted a comprehensive review of efforts to develop safer hydraulic fracturing
chemicals. However, the following are some specific examples of efforts that companies cite as part
of their efforts toward safer chemical use:
• A renewable citrus-based replacement for conventional surfactants (Fisher. 20121:
5-63
-------
Chapter 5 - Chemical Mixing
• A crosslinked gel system comprised of chemicals designated as safe food additives by the
U.S. Food and Drug Administration (Holtsclaw etal.. 2011):
• A polymer-free gel additive (Al-Ghazal etal.. 2013):
• A dry, hydrocarbon-free powder to replace liquid gel concentrate fWeinstein etal.. 20091:
• Biodegradable polymers fIrwin. 20131:
• The use of ultraviolet light to control bacteria fRodveltetal.. 20131:
• New chelating agents that reduce the use of strong acids (LePage etal.. 2013):
• Eco-friendly viscoelastic surfactant (VES) polymer-free fluid reduces fracture cleanup time
with 95% retrieved fluids compared to 40 - 60% and is less toxic than polymer-based
fluids (AlKhowaildi etal.. 2016): and
• The recovery and reuse of produced water as hydraulic fracturing fluids, which can reduce
the need to add additional chemicals (Horn etal.. 2013).
A review of the EPA's new chemicals program found that, from 2009 to April 2015, the Agency
received pre-manufacturing notices (PMN) for about 110 chemicals that have the potential for use
as additives. Examples include chemicals intended for use as clay control agents, corrosion
inhibitors, gel crosslinkers, emulsifiers, foaming agents, hydrate inhibitors, scale inhibitors, and
surfactants. At the time of PMN submission, these chemicals were not in commercial use in the
United States. As of April 2015, the EPA had received 30 notices of commencement, indicating that
some of the chemicals are now used commercially.
As different hydraulic fracturing fluids are developed, they have corresponding effects on different
stages of the hydraulic fracturing water cycle. For example, in Figure 5-4(b) an example of an
energized fluid uses a total water volume of 105,000 gal (397,000 L), which means less water is
required in the water acquisition stage and less produced water results in less wastewater. Figure
5-4(a) shows slickwater with 4,763,000 gal (18,030,000 L) of water, yet a larger fraction of
slickwater may be reused, reducing the need for more water for another frac job and requiring the
treatment of less wastewater.
5.10 Synthesis
The chemical mixing stage includes the mixing of base fluid, proppant, and additives on the well pad
to make hydraulic fracturing fluid. This chapter provided an analysis of the factors affecting
potential impacts on drinking water resources during the chemical mixing stage of the hydraulic
fracturing water cycle and the factors governing the frequency and severity of these impacts.
5.10.1 Summary of Findings
Reports have demonstrated that spills and releases of chemicals and fluids have occurred during
the chemical mixing stage and have reached soils and surface water receptors. Spill reports have
not documented impacts on groundwater related to the chemical mixing stage. Spill reports have
little information on post-spill testing and sampling. Impacts on groundwater may remain
undocumented. The potential pathway for impact on groundwater has been demonstrated and
5-64
-------
Chapter 5- Chemical Mixing
documented for chemicals spilled during other parts of the hydraulic fracture water cycle.
(Evidence of groundwater impact from produced water spills is discussed Chapter 7.)
The hydraulic fracturing fluid generally consists of a base fluid (typically water), a proppant
(typically sand), and additives (chemicals), although there is no standard or single composition of
hydraulic fracturing fluid used. According to the analysis of the EPA FracFocus 1.0 project database,
based on FracFocus disclosure data from January 2011 to February 2013, approximately 93% of
hydraulic fracturing fluids use water as a base fluid. Non-aqueous fluids, such as nitrogen, carbon
dioxide, and hydrocarbons, are also used as base fluids or used in combination with water as base
fluids. The number of chemicals injected into a well typically ranges from 4 to 28, with a median of
14 fU.S. EPA. 2015al In water-based hydraulic fracturing, the composition, by volume, of a typical
hydraulic fracturing fluid is 90% to 97% water, 2% to 10% proppant, and 2% or less additives
(Carter etal.. 2013: Knappe and Fireline. 2012).
The EPA has identified 1,084 different chemicals used in chemical mixing. A recent study of
FracFocus disclosure data, covering January 2011 to April 2015, has reported 263 new CASRNs,
increasing the number of chemicals identified for use by approximately 24% fKonschnik and
Davalu. 2016). Hydraulic fracturing chemicals cover a wide range of chemical classes and a wide
range of physicochemical properties. The chemicals include acids, aromatic hydrocarbons, bases,
hydrocarbon mixtures, polymers, and surfactants. The use of 32 chemicals, excluding water, quartz,
and sodium chloride, is reported in 10% or more of disclosures in the EPA FracFocus 1.0 project
database. The ten most common chemicals (excluding quartz) are methanol, hydrotreated light
petroleum distillates, hydrochloric acid, isopropanol, ethylene glycol, peroxydisulfuric acid
diammonium salt, sodium hydroxide, guar gum, glutaraldehyde, and propargyl alcohol (U.S. EPA.
2015c). These chemicals can be present in multiple additives. Methanol, hydrotreated light
petroleum distillates, and hydrochloric acid are the three chemicals reported to be used in more
than half of all hydraulic fracturing jobs, with methanol being used at 72% of all sites.
An EPA analysis of spills data (January 2006 to April 2012, from nine states, nine service
companies, and nine operators) identified over 36,000 spills, with 457 spills (~1%) that were on or
near the well pad and definitively associated with hydraulic fracturing. Of these spills, 151 were of
chemicals or hydraulic fracturing fluid and thus assumed to be associated with the chemical mixing
stage. Chemical spills during the chemical mixing stage were primarily caused by equipment failure
(34%), followed by human error (25%), although 26% spills had an unknown source. The
remaining spills were caused by a failure of container integrity, weather, vandalism, and well
communication. Reported spills covered a large range of volumes, from 5 to 19,320 gal (19 to
73,130 L), with a median of 420 gal (1,600 L) flJ.S. EPA. 2015ml.
The rate of reported spills during the hydraulic fracturing water cycle is estimated to range from
0.4 to 12.2 reported spills for every 100 wells, based on spills data from North Dakota,
Pennsylvania, and Colorado, with a median rate of 2.6 reported spills for every 100 wells (See
Appendix C). The estimated rates provide an approximate estimate of the potential frequency of the
number of spills at a site. It is uncertain how representative these rates are of national spill rates or
rates in other states. These numbers are not specific to the chemical mixing stage. In 2015, there
5-65
-------
Chapter 5 - Chemical Mixing
are 2.6 reported spills occurring during the chemical mixing stage per 100 wells hydraulically
fractured in North Dakota.
The total volume of chemicals used on site are estimated to range from 2,600 to 30,000 gal (9,800
to 114,000L). An estimate for the mean volume for any chemical used on-site is 650 gal (2,500 L)
with a mean mass of 1500 kg (3,200 lb). An estimate of 2,300 to 6,500 gal (8,800 to 25,000 L) of
additives are stored on site, typically in multiple totes of 200 to 375 gal (760 to 1,420 L). These
volumes provide insight on how much potentially could spill at any given hydraulic fracturing site
and what the volume of a spill might be depending on where/when it occurs during the chemical
mixing process.
The potential of spills to reach drinking water resources depends on site and chemical properties.
The fate and transport of spilled hydraulic fracturing chemicals is complex, particularly because
chemicals are generally present as diverse, complex mixtures. There are different pathways for a
spill to reach ground and surface water and to serve as a long term source. Roughly 40% of
hydraulic fracturing chemicals are organic chemicals, which have physicochemical properties that
cover the parameter space, from fully miscible to insoluble and from highly hydrophobic to highly
hydrophilic. Of the 20 most frequently used chemicals used at hydraulic fracturing sites, three
chemicals have low mobility: hydrotreated light petroleum distillates, heavy aromatic petroleum
solvent naphtha, and naphthalene. These chemicals have the potential to act as long term sources of
contamination if spilled on-site.
5.10.2 Factors Affecting the Frequency or Severity of Impacts
The specific factors that have the potential to affect the frequency and severity of impacts include
the size and type of the fracturing operation; volume, mass, and concentration of chemicals spilled;
type of chemicals and their properties; combination of chemicals spilled; environmental conditions;
proximity to drinking water resources; employee training and experience; quality and maintenance
of equipment; and spill containment and mitigation.
The size and type of a fracturing operation, including the number of wellheads, the depth of the
well, the length of the leg(s), and the number of stages and phases, affect the potential frequency
and severity spills. Larger operations can require larger volumes of chemicals, more storage
containers, more equipment, and additional transfers between different pieces of equipment
Larger storage containers increase the maximum volume of a spill or leak from a storage container.
Additional transfers between equipment increase the possibility of human error and potential
frequency of spills.
The volume, mass, and concentration of spilled chemicals affect the frequency and severity of
impacts. A larger volume increases the potential for a spill to travel a longer distance and reach a
drinking water resource. The severity of the spill will be affected by the spill volume, the total mass
of chemicals released, and the concentration with which it reaches the drinking water resource.
The type of chemicals spilled affects how the chemicals will move and transform in the
environment and the type of impact it will have on a drinking water resource. More mobile
chemicals move faster through the environment, which can increase the frequency of impact More
5-66
-------
Chapter 5- Chemical Mixing
soluble chemicals can reach a drinking water resource at higher concentrations, thereby increasing
the potential severity of an impact Less mobile chemicals will move more slowly, and can have
delayed and longer-term impacts at lower concentrations. The potential severity of impact is
affected by how the chemical adversely impacts water quality. Some chemicals can have severe
impacts at low concentrations, while some chemicals can have minimal impacts even at high
concentrations. Water quality impacts can range from aesthetic effects (e.g., taste, smell) to adverse
health effects.
The environmental conditions at and around the spill site affect the fate and transport of a given
chemical and thus affect the frequency of impacts as well as potential severity. Conditions include
soil properties, climate, weather, and terrain. Permeable soils allow for rapid transport of the
spilled fluid through the subsurface and to groundwater. The presence of preferential flow paths
(e.g., fractures, animal burrows) may provide rapid transport through the subsurface in what might
appear to have low permeability. The presence of complexing agents and colloids may further
increase transport of less soluble chemicals. Precipitation can re-mobilize trapped chemicals and
move them over land or through the subsurface.
The proximity of a spill to drinking water resources affects the frequency and severity of impact.
The closer a spill is to a drinking water resource, the higher the potential to reach it As a fluid
moves toward a drinking water resource, it can decrease in concentration, which can reduce the
severity of an impact. The characteristics of the drinking water resource will also influence the
severity of the impact of a spill. For example, a slow release into a fast moving stream will result in
large dilution and lower concentrations of chemicals (less severe impact). The transport of a
chemical to groundwater may have a more severe impact, as there may be less dispersion of the
chemical (higher concentrations in the groundwater, more severe impact) and the chemical could
serve as a long-term source of contamination (resulting in a chronic exposure versus an acute
exposure).
Effective spill containment and mitigation measures can prevent or reduce the frequency and
severity of impacts. Spill containment measures include well pad containment liners, diversion
ditches, berms, dikes, overflow prevention devices, drip pans, and secondary containers. These may
prevent a spill from reaching soil and water receptors. Spill mitigation, including removing
contaminated soils, vacuuming up spilled fluids, and using sorbent materials can limit the severity
of a spill. It is unclear how effective these practices are and to what extent they are implemented.
5.10.3 Uncertainties
The lack of information and the uncertainty around information having to do with the composition
of additives and fracturing fluids, containment and mitigation measures in use, the proximity of
chemical mixing to drinking water resources, and the fate and transport of spilled fluids limits our
ability to fully assess potential impacts on drinking water resources and the factors affecting their
frequency and severity.
There is no standard design for hydraulic fracturing fluids. Detailed information on the chemicals
used is limited. Volumes, concentrations, and mass, as well as the identity of some of chemicals
5-67
-------
Chapter 5 - Chemical Mixing
stored on-site, are generally not publicly available. The FracFocus national registry, which currently
holds the most comprehensive information on water and chemicals used in hydraulic fracturing
fluids, is structured so as to input chemical information as a maximum percentage of the mass of
fracturing fluid and the given additive. This does not provide exact information on the volume of a
chemical, the mass of a chemical, or the actual composition of an additive. The accuracy and
completeness of original FracFocus disclosure information has not been verified. In applying the
EPA-standardized chemical list to the ingredient records in the EPA FracFocus 1.0 project database,
standardized chemical names were assigned to only 65% of the ingredient records from the more
than 36,000 unique, fully parsed disclosures. The remaining ingredient records could not be
assigned a standardized chemical name and were excluded from analyses (U.S. EPA. 2015a).
Operators may specify certain ingredients as confidential business information (CBI) and not
disclose the chemical used. More than 70% of disclosures in the EPA FracFocus 1.0 project database
contained at least one CBI chemical. Of disclosures with at least one CBI chemical, the average
number of CBI chemicals per disclosure was five. Approximately 11% of all chemicals reported in
the disclosures in the EPA FracFocus 1.0 project database were reported as CBI fU.S. EPA. 2015al
The rate of withholding in FracFocus 2.0 data has increased to 16.5% fKonschnik and Davalu.
2016). No data are available in FracFocus disclosures for any chemical listed as CBI. Therefore,
chemicals identified as CBI in FracFocus disclosures are not included in any of the analyses in this
assessment including estimates of chemical volume, physicochemical properties, or frequency of
use. It is feasible that the same chemicals are repeatedly reported as CBI. Each reported CBI
chemical could also be unique, which would mean there is a very large number of chemicals that we
know nothing about This results in an unknown amount of uncertainty regarding CBI chemicals
and their potential impact on drinking water resources.
Of the 1,084 hydraulic fracturing fluid chemicals identified by the EPA, 629 were inorganic
chemicals, mixtures, or polymers, and thus they did not have estimated physicochemical properties
reported in the EPI Suite™ database. Knowing the chemical properties of a spilled fluid is essential
to predicting how and where it will travel in the environment. Although we can make some
generalizations about the physicochemical properties of these chemicals and how spilled chemicals
may move in the environment, the distribution of properties could change if we obtained data for
all known fracturing fluid chemicals (as well as for those listed as CBI).
There has been limited research on the fate and transport of spilled chemicals on site. We have
provided a limited overview discussing the processes that may be important, but the processes are
complex. There is great uncertainty in how these chemicals will move in the environment. These
processes are complicated by the data gaps in fluid characteristics, especially present in mixtures,
and there is limited understanding on how chemicals act in mixture in the environment. Hydraulic
fluid mixtures are different than other previously studied mixtures (like petroleums, coal tars, and
polychlorinated biphenyls (PCBs). Those mixtures are of chemicals of similar classes, while
hydraulic fracturing fluids are chemicals covering a range of different chemical classes.
There is a lack of field data at hydraulic fracturing sites. There is a lack of baseline ground and
surface water quality data. This lack of data limits our ability to assess the relative change to water
quality from a spill or attribute the presence of a contaminant to a specific source. There is a lack of
5-68
-------
Chapter 5- Chemical Mixing
publicly or readily accessible sampling of soils and groundwater after a fracturing job is complete.
The lack of data and uncertainty on what chemicals are used for hydraulic fracturing makes it
unclear what chemicals to measure. Further uncertainty lies in the limited analytical techniques for
chemicals used in hydraulic fracturing.
There are uncertainties and data gaps in the current information on spills. The EPA spills report
included data from January 2006 to April 2012 from nine states, nine service companies, and nine
oil and gas production well operators fU.S. EPA. 2015al. This data contained over 36,000 reported
spills. From this data set, only 457 were determined to be definitively associated with hydraulic
fracturing and occurred on or near the well pad. With these data, it is impossible to know if all these
spill reports capture all spills occurring at hydraulic fracturing sites. The available data might not
extrapolate to the rest of the nation. Spill reports had limited information on spill causes,
containment and mitigation measures, and sources of spills. The actual chemicals spilled, the total
mass, and the composition are generally not included. There are little available data on impacts of
spills, due to a lack of baseline data and incomplete documentation of follow-up actions and testing.
In general, then, we are limited in our ability to fully assess potential impacts on drinking water
resources from chemical spills, based on current available information. To improve our
understanding we need: more information on the chemical composition of additives and fracturing
fluids and the physicochemical properties of chemicals used; baseline monitoring and field studies
of spilled chemicals; ground and surface water drinking water resources located and identified,
with quality conditions performed before and after hydraulic fracturing; detailed site-specific
environmental conditions; more information on containment and mitigation measures and their
effectiveness; and more detail on the characteristics of spills, such as the exact chemicals and the
amount spilled (mass, concentration, volume).
5.10.4 Conclusions
This chapter discusses the factors that affect the potential for the chemical mixing stage of the
hydraulic fracturing water cycle to impact drinking water resources. Reports have demonstrated
that spills and releases of chemicals and fluids have occurred during the chemical mixing stage and
have reached soils and surface waters with the potential to reach groundwater. The potential for
spilled fluids to reach, and therefore impact, ground or surface water resources depends on the
composition of the spilled fluid, spill characteristics, spill response activities, and the fate and
transport of the spilled fluid. There is no standard composition for a hydraulic fracturing fluid,
which consists of base fluid, proppant, and additives. The EPA identified 1,084 chemicals that have
been reported to be used nationwide, and these chemicals cover a wide variety of chemical classes
and physicochemical properties, and this number is increasing. These chemicals cover a range of
classes and physicochemical properties. The type of fluid and the number, volume, and type of
chemicals used vary from site to site. Hydraulic fracturing fluids generally consist of a mixture of
chemicals, which affects the potential for a release to reach a drinking water resource and the
severity of the potential impact State and industry spill data collected and reviewed by the EPA and
others indicate that small (approximately 30 gal or 100 L) and large spills (greater than 1,000 gal or
4,000 L) can reach surface water resources. While small spills have reached surface water
resources (and have the potential to reach groundwater resources), large volume spills are more
5-69
-------
Chapter 5 - Chemical Mixing
likely to travel longer distances and thus have a greater potential to reach ground and surface water
resources. Large volume spills, particularly of concentrated additives, also have a greater potential
to result in more severe impacts on drinking water resources, because they can deliver a large
quantity of potentially hazardous chemicals to ground or surface water resources.
5-70
-------
Chapter 6 - Well Injection
Chapter 6. Well Injection
Abstract
The well injection stage of the hydraulic fracturing water cycle involves the injection of hydraulic
fracturing fluids through the oil and gas production well and their movement in the production zone.
Subsurface pathways created during this stage—including the production well and newly created
fractures—can allow hydraulic fracturing fluids or naturally occurring fluids to reach groundwater
resources.
This chapter examines two types of pathways by which hydraulic fracturing fluids and liquids and/or
gases that exist in the subsurface can move to, and affect the quality of, subsurface drinking water
resources. First, fluids can move via pathways adjacent to or through the production well as a result of
inadequate design, construction, or degradation of the casing or cement. Second, fluid movement can
occur within the subsurface geologic formations via fractures extending out of oil/gas-containing
formations, by intersecting abandoned or active offset wells, or via naturally occurring faults and
fractures.
The primary factors that can affect the frequency or severity of impacts to drinking water associated
with injection for hydraulic fracturing are: (1) the condition of the well's casing and cement and their
placement relative to drinking water resources, (2) the vertical separation between the production zone
and formations that contain drinking water resources, and (3) the presence/proximity and condition of
wells near the hydraulic fracturing operation.
We identified two cases where hydraulic fracturing activities affected the quality of drinking water
resources due to well construction issues, including inadequate cement or ruptured casing. Additionally,
there are places where oil and gas reservoirs and drinking water resources co-exist in the same
formation and hydraulic fracturing operations occur, which results in the introduction of hydraulic
fracturing fluids into the drinking water resource. There are other cases involving the migration of stray
gas where hydraulic fracturing could be a contributing cause to impacts on drinking water resources.
While there is evidence that these pathways have formed and that groundwater quality has been
impacted, there are limited nationally available data on the performance of wells used in hydraulic
fracturing operations, pre- and post-hydraulic fracturing groundwater quality, and the extent of the
fractures that develop during hydraulic fracturing operations.
These data limits, in combination with the geologic complexity of the subsurface environment and the
fact that these processes cannot be directly observed, make determining the frequency of such impacts
challenging.
6-1
-------
Chapter 6 - Well Injection
This page is intentionally left blank.
6-2
-------
Chapter 6 - Well Injection
6. Well Injection
6.1 Introduction
In the well injection stage of the hydraulic fracturing water cycle, hydraulic fracturing fluids
(primarily water, mixed with the types of chemicals and proppant described in Chapter 5) are
injected into a well under pressure.1 These fluids flow under pressure through the well, then exit
the well and move into the formation, where they create fractures in the rock. This process is also
known as a fracture treatment or a type of stimulation.2 The fractures, which typically extend
hundreds of feet away from the well, are designed to remain within the production zone to access
as much oil or gas as possible by using an appropriate amount of water and chemicals to complete
the operation.3
Production wells are sited and designed primarily to optimize production of oil or gas, which
requires isolating water-bearing formations from hydrocarbon-bearing formations in order to
prevent the water from diluting the hydrocarbons and to protect drinking water resources.4
However, problems with the well's components or improperly sited, designed, or executed
hydraulic fracturing operations (or combinations of these) could adversely impact the quality of
drinking water resources. (Note that, due to the subsurface nature of activities in the well injection
stage, the drinking water resources that may be directly impacted are groundwater resources; see
Chapter 2 for additional information about groundwater.5)
The well and the geologic environment in which it is located are a closely linked system. Wells are
often designed with multiple barriers (i.e., isolation afforded by the well's casing and cement and
the presence of subsurface rock formations) to prevent fluid movement between oil/gas zones and
drinking water resources. Therefore, this chapter discusses (1) the well (including its construction
and operation) and (2) the characteristics of or features in the subsurface geologic formations that
could provide or have provided pathways for migration of fluids to drinking water resources. If
present, and in combination with the existence of a fluid and a physical force that moves the fluid,
these pathways can lead to impacts on the quality of drinking water resources throughout the life of
the well, including during and after hydraulic fracturing.6
1A fluid is a substance that flows when exposed to an external pressure; fluids include both liquids and gases.
2 In the oil and gas industry, "stimulation" has two meanings—it refers to (1] injecting fluids to clear the well or pore
spaces near the well of drilling mud or other materials that block or inhibit optimal production (i.e., matrix treatment]
and (2] injecting fluid to fracture the rock to optimize the production of oil or gas. This chapter focuses on the latter.
3 The "production zone" (sometimes referred to as the target zone or the targeted rock formation] refers to the portion of
a subsurface rock zone that contains oil or gas to be extracted (sometimes using hydraulic fracturing]. "Producing
formation" refers to the larger geologic unit in which the production zone occurs.
4 A subsurface formation (or "formation"] is a mappable body of rock of distinctive rocktype(s] and characteristics (such
as permeability and porosity] with a unique stratigraphic position.
5 Government agencies and other organizations use a variety of terms to describe potable groundwater and groundwater
resources. In this chapter, we use the general term "groundwater resources" to refer to drinking water resources that
occur underground. However, other terms are used in specific contexts to reflect the language used in cited materials.
6 The primary physical force that moves fluids within the subsurface is a difference in pressure. Fluids move from areas of
higher pressure to areas of lower pressure when a pathway exists. Density-driven buoyancy may also serve as a driving
force; see Section 6.3 for more information.
6-3
-------
Chapter 6 - Well Injection
Fluids can move via pathways adjacent to or through the production well that are created in
response to the stresses exerted during hydraulic fracturing operations if the well is not able to
withstand these stresses (Section 6.2). While wells are designed and constructed to isolate fluids
and maximize the production of oil and gas, inadequate construction or degradation of the casing or
cement can allow fluid movement that can impact drinking water quality. Potential issues
associated with wells may be related to the following:
• Inadequate or degraded casing. This may be influenced by the number of casing strings and
the depths to which they are set, compatibility with the geochemistry of intersected
formations, the age of the well, whether re-fracturing is performed, and other operational
factors.
• Inadequate or degraded cement. This may be influenced by a lack of cement in key
subsurface intervals, poor-quality cement, improperly placed cement, or degradation of
cement over time.
Fluid movement can also occur via induced fractures and/or other features within subsurface
formations (Section 6.3). While the hydraulic fracturing operation may be designed so that the
fractures will remain within the production zone, it is possible that, in the execution of the
hydraulic fracturing treatment, fractures can extend beyond their designed extent. Four scenarios
associated with induced fractures may contribute to fluid migration or communication between
zones:
• Flow of injected and/or displaced fluids through pore spaces in adjacent rock formations
out of the production zone due to pressure differences and buoyancy effects.
• Fractures extending out of oil/gas formations into drinking water resources or zones that
are in communication with drinking water resources or fracturing into zones containing
drinking water resources.
• Fractures intersecting artificial structures, including active (producing) or inactive offset
wells near the well that is being stimulated (i.e., well communication) or abandoned or
active mines.
• Fractures intersecting geologic features that can act as pathways for fluid migration, such as
existing permeable faults and fractures.
This chapter describes the conditions that can contribute to or cause the development of the
pathways listed above, the evidence for the existence of these pathways, examples of impacts on the
quality of drinking water resources associated with these pathways that have been documented in
the literature, and the factors that can affect the frequency or severity of those impacts. (See
Chapter 10 for a discussion of factors and practices that can reduce the frequency or severity of
impacts to drinking water quality.)
The interplay between the well and the subsurface features is complex and not directly observable;
therefore, sometimes it is not possible to identify what specific element is contributing to or is the
primary cause of an impact on drinking water resources. For example, concerns have been raised
6-4
-------
Chapter 6 - Well Injection
regarding stray gas detected in groundwater in natural gas production areas (for additional
information about stray gas, see Sections 6.2.2 and 6.3.2.4).1 Stray gas migration is a technically
complex phenomenon, because there are many potential naturally occurring or artificially created
routes for migration of gas into aquifers, including along production wells and via naturally existing
or induced fractures. It is also challenging to determine the source of the natural gas and whether
the mobilization is related to oil or gas production activities.
Furthermore, identifying cases where contamination of drinking water resources occurs due to oil
and gas production activities—including hydraulic fracturing operations—requires extensive
amounts of site and operational data, collected before and after hydraulic fracturing operations.
(See Section 6.4 for additional information on data limitations.) Where such data do exist and
provide evidence of contamination, we present it in the following sections. We do not attempt to
predict which of these pathways is most likely to occur or to lead to a drinking water impact, or the
magnitude of an impact that might occur as a result of migration via any single pathway, unless the
information is available and documented based on collected data. However, a qualitative
assessment of the factors that can affect the frequency or severity of impacts on drinking water
quality associated with the well injection stage is possible; see Section 6.4.
6.2 Fluid Migration Pathways Within and Along the Production Well
In this section, we discuss pathways for fluid movement along or through the production well used
in the hydraulic fracturing operation. While these pathways can form during other times within the
life of an oil and gas well, the repeated high pressure stresses exerted during hydraulic fracturing
operations can make maintaining the mechanical integrity of the well more difficult (Council of
Canadian Academies. 2014).2 Section 6.2.1 presents the purpose of the various well components
and typical well construction configurations. Section 6.2.2 describes the pathways for fluid
movement that can potentially develop within the production well and wellbore and the conditions
that lead to pathway development, either as a result of the original design of the well, degradation
over time or use, or hydraulic fracturing operations.
While we discuss casing and cement separately, it is important to note that these are related—
inadequacies in one of these components can lead to stresses on the other. For example, flaws in
cement may expose the casing to corrosive fluids. Furthermore, casing and cement work together in
the subsurface to form a barrier to fluid movement, and it may not be possible to distinguish
whether mechanical integrity problems are related to the casing, the cement, or both. For additional
information on well design and construction, see Appendix D.
6.2.1 Overview of Well Construction
Production wells are constructed to transport hydrocarbon resources from the reservoirs in which
they are found to the surface. They are also used to isolate fluid-bearing zones (containing oil, gas,
1 Stray gas refers to the phenomenon of natural gas (primarily methane] migrating into shallow drinking water resources
or to the surface.
2 Mechanical integrity of a well refers to the absence of significant leakage within the injection tubing, casing, or packer
(referred to as internal mechanical integrity] or outside of the casing (referred to as external mechanical integrity].
6-5
-------
Chapter 6 - Well Injection
or fresh water) from each other. Multiple barriers (i.e., casing and cement) are often present, and
they act together to prevent both horizontal fluid movement (in or out of the well) and vertical fluid
movement (along the wellbore from deeper oil- or gas-bearing formations to drinking water
resources). Proper design and construction of the casing cement, and other well components in the
context of the location of drinking water resources and maintaining mechanical integrity
throughout the life of a well are necessary to prevent migration of hydraulic fracturing fluids and
formation fluids into drinking water resources.
A well is a multiple-component system that typically includes casing cement, and a completion
assembly, and it may be drilled vertically, horizontally, or in a deviated orientation (Figure 6-1).
These components work together to prevent unintended fluid movement into, out of, or along the
well. Due to the presence of multiple barriers within the well and the geologic system in which it is
placed, the existence of a pathway for fluid movement through a component of this system does not
necessarily mean that an impact on a drinking water resource has occurred or will occur.
land surface
vertical
well
deviated
well
horizontal
well
t
coalbed
with methane
tight sand
gas
confining
sandstone
tight sand
oil
Figure Not to Scale
migration of oil
and gas over
geologic time
Figure 6-1. Schematic cross-section of general types of oil and gas resources and the
orientations of production wells used in hydraulic fracturing.
1 Completion is a term used to describe the assembly of equipment at the bottom of the well that is needed to enable
production from an oil or gas well. It can also refer to the activities and methods (including hydraulic fracturing) used to
prepare a well for production following drilling.
2 For the purposes of this assessment, a well's orientation refers to its inclination from verticality. Wells drilled straight
downward are considered to be vertical, wells drilled directionally to end up parallel to the production zone's bedding
plane are considered horizontal, and directionally drilled wells that are neither vertical nor horizontal are referred to as
deviated. In industiy usage, a well's orientation commonly refers both to its inclination from vertical and the azimuthal
(compass) direction of a directionally drilled wellbores.
6-6
-------
Chapter 6 - Well Injection
Casing primarily acts as a barrier to lateral movement of fluids, and cement primarily acts as a
barrier to unintended vertical movement of fluids. Together, casing and cement are important in
preventing fluid movement into drinking water resources, and are the focus of this section. Figure
6-2 illustrates the configurations and types of casing and cement and other features that may occur
in oil and gas production wells. The figure depicts an idealized representation of the components of
a production well; it is important to note that there is a wide variety in the design of hydraulically
fractured oil and gas wells in the United States fU.S. EPA. 2015nl. and the descriptions in the figure
or in this chapter do not represent every possible well design.
6.2.1.1 Casing
Casing is steel pipe that is placed into the drilled wellbore to maintain the stability of the wellbore,
to transport the hydrocarbons from the subsurface to the surface, and to prevent intrusion of other
fluids into the well and wellbore (Hvne. 2012: Renpu. 2011). A long continuous section of casing is
referred to as a casing string, which is composed of individual lengths of casing (known as casing
joints) that are threaded together using casing collars. In different sections of the well, multiple
concentric casing strings of different diameters can be used, depending on the construction of the
well.
The presence of multiple layers of casing strings can isolate and protect geologic zones containing
drinking water. In addition to conductor casing, which prevents the hole from collapsing during
drilling, one to three other types of casing may be also present in a well. The types of casing include
(from largest to smallest diameter) surface casing, intermediate casing, and production casing
(GWPC. 2014: Hvne. 2012: Renpu. 2011). One or more of any of these types of casing (but not
necessarily all of them) may be present in a well. Surface casing often extends from the wellhead
down to the base (i.e., the bottom or lowest part) of the drinking water resource to be protected.
Wells also may be constructed with production liners, which are anchored or suspended from
inside the bottom of the previous casing string. Production liners serve the same purpose as
production casing but extend only to the end of the previous casing, rather than all the way to the
surface. Wells may also have production tubing, which is used to transport the hydrocarbons to the
surface. Tie-back liners may be used to extend a production liner to the surface when downhole
pressure or corrosive conditions warrant additional protection of the intermediate or production
casing.
Among the wells represented by the Well File Review (described in Text Box 6-1), between one and
four casing strings were present (the Well File Review did not evaluate conductor casings). A
combination of surface and production casings was most often reported, followed by a combination
of surface, intermediate, and production strings. All of the production wells used in hydraulic
fracturing operations in the Well File Review had surface casing, while approximately 39% of the
wells (an estimated 9,100 wells) had intermediate casing, and 94% (an estimated 21,900 wells) had
production casing (U.S. EPA. 2015n).1'2
19,100 wells (95% confidence interval: 2,900 - 15,400 wells].
2 21,900 wells (95% confidence interval: 19,200 - 24,600 wells].
6-7
-------
Chapter 6 - Well Injection
Wells used in hydraulic fracturing operations are often constructed with multiple
layers of casing and cement to isolate fluid-bearing zones from each other.
While there is no "typical well," some of the more common elements are
presented in this figure. This image depicts a well that has multiple layers of
casing and cement. For additional information on the design and construction of
wells, see Appendix D.
Wells are constructed to transport hydrocarbons to the surface, and prevent
unwanted fluid movement into or out of the well and the wellbore (1), which is
the drilled hole into which the well is placed.
Surface casing (2) often extends from the wellhead (3) down to or below the
base (bottom) of the ground water resource (4) to be protected. The surface
casing is cemented from its base to the surface to isolate the ground water
resource and prevent fluid movement. (Wells may also have a conductor casing
(not shown) to prevent unconsolidated material from collapsing into the
wellbore.) The cement shoe (5) controls the placement of cement and prevents
it from flowing back into the casing after the cement has been placed.
When used, intermediate casing (6) can reduce pressure on weak formations or
allow better control of over-pressured formations, and it extends from the
surface through the formation(s) of concern.
The production casing (7) extendsto the end of the wellbore in the production
zone (8) and is cemented in place. In some wells, a production liner is used in
place of production casing. The production liner is hung from the next largest
casing string by a hanger (9) that is attached to the casing and typically is
cemented in place from the surface. Tubing (10), when used, conveys
hydrocarbons to the surface; it is installed after hydraulic fracturing operations
and is not cemented in place. If the well has an open hole completion (not
shown), the production casing extends just into the production zone, and the
entire length of the wellbore through the production zone is uncased.
A packer (11), a mechanical sealing device, maybe set at the lower end of the
tubing to create and seal off the annulus (12), the space between the tubing and
casing (or between two casing strings), and to keep fluid from migrating within
the annulus. Perforations (13) may be made through the casing and cement
using explosive charges. Hydraulic fracturing fluids are pumped down the
innermost casing and through the perforations to create and extend fractures.
Fracturing fluidsare pumped into the well through the tubing, if present, or
through the production casing (if the well has an open hole completion).
Cement (14) protects the casing from exposure to formation fluids, adds
strength to the casing, and, when placed correctly, prevents fluid movement
along the wellbore between different fluid-containing formations.
¦
Note: Figure not to scale; other configurations are possible.
Figure 6-2. Overview of well construction.
6-8
-------
Chapter 6 - Well Injection
Hydraulic fracturing operations impose a variety of stresses on the well components. In order to
prevent the formation of pathways to drinking water resources, the casing should be designed with
sufficient strength to withstand the stresses it will encounter during the installation, cementing,
hydraulic fracturing, production, and post-production phases of the life of the well. These stresses,
illustrated in Figure 6-3, include burst pressure (the interior pipe pressure that will cause the
casing to burst), collapse pressure (the pressure applied to the outside of the casing that will cause
it to collapse), tensile stress (the stress related to stretching exerted by the weight of the casing or
tubing being raised or lowered in the hole), compression and bending (the stresses that result from
pushing along the axis of the casing or bending the casing), and cyclic stress (the stress caused by
frequent or rapid changes in temperature or pressure). While the injection stage represents a
relatively brief portion of the life of a hydraulic fracturing well (Section 3.3), injection imposes the
highest stresses the well is likely to encounter.
Text Box 6-1. The Well File Review.
The EPA conducted a survey of onshore oil and gas production wells that were hydraulically fractured by nine
oil and gas service companies in the continental United States between approximately September 2009 and
September 2010. This effort, known as the "Well File Review," produced two reports. The first report, Review
of Well Operator Files for Hydraulically Fractured Oil and Gas Production Wells: Well Design and Construction
fU.S. EPA. 2015n) describes well design and construction characteristics and their relationships to the
location of operator-reported drinking water resources and the number and relative location of constructed
barriers (i.e., casing and cement) that can block pathways for potential subsurface fluid movement. A second
report, Review of Well Operator Files for Hydraulically Fractured Oil and Gas Production Wells: Hydraulic
Fracturing Operations fU.S. EPA. 2016c) presents information on hydraulic fracturing job characteristics and
the reported use of casing pressure tests, annular pressure monitoring, surface treating pressure monitoring,
and microseismic monitoring conducted before or during hydraulic fracturing operations; it also explores the
roles of well mechanical integrity and induced fracture growth as they relate to the potential for subsurface
fluid movement to intersect protected groundwater resources.
The survey was based on a sample of 323 hydraulically fractured oil and gas production wells. Results of the
research are presented as rounded estimates of the frequency of occurrence of hydraulically fractured
production well design, construction, and operational characteristics with 95% confidence intervals (CIs).
The results are statistically representative of an estimated 23,200 onshore oil and gas production wells
hydraulically fractured in 2009 and 2010 by nine service companies where an estimated 28,500 hydraulic
fracturing jobs were performed.
In addition, the casing must be resistant to corrosion from contact with the formations and any
fluids that might be transported through the casing, including hydraulic fracturing fluids, brines,
and oil or gas. Casing strength or corrosion resistance can be increased by using fiberglass or high-
strength alloys or by increasing the thickness of the casing.
One way to ensure that the strength of the casing is sufficient to withstand the stresses imposed by
hydraulic fracturing operations is to pressure test the casing. The casing can be pressurized to the
pressure anticipated during hydraulic fracturing operations and shut-in periods; if the well can
hold the pressure, it is considered to be leak-free and therefore should be able to withstand the
pressures of hydraulic fracturing. However, if the test pressure is less than the hydraulic fracturing
6-9
-------
Chapter 6 - Well Injection
pressure, the casing is determined to be leak-free, but its suitability to resist the stresses associated
with the planned fracturing operation is less certain.
The Well File Review (U.S. EPA. 2016c) found that pressure tests were performed prior to an
estimated 15,600 of 28,500 hydraulic fracturing jobs the EPA studied, including cases where a frac
string was pressure tested.1 In 52% of those pressure tests performed (representing 28% of the
hydraulic fracturing jobs studied), the well was tested to a pressure equal to or greater than the
maximum pressure that occurred during the hydraulic fracturing job fU.S. EPA. 2016cl.2 Thus, in a
significant number of hydraulic fracturing jobs (i.e., 72% of the wells studied), there are no data in
the well files to indicate that the casing was tested in a manner that could ensure the adequacy of
the casing to withstand the pressures of hydraulic fracturing. While, in some cases, casing may not
have been pressure tested because a frac string was to be installed to protect the casing from the
increased pressure, only 10% of fracturing jobs were conducted using a frac string.
Tensile
H
Compressive
I Collapse
Burst
Note: Figure not to scale
Stresses Exerted on Well Casings
Figure 6-3. The various stresses to which the casing will be exposed.
In addition to the stresses illustrated, the casing will be subjected to bending and cyclic stresses. Source: U.S. EPA
(2012d).
1 15,600 jobs (95% confidence interval: 11,800 - 19,300 jobs],
2 52% of pressure tests (95% confidence interval: 20 - 82% of tests].
6-10
-------
Chapter 6 - Well Injection
6.2.1.2 Cement
Cement is one of the most important components of a well for providing zonal isolation and
reducing impacts on drinking water. Cement in the space between the casing and formation isolates
fluid-containing formations from each other, protects the casing from exposure to formation fluids,
and provides additional strength to the casing. The strength of the cement and its compatibility
with the formation and fluids encountered are important for maintaining mechanical integrity
throughout the life of the well.
A variety of methods are available for placing the cement, evaluating the adequacy of the cementing
process and the resulting cement job, and repairing any identified deficiencies. Cement is most
commonly emplaced by pumping the cement down the inside of the casing to the bottom of the
wellbore and then up the space between the outside of the casing and the formation (or the next
largest casing string). This method is referred to as the primary cement job and can be performed
as a continuous event in a single stage (i.e., "continuous cementing") or in multiple stages (i.e.,
"staged cementing"). Staged cementing may be used when, for example, the estimated weight and
pressure associated with standard cement placement could damage weak zones in the formation
fCrook. 20081.
Deficiencies in the cementing process can result from poorly centered casing, poor removal of
drilling mud behind the casing, cement shrinkage, premature gelation, excessive fluid loss,
improper mixing, or lost cement.1'2 Cement deficiencies can be reduced by proper design of the
cementing process including use of casing centralizers, proper design of the cement, proper mud
removal, and use of cement additives (Kirksev. 2013).3 If any deficiencies or defects in the primary
cement job are identified, remedial cementing may be performed. See Text Box 6-2 for an example
of an incident where cementing issues were studied as part of an evaluation of drinking water well
impacts.
Text Box 6-2. Dimock, Pennsylvania.
In 2009, shortly after drilling and hydraulic fracturing in the Marcellus Shale commenced in the area,
residents near the township of Dimock, Pennsylvania reported that natural gas was appearing or increasing
in their water wells (Hammond. 2016: PA DEP. 2009a).
Water wells in the area largely draw from the Catskill Formation and range in depth from less than 50 ft (15
m) to more than 500 ft (150 m) (Molofsky et al.. 2013). In this area, the Marcellus Shale is about 7,000 ft
(2,000 m) below the surface and its natural gas is extracted through vertical and horizontal wells (Hammond,
2016). Methane exists naturally in the subsurface in this part of Pennsylvania, including in the Catskill
Formation and the geologic formations below it (Baldassare et al.. 2014: Molofsky et al.. 2013: Molofsky et al..
20111
(Text Box 6-2 is continued on the following page.)
1 Gelation is the process in the setting of the cement where it begins to solidify and lose its ability to transmit pressure to
the formation.
2 Lost cement refers to a failure of the cement or the spacer fluid used to wash the drilling fluid out of the wellbore to be
circulated back to the surface, indicating that the cement has escaped into the formation.
3 Centralizers are used to keep the casing in the center of the hole and allow an even cement job.
6-11
-------
Chapter 6 - Well Injection
Text Box 6-2 (continued). Dimock, Pennsylvania.
The Pennsylvania Department of Environmental Protection (PA DEP] investigated and made a determination
that 18 water wells located within a 9 mi2 (23 km2) area had been negatively affected as a result of natural gas
extraction activities. For approximately two years, during which there was a partial ban on gas well drilling
and hydraulic fracturing in the vicinity, the gas company plugged four gas wells and undertook remedial
construction actions at 18 additional gas wells (including remedial cementing at several wells, adding as
much as 6,300 ft (1,900 m) of cement behind the production casings) (PA DEP. 2010b. d, 2009a).
The figure below presents a simplified geologic representation of water wells and one type of horizontal gas
well completed within the geologic formations in the area. The location of remedial cementing performed in
some gas wells is indicated.
Gas Well
Water Well
Approximate depth
below ground (ft)
Water well depths:
50-500 feet
Surface casing
length: 400-1,200 ft
Catski Formation
Other intervening geologic formations
Intermediate casing
length: 1,500-1,900 ft
2,000
Remedial cementing in
well annulus after
initial well construction
Wellbore
J
Casing
Cement, Grout
II
Packer
Marcellus Shale
la a a a b la a h h
7,000
M M IH M M M H
Production
casing
7.300
Not to scale
Several studies in this and surrounding areas have focused on the geochemistry of the groundwater, in
particular on gas composition, and noble and natural gas isotopes in the water. Results are consistent with an
accumulation of stray gas originating from greater depth and moving to the Catskill Formation flackson et al..
2013c: Molofsky et al.. 2013: Molofsky et al.. 2011). However, the identity of the geologic formation(s)
sourcing the natural gas is not always certain and may be consistent with sourcing from either the Marcellus
(as suggested by lackson etal. (2013c)). or the intervening geologic formations (Molofsky et al.. 2013).
The role of hydraulic fracturing in the migration of gas to the Catskill Formation, and the specific pathways by
which this migration occurred, is even less certain. Some investigators suspect that the initial gas well
construction allowed natural gases from deeper formations to move upward along uncemented wellbores
(Hammond. 2016: PA DEP. 2010b. d, 2009a). However, no publicly available information exists to document
whether hydraulic fracturing may have aided fluid movement along wellbores to enter drinking water
resources from greater depths. Reviews of information, such as hydraulic fracturing job reports showing the
intervals hydraulically fractured, injection rates, and pressure monitoring, would support an evaluation of
whether hydraulic fracturing might have played a role in the migration of natural gas to drinking water wells
in the area.
6-12
-------
Chapter 6 - Well Injection
Among the wells represented in the Well File Review, over 90% of cemented casings were
cemented using primary cementing methods. Secondary or remedial cementing was used on an
estimated 8% of casings (most often on surface and production casings and less often on
intermediate casings).1 The remedial cementing techniques employed in these wells included
cement squeezes, cement baskets, and pumping cement down the annulus (U.S. EPA. 2015n). See
Appendix D for more information on remedial cementing techniques.
The cement does not always need to be continuous along the entire length of the well to protect
drinking water resources; rather, protection of drinking water resources depends on a good cement
seal across the appropriate subsurface zones, including all fresh water- and hydrocarbon-bearing
zones. One study of wells in the Gulf of Mexico found that, if at least 50 ft (15 m) of high quality
cement was present, pressure differentials as high as 14,000 psi (97 MPa) would not lead to
breakdown in isolation between geologic zones (King and King. 2013).
Most wells have cement behind the surface casing, which is a key barrier to contamination of
drinking water resources. The surface casings in nearly all of the wells used in hydraulic fracturing
operations represented in the Well File Review (93% of the wells, or an estimated 21,500 wells)
were fully cemented.23 None of the wells studied in the Well File Review had completely
uncemented surface casings.
The length and location of cement behind intermediate and production casings can vary based on
the presence and locations of over-pressured formations, formations containing fluids, or
geologically weak formations (i.e., those that are prone to structural failure when exposed to
changes in subsurface stresses). State regulations and economics also play a role.
In the Well File Review, the intermediate casings of most of the wells studied were fully cemented,
although there were relatively wide 95% confidence intervals in the results. Among production
casings, about half were partially cemented, about a third were fully cemented, and the remainder
were either uncemented or their cementing status was undetermined. Among the approximately
9,100 wells represented in the Well File Review that are estimated to have intermediate casing, the
intermediate casing was fully cemented in an estimated approximately 7,300 wells (80%) and
partially cemented in an estimated 1,700 wells (19%).4 5 Production casings were partially
cemented in 47% of the wells, or approximately 10,900 wells fU.S. EPA. 2015nl.6
18% of casings (95% confidence interval: 3% - 14% of casings].
2 The Well File Review defined fully cemented casings 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]. 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. 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.
3 21,500 wells (95%o confidence interval: 19,500 - 23,600 wells].
4 9,100 wells (95%o confidence interval: 2,900 - 15,400 wells].
5 7,300 wells (95%o confidence interval: 600 - 13,900 wells].
6 10,900 wells (95%o confidence interval: 6,900 - 14,900 wells].
6-13
-------
Chapter 6 - Well Injection
The Well File Review also estimated the number of wells with a continuous cement sheath along the
outside of the well. An estimated 6,800 of the wells represented in the study (29%) had cement
from the bottom of the well to the ground surface, and approximately 15,300 wells (66%) had one
or more uncemented intervals between the bottom of the well and the surface.12 In the remaining
wells, the location of the top of the cement was uncertain, so no determination could be made
regarding whether the well had a continuous cement sheath along the outside of the well (U.S. EPA.
201 5nl.
A variety of logs are available to evaluate the quality of cement behind the well casing. Among wells
in the Well File Review, the most common type of cement evaluation log run was a standard
acoustic cement bond log fU.S. EPA. 2015nl. Standard acoustic cement bond logs are used to
evaluate both the extent of the cement placed along the casing and the cement bond between the
cement, casing, and wellbore. Cement bond indices calculated from standard acoustic cement bond
logs on the wells in the Well File Review showed a median bond index of 0.7 just above the
hydraulic fracturing zone; this value decreased to 0.4 over a measured distance of 5,000 ft (2,000
m) above the hydraulic fracturing zone fU.S. EPA. 2015nl.3 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: Smolen. 2006). One study of 28 wells
found that cement bond logs failed to predict communication between formations 11% of the time
(Boyd etal.. 2006). In addition, they cannot discriminate between full circumferential cement
coverage by weaker cement and lack of circumferential coverage by stronger cement fKing and
King. 2013: Smolen. 20061. 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 in (14 cm) diameter casing with a
bond index of 0.8 along as little as 5 ft (1.5 m) can act as an effective seal. The authors also suggest
that an effective seal in wells having calculated bond indices differing from 0.8 are expected to have
an inverse relationship between bond index and requisite length of the cemented interval, with
longer lengths needed along casing having a lower bond index. Another study recommends that
wells undergoing hydraulic fracturing should have a given cement bond over an interval three
times the length that would otherwise be considered adequate for zonal isolation fFitzgerald etal..
19851. Conversely, King and King f20131 concluded field tests from wells studied by Flournov and
Feaster (1963) had effective isolation when the cement bond index ranged from 0.31 to 0.75.
External mechanical integrity tests (MITs), including temperature logs, noise logs, and radioactive
tracer logs, are another means to evaluate the zonal isolation performance of well cement. Instead
of measuring the apparent quality of the cement, external MITs measure whether there is evidence
of fluid movement along the wellbore (and potentially to a drinking water resource). An external
MIT conducted before the hydraulic fracturing job can allow detection of channels in the cement
that could allow injected fluids to move out of the production zone. An external MIT performed
1 6,800 wells (95% confidence interval: 1,600 - 11,900 wells].
215,300 wells (95% confidence interval: 10,500 - 20,100 wells].
3 Cement bond logs are used to calculate a bond index, which varies between 0 and 1, with 1 representing the strongest
bond and 0 representing the weakest bond.
6-14
-------
Chapter 6 - Well Injection
after hydraulic fracturing operations can detect any fluid movement resulting from cement damage
caused by the hydraulic fracturing job. It is important to note that, if a well fails an MIT, this does
not mean there is a failure of the well or that drinking water resources are impacted. An MIT failure
is a warning that something needs to be addressed, and a loss of mechanical integrity is an event
that can result in fluid movement from the well if remediation is not performed. More details on
MITs are available in Appendix D.
Monitoring the treatment pressure of the hydraulic fracturing operation can also detect problems
occurring during fracturing. Sudden changes in pressure during hydraulic fracturing operations can
be indicative of failures in the cement or casing. This type of monitoring is performed in nearly all
hydraulic fracturing jobs: the Well File Review fU.S. EPA. 2016cl found that the treatment pressure
was monitored in 97% (or 27,700) of all hydraulic fracturing jobs studied.1
6.2.1.3 Well Orientation
A well can be drilled and constructed with any of several different orientations: vertical, horizontal,
and deviated. The well's orientation can be important, because it affects the difficulty of drilling,
constructing, and cementing the well. In particular, as described in Section 6.2.2, constructing and
cementing horizontal wells present unique challenges (Sabins. 1990). In a vertical well, the
wellbore is vertical throughout its entire length, from the wellhead at the surface to the production
zone. Deviated wells are usually drilled vertically in the shallowest part of the well but are then
drilled directionally, deviating from the vertical direction at some point such that the bottom of the
well is at a significant lateral distance away from the point in the subsurface directly under the
wellhead. In a horizontal well, the well is drilled vertically to a point known as the kickoff point,
where the well turns toward the horizontal, extending into and parallel with the approximately
horizontal targeted producing formation (Figure 6-2).
Among wells evaluated in the Well File Review, about 65% were vertical, 11% were horizontal, and
24% were deviated wells (U.S. EPA. 2015n).2This is generally consistent with information available
in industry databases—of the approximately 16,000 oil and gas wells used in hydraulic fracturing
operations in 2009 (one of the years for which the data for the Well File Review were collected),
39% were vertical, 33% were horizontal, and 28% were either deviated or the orientation was
unknown (Drillinglnfo. 2014b). See Section 3.3 for additional information on the use of horizontal
wells in the United States.
6.2.1.4 Well Completion
Another important aspect of well construction is the way in which the well is completed into the
production zone, because the well's completion is part of the system of barriers and must be intact
for the well to operate properly. A variety of completion configurations are available. The most
common configuration is for casing to extend to the end of the wellbore and be cemented in place
(U.S. EPA. 2015n: George etal.. 2011: Renpu. 2011). In these cased and cemented completions, the
127,700 jobs (95% confidence interval: 24,800 - 30,600 jobs].
2 The Well File Review considered any non-horizontal well in which the well bottom was located more than 500 ft (152
m] laterally from the wellhead as being deviated.
6-15
-------
Chapter 6 - Well Injection
cement provides the primary containment of fluids to the production zone. Before hydraulic
fracturing begins, perforations are made through the casing and cement into the production zone. It
is through the perforated casing and cement that hydraulic fracturing is conducted. In some cases, a
smaller temporary casing, known as a temporary frac string, is inserted inside the production
casing to protect the casing from the high pressures imposed during hydraulic fracturing
operations.
A different type of a cased completion uses production casing set on formation packers, where the
production casing extends through the production zone and the length of the casing extending
through the drilled horizontal wellbore is leftuncemented, but has a series of formation packers
that swell to seal the annulus between the casing and the formation.1 With these completions, the
production zone is fractured in separate stages through ports that open between the formation
packers. When formation packers are used, they provide the primary isolation of hydraulic
fracturing fluids during hydraulic fracturing.
Another type of completion is an open hole completion. When open hole completions are used, the
entire production zone is fractured all at once in a single stage or may be fractured in separate
stages using a temporary frac string set on one or more temporary formation packers that are
positioned to a different interval for each stage. If a temporary frac string is used in an open hole
completion, its packer (s) provide the primary isolation of hydraulic fracturing fluids during
hydraulic fracturing and if no temporary frac string is used, then the next higher casing in the well
provides the primary isolation of hydraulic fracturing fluids during the treatment
Among wells represented in the Well File Review, an estimated 6% of wells (1,500 wells) had open
hole completions, 6% of wells (1,500 wells) used formation packers, and the rest were cased and
cemented (U.S. EPA. 2015n).2'3
In some cases, wells may be re-completed after the initial construction, with re-fracturing if
production has decreased (Vincent. 2011). Re-completion also may include additional perforations
in the well at a different interval to produce from a new formation, lengthening the wellbore, or
drilling new laterals from an existing wellbore. In 95% of the re-completions represented in the
Well File Review, hydraulic fracturing occurred at shallower depths than the previous job (U.S. EPA.
2016c).4
6.2.2 Factors that can Affect Fluid Movement to Drinking Water Resources
The following sections describe the pathways for fluid movement that can develop within the
production well and wellbore. We also describe the conditions leading to the development of fluid
movement pathways and, where available, evidence that a pathway has allowed fluid movement to
1A formation packer is a specialized casing part that has the same inner diameter as the casing but whose outer diameter
expands to make contact with the formation and seal the annulus between the uncemented casing and formation,
preventing migration of fluids.
21,500 wells with open hole completions (95% confidence interval: 10 - 4,800 wells].
31,500 wells using formation packers (95% confidence interval: 1,400 - 1,600 wells].
4 95%o of jobs (95%o confidence interval: 75 - 99% of jobs].
6-16
-------
Chapter 6 - Well Injection
occur within the casing or cement, and—in the case of sustained casing pressure (Section 6,2.2.4)—
a combination of factors within the casing and cement. (See Figure 6-4 for an illustration of
potential fluid movement pathways related to casing and cement.}
Figure 6-4. Potential pathways for fluid movement in a cemented wellbore.
These pathways (represented by the white arrows) include: (1) casing and tubing leak into a permeable formation,
(2) migration along an uncemented annulus, (3) migration along microannuli between the casing and cement, (4)
migration through poor cement, and (5) migration along microannuli between the cement and formation. Note:
the figure is not to scale and is intended to provide a conceptual illustration of pathways that may develop within
the well.
6-17
-------
Chapter 6 - Well Injection
We describe information regarding the rate at which these pathways have been identified in
hydraulic fracturing wells when it is available. Where such information does not exist, we present
the results of research on oil and gas production wells in general or on injection wells, including
those used for the geologic sequestration of carbon dioxide.1 Publicly accessible information is
insufficient to determine whether wells intended for hydraulic fracturing are constructed
differently from production wells where no fracturing is conducted. See Chapter 10 for additional
discussion of data gaps. It is not generally possible, based on the literature reviewed for this
assessment, to determine the precise degree to which hydraulic fracturing created, or moved fluids
along, the pathways described or whether all of the wells studied were hydraulically fractured. Nor
is it generally possible to estimate the degree to which wells that were hydraulically fractured have
a significantly different number of redundant barriers to protect drinking water resources than
other production wells. However, given the applicability of well construction technology to address
the subsurface conditions encountered in hydraulic fracturing operations and production or
injection operations in general, the information presented here is considered relevant to the
assessment.
6.2.2.1 Pathways Related to Well Casing
High pressures associated with hydraulic fracturing operations can damage casing and lead to fluid
movement that can impact drinking water quality. As noted above, the casing string through which
hydraulic fracturing fluids are injected is subject to higher internal pressures during hydraulic
fracturing operations than during other phases in the life of a production well. To withstand the
stresses created by the high pressure of hydraulic fracturing, the well and its components must
have adequate strength and elasticity. If the casing is compromised or is otherwise not strong
enough to withstand these stresses (Figure 6-3), a casing failure can result. If undetected or not
repaired, casing failures can serve as pathways for hydraulic fracturing fluids to leak out of the
casing. Below we present data or information suggesting that pathways along the casing are
present or allowing fluid movement. See Chapter 10 for more information on factors that can
increase or decrease the frequency or severity of impacts to drinking water quality associated with
well construction.
Hydraulic fracturing fluids or fluids present within the well casing could flow into other zones in
the subsurface if there is a leak in the casing, and cement is inadequate or not present As described
below, pathways for fluid movement associated with well casing can be related to the original
design or construction of the well, degradation of the casing over time, or problems that can arise
through extended use as the casing succumbs to stresses.
Casing failure can also occur if the wellbore passes through a structurally weak geologic zone that
shears and deforms the well casing. Such shearing is common when drilling through zones
containing salt fRenpu. 20111. The changes in the pressure field in the portions of the formation
near the wellbore during hydraulic fracturing can also cause mechanically weak formations to
shear, potentially damaging the well's casing or cement. Palmer et al. (2005) demonstrated through
modeling that hydraulic fracturing within coal that had a low unconfined compressive strength
1 An injection well is a well into which fluids are being injected (40 Code of Federal Regulations (CFR] 144.3].
6-18
-------
Chapter 6 - Well Injection
could cause shear failure of the coalbeds surrounding the wellbore. Shearing of the coalbed layers
can cause the casing to deform and potentially fail.
Corrosion in uncemented zones is the most common cause of casing failure. This can occur if
uncemented sections of the casing are exposed to corrosive substances such as brine or hydrogen
sulfide fRenpu. 20111. Corrosion commonly occurs at the collars that connect sections of casing or
where equipment is attached to the casing. Corrosion at collars can exacerbate problems with loose
or poorly designed connections, which are another common cause of casing leaks fAgbalagba et al..
2013: Brufatto etal.. 2003). Watson and Bachu (2009) found that 66% of all casing corrosion
occurred in uncemented well sections, as shown in Pathway 1 of Figure 6-4.
As noted above, the casing and cement work together to strengthen the well and provide zonal
isolation. Uncemented casing does not necessarily lead to fluid migration. However, migration can
occur if the casing in an uncemented zone fails during hydraulic fracturing operations.
Other mechanical integrity problems have been found to vary with the well environment,
particularly environments with high pressures and temperatures. Wells in high pressure/high
temperature environments, wells with thermal cycling, and wells in corrosive environments can
have life expectancies of less than 10 years fAgbalagba etal.. 20131.
The depth of the surface casing relative to the base of the drinking water resource to be protected is
an important factor in protecting the drinking water resource. In a limited risk modeling study of
selected injection wells in the Williston Basin, Michie and Koch (1991) found the risk of aquifer
contamination from leaks from the inside of the well to the drinking water resource was seven in
1,000,000 injection wells if the surface casing was set deep enough to cover the drinking water
resource, and that the risk increased to six in 1,000 wells if the surface casing was not set deeper
than the bottom of the drinking water resource. An example where surface casing did not extend
below drinking water resources comes from an investigation of 14 selected drinking water wells
with alleged water quality problems in the Wind River and Fort Union formations near Pavillion,
Wyoming (WYOGCC, 2014b). The state found that the surface casing of oil and gas wells was
shallower than the depth of three of the 14 drinking water wells. Some of the oil and gas wells with
shallow surface casing had elevated gas pressures in their annuli ("WYOGCC. 2014bl. The presence
of gas in the annuli, combined with surface casing that is set above the lowest drinking water
resource, could allow migration of gas into drinking water resources.
Fleckenstein et al. (2015) found that the depth of surface casing and the presence of uncemented
gas zones are major factors in determining the likelihood of well failures and contamination. Their
study in the Wattenberg field in Colorado classified the wells in the field into seven categories
based on the depth of surface casing, the presence of cement, and the presence of intermediate gas
zones above the production zone (Table 6-1). The categories were arranged in order of risk, with
category 1 wells being at the highest risk of allowing fluid migration and category 7 wells being the
least likely to allow migration. The overall barrier failure rate was 2.4% of all wells, and the overall
catastrophic failure rate was 0.06% of all wells. A remediation effort was made in order to decrease
the likelihood of fluid migration, which included the plugging of 1,103 of the 17,948 wells studied.
All the wells shown in the table are vertical wells that were drilled between 1970 and 2013. Similar
6-19
-------
Chapter 6 - Well Injection
categories were created for the 973 horizontal wells in the field. No failures were recorded for any
of the horizontal wells.
Table 6-1. Failure rates of vertical wells in the Wattenberg field, Colorado.
From Fleckenstein et al. (2015).
Category and description3'15
Total wells
Wells with
barrier
failures0
Wells with
catastrophic
failures'1
l-Shallow surface casing and exposed (uncemented) over-
pressured intermediate gas zones
399
92 (23.06%)
3 (0.75%)
2 - Shallow surface casing and exposed under-pressured
intermediate gas zones
7,811
276 (3.53%)
6 (0.08%)
3 - Shallow surface casing but no exposed gas zones
3,407
20 (0.59%)
1 (0.03%)
4 - Shallow surface casing with production casing cemented
to bottom of surface casing
1,063
0 (0%)
0 (0%)
5 - Deep surface casing with production casing cement below
top of gas
1,374
13 (0.95%)
0 (0%)
6 - Deep surface casing with production casing cement above
top of gas
2,069
0 (0%)
0 (0%)
7 - Deep surface casing with production casing cement to
bottom of surface casing
705
0 (0%)
0 (0%)
Total
16,828
401 (2.4%)
10 (0.06%)
a The study defined shallow surface casing as casing that did not extend below the Fox Hills Aquifer, a deep aquifer that had not
been identified and protected by the state prior to 1994.
b Uncemented zones could be located along the intermediate or production casings.
c Barrier failures were considered to have occurred when there were signs of a failure, but no contamination.
d A catastrophic failure was considered to have occurred where there was contamination of drinking water aquifers (i.e., the
presence of thermogenic gas in a drinking water well) and evidence of a well defect such as exposed intermediate gas zone or
casing leaks.
Sherwood etal. (2016) examined complaint records in the same field. They reviewed 29 Colorado
Oil and Gas Commission complaint records associated with 32 incidents at 42 drinking water wells
in which thermogenic methane was detected. (See Text Box 6-3 for more information on
thermogenic and biogenic methane.) Of the 29 complaints, 10 were determined to be caused by oil
and gas wellbore failures, one was suspected of being a wellbore failure but not confirmed, three
were settled in court with documents being sealed, and the remaining 15 were unresolved.1 If all 32
cases are assumed to be associated with an individual oil and gas well, that would result in a failure
rate of 0.06% of all oil and gas wells in the basin, the same failure rate as found in the Fleckenstein
etal. f20151 study. As in the Fleckenstein study, surface casing that was set too shallow and
1 This paper defined a wellbore failure as the failure of one or more barriers to fluid movement in the wellbore (e.g.,
cement, casing, etc.].
6-20
-------
Chapter 6 - Well Injection
uncemented intermediate zones were the main contributing factors to wellbore failure. All 11 of the
confirmed or suspected wellbore failures involved vertical wells that were drilled before 1933 and
had surface casing shallower than nearby aquifers. Of these wells, seven had been hydraulically
fractured. The study noted that the failure rate was fairly constant over time with about two new
cases per year since 2000 and that the rate had not changed since high rates of hydraulic fracturing
of horizontal wells became prevalent around 2010. This is consistent with the study's finding of no
failures in horizontal wells.
During hydraulic fracturing operations in September of 2010 near Killdeer, in Dunn County, North
Dakota, the production, surface, and conductor casing of the Franchuk 44-20 SWH well ruptured,
causing fluids to spill to the surface flacob. 20111. The rupture occurred during the 5th of
23 planned stages of hydraulic fracturing when the pressure spiked to over 8,390 psi (58 MPa).
Ruptures were found in two locations along the production casing—one just below the surface and
one at about 55 ft (17 m) below ground surface. The surface casing ruptured in three places down
to a depth of 188 ft (57 m), and the conductor casing ruptured in one place. Despite a shutdown of
the pumps, the pressure was still sufficient to cause fluid to travel through the ruptured casings and
to flow to the surface. Ultimately, over 166,000 gal (628,000 L) of fluids and approximately
2,860 tons (2,595 metric tons) of contaminated soil were removed from the site (Tacob. 20111.
The EPA investigated the Killdeer site as part of its Retrospective Case Study in Killdeer, North
Dakota: Study of the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources fU.S.
EPA. 2015il. As part of the study, water quality samples were collected from three domestic wells,
nine monitoring wells, two supply wells, one municipal well, and one state well in July 2011,
October 2011, and October 2012. Two study wells installed less than 60 ft (20 m) from the
production well (NDGW08 and NDGW07) had significant differences in water quality compared to
the remaining study wells.1 These two wells showed differences in ion concentrations (e.g.,
chloride, calcium, magnesium, sodium, strontium) and tert-butyl alcohol (TBA). The sampling
identified brine contamination that was consistent with mixing of local groundwater with brine
from Madison Group formations, which the well had penetrated. The TBA was consistent with
degradation of tert-butyl hydroperoxide, a component of the hydraulic fracturing fluid used in the
Franchuk well. Based on the analysis of potential sources of contamination, the EPA determined
that the only potential sources of TBA were gasoline spills, leaky underground storage tanks, and
hydraulic fracturing fluids. However, the lack of MTBE and other signature compounds associated
with gasoline or fuels strongly suggests that the rupture (blowout) was the only source consistent
with findings of high brine and TBA concentrations in the two wells.2 For additional information
about impacts at the Killdeer site, see Section 6.3.2.2.
1 Based on comparison with historical Killdeer aquifer water quality data, the remaining study wells were in general
consistent with historical background data; these wells were then used for the data analysis as background wells.
Comparisons of TBA between the study data and historical data could not be made since no historical data for TBA were
found for the Killdeer aquifer.
2 A well blowout is the uncontrolled flow of fluids out of a well.
6-21
-------
Chapter 6 - Well Injection
Inadequate casing or cement can respond poorly when blowout preventers activate.1 When
blowout preventers are activated, they immediately stop the flow in the well, which can create a
sudden pressure increase in the well. If the casing or cement are not strong enough to withstand
the increased pressure when this occurs, well components can be damaged fThe Royal Society and
the Royal Academy of Engineering. 2012) and the potential for fluid release and migration in the
subsurface increases. Blowouts can also occur during the production phase, and cause spills on the
surface that can affect drinking water resources; see Section 7.4.2.2.
While well construction and hydraulic fracturing techniques continue to change, the pressure- and
temperature-related stresses associated with hydraulic fracturing remain as factors that can affect
the integrity of the well casing. Tian etal. f20151 investigated one such case where temperature
effects led to casing damage in China. In the Changning-Weiyuan basin in China, 13 of 33 wells
(39.4%) suffered casing damage, with most of the wells experiencing the damage after fracturing.
The authors found that injection of the cooler hydraulic fracturing fluid led the casing temperature
to drop from the formation temperature of 212°F to 64°F (100°C to 18°C) in some cases. This drop
in temperature, in turn, caused pockets of high pressure fluid outside the casing to contract If the
temperature dropped below 136°F (58°C), the effect was sufficient to form a vacuum outside the
casing, potentially leading to casing deformation. Areas of the casing with severe doglegs (i.e., bends
in the well) and where there was a smaller space between the casing and formation were more
prone to this type of damage. While the conditions in this Chinese basin may or may not represent
conditions in U.S. basins, they do demonstrate that temperature changes during hydraulic
fracturing can place additional stress on the well and highlight their importance as a consideration
in casing design. In the case mentioned, increasing the space around the casing, decreasing dogleg
angles, properly removing drilling mud, and using high strength, low elasticity cement were found
to improve performance.
Sugden etal. f20131 used numerical simulation to examine a similar problem using parameters
chosen to represent the Haynesville Shale. They found that injecting a fluid at 70°F (21°C) could
cool the wellbore temperature from 320°F to 96°F (160°C to 36°C). The temperature change was
90% complete within the first half hour of hydraulic fracturing operations. They also found that a
well with a 20 degree per 100 ft (31 m) dogleg decreased the pressure required to damage the well
casing by 850 psi (5.9 MPa). The study also reported that cooling of fluids in voids in the cement can
lead to contraction of the fluids. In low permeability shales, fluid cannot flow in fast enough to
compensate, and the pressure in the void can drop significantly. Sugden etal. f20131 report that
such cement voids can reduce the pressure needed to rupture the casing by 40%.
Emerging isotopic techniques can be used to identify the extent to which stray gas occurring in
drinking water resources is linked to casing failure. (See Text Box 6-3 for more information on stray
gas.) Darrah etal. f20141 used hydrocarbon and noble gas isotope data to investigate the source of
gas in eight identified "contamination clusters" that occurred in the Marcellus and Barnett shales.
Seven of these clusters were stripped of atmospheric gases (Argon-36 and Neon-20) and were
1A blowout preventer (BOP] is casinghead equipment that prevents the uncontrolled flow of oil, gas, and mud from the
well by closing around the drill pipe or sealing the hole f Oil and Gas Mineral Services. 20101 BOPs are typically a
temporary component of the well, in place only during drilling and perhaps through hydraulic fracturing operations.
6-22
-------
Chapter 6 - Well Injection
enriched in crustal gases, indicating the gas migrated quickly from depth without equilibrating with
intervening formations. The rapid transport was interpreted to mean that the migration did not
occur along natural fractures or pathways, which would have allowed equilibration to take place.
Based on the isotopic results, the authors also ruled out the possibility that the gas was carried
upward (relative to the surface) as the geologic formation in which it formed was uplifted over
geologic time. Possible explanations for the rapid migration include transport up the well and
through a leaky casing (Pathway 1 in Figure 6-4) or along uncemented or poorly cemented
intervals from shallower depths (Pathways 2 through 5 in Figure 6-4). In four Marcellus Shale
clusters, gas found in drinking water wells had isotopic signatures and ratios of ethane to methane
that were consistent with those in the producing formation. The authors conclude that this suggests
that gas migrated along poorly constructed wells from the producing formation, likely with
improper, faulty, or failing production casings. In three clusters, the isotopic signatures and ethane
to methane ratios were consistent with formations overlying the Marcellus. The authors suggest
that this migration occurred from the shallower gas formations along uncemented or improperly
cemented wellbores. In another Marcellus cluster in the study, deep gas migration was linked to a
subsurface well, likely from a failed well packer.
Text Box 6-3. Stray Gas Migration.
Stray gas refers to the phenomenon of natural gas (primarily methane) migrating into shallow drinking water
resources, into water wells or other types of wells, to the surface, or to near-surface features (e.g., basements,
streams, or springs). The source of the migrating gas can be natural gas reservoirs (either conventional or
unconventional), or from coal mines, landfills, leaking gas wells, leaking gas pipelines, buried organic matter,
or natural microbial processes (Li and Carlson. 2014: Baldassare. 2011). Although methane is not a regulated
drinking water contaminant, its presence in drinking water resources can initiate chemical and biological
reactions that release or mobilize other contaminants. Over time, it can promote more reducing conditions in
groundwater, potentially leading to reductive dissolution of iron and manganese and the possible liberation
of naturally occurring contaminants, such as arsenic, that are potentially associated with iron and manganese
(U.S. EPA. 2014f). In addition, methane can accumulate to explosive levels in confined spaces (like basements
or cellars) if it exsolves (degases) from groundwater into those spaces. (See Section 9.5.5 for information
about the hazards associated with methane exposure.)
Detectable levels of dissolved natural gas exist in some aquifers, even in the absence of human activity
fGorodv. 20121 In northern Pennsylvania and New York, low levels of methane are frequently found in water
wells in baseline studies, prior to commercial oil or gas development (Christian et al.. 2016: Kappel. 2013:
Kappel and Nvstrom. 2012): for example, one USGS study detected methane in 80% of sampled wells in Pike
County, Pennsylvania (Senior. 2014). The origin of methane in groundwater can be either thermogenic
(produced by high temperatures and pressures in deeper formations, such as the gas found in the Marcellus
Shale) or biogenic (produced in shallower formations by bacterial activity in anaerobic conditions).
Gas occurrence is linked to local and regional geologic characteristics. In some cases, thermogenic methane
occurs naturally in shallow formations because the formation itself was uplifted (relative to the surface) over
geologic time. In other cases, it has migrated there via one or more pathways. For example, Brantley et al.
(2014) suggest that northern Pennsylvania's glacial history can help explain why stray gas is more common
(Text Box 6-3 is continued on the following page.)
6-23
-------
Chapter 6 - Well Injection
Text Box 6-3 (continued). Stray Gas Migration.
there than in the southern part of the state. Christian et al. f20161. Mcphillips etal. f20141. Molofskv et al. fin
Press], and Wilson (20141 all identified correlations between the presence of methane in water wells and
certain geologic, hydrographic, and geochemical parameters, such as valley locations and the presence of coal
beds.
Stray gas migration can be a technically complex phenomenon to study, in part because there are many
potential sources and routes for migration. When a particular site lacks detailed monitoring data, especially
baseline measurements, determination of sources and migration routes is complicated and challenging.
Examining the concentrations and isotopic compositions of methane and higher molecular weight
hydrocarbons such as ethane and propane can aid in determining the source of stray gas (Tillev and
Muehlenbachs. 2012: Baldassare. 2011: Rowe and Muehlenbachs. 19991 Isotopic composition and
methane-to-ethane ratios can help determine whether the gas is thermogenic or biogenic in origin and
whether it is derived from shale or other formations fGorodv. 2012: Muehlenbachs etal.. 2012: Barker and
Fritz. 19811 Isotopic analysis can also be used to identify the strata where the gas originated and provide
evidence for migration mechanisms (Darrah et al.. 20141 For example, isotope-based techniques have been
used to investigate the potential sources of methane in drinking water wells in Dimock, Pennsylvania
(Hammond. 20161 and lackson et al. (2013c) found evidence of potential Marcellus gas contamination in
some Pennsylvania drinking water wells using stable-isotopic ratios, while other wells in the area appeared
to be contaminated by shallower sources (not associated with gas production).
However, determining the source of methane does not necessarily establish the migration pathway. Multiple
researchers (e.g., Siegel etal.. 2015: lackson et al.. 2013c: Molofskv et al.. 2013: Revesz et al.. 2012: Osborn et
al.. 20111 have described biogenic and/or thermogenic methane in groundwater supplies in Marcellus gas
production areas, although the sources and pathways of migration are generally unknown. Well casing and
cementing issues may be an important source of stray gas problems (lackson et al.. 2013cl however, other
potential subsurface pathways are also discussed in the literature. Zhang and Soeder (20161 suggested that
air-drilling practices used to construct the vertical component of gas wells can affect methane migration by
creating groundwater surges in the shallow subsurface. The type of well may also play a role; in one study,
deviated gas wells in Canada were three to four times more likely than vertical wells to have evidence of gas
migration to the surface (lackson et al.. 2013bl
In the absence of data on specific pathways, some researchers have investigated geographic correlations.
lackson et al. f2013c1 and Osborn et al. f20111 found that thermogenic methane concentrations in well water
increased with proximity to Marcellus Shale production sites. In contrast, Molofskv et al. (20131 found the
presence of gas to be more closely correlated with topography and elevation, and (Siegel et al.. 20151found no
correlation between methane in groundwater and proximity to production wells. Kresse et al. T20121
investigated methane concentration and isotopic geochemistry in shallow groundwater in the Fayetteville
Shale area, and found no evidence that the water had been influenced by shale gas activities. Similarly, Li and
Carlson (20141. while not ruling out potential leakage pathways from deeper reservoirs, found no systematic
correlation between increasing well drilling density in the Wattenberg Field in Colorado and near-surface
stray gas concentrations.
EPA conducted retrospective case studies to investigate stray gas in northeastern Pennsylvania and the Raton
Basin of Colorado. As described in the northeastern Pennsylvania case study report, Retrospective Case Study
in Northeastern Pennsylvania: Study of the Potential Impacts of Hydraulic Fracturing on Drinking Water
Resources (U.S. EPA. 2014fl 27 of 36 drinking water wells within the study area (75%) contained elevated
methane concentrations. For some of the wells, the EPA concluded that the methane (of both thermogenic
and biogenic origin) was naturally occurring gas, not attributable to gas exploration activities. In others, it
(Text Box 6-3 is continued on the following page.)
6-24
-------
Chapter 6 - Well Injection
Text Box 6-3 (continued). Stray Gas Migration.
appeared that methane had entered the water wells following well drilling and hydraulic fracturing. In most
cases, the methane in the wells likely originated from intermediate formations between the production zone
and the surface; however, in some cases, the methane appears to have originated from deeper layers such as
those where the Marcellus Shale is found (U.S. EPA. 2014f). The Raton Basin case study examined the Little
Creek Field, where potentially explosive quantities of methane entered drinking water wells in 2007. As
described in the EPA's Retrospective Case Study in the Raton Basin, Colorado: Study of the Potential Impacts of
Hydraulic Fracturing on Drinking Water Resources (U.S. EPA. 2015k). the methane was found to be primarily
thermogenic in origin, modified by biologic oxidation (U.S. EPA. 2015k). Secondary biogeochemical changes
related to the migration and reaction of methane within the shallow drinking water aquifer were reflected in
the characteristics of the Little Creek Field groundwater (U.S. EPA. 2015k).
The sources of methane in the two studies could be determined with varying degrees of certainty. Narrowly
identifying the most likely pathway(s) of migration has been more difficult. In northeastern Pennsylvania,
while the sources could not be definitively determined, the Marcellus Shale could not be excluded as a
potential source in some wells based on isotopic signatures, methane-to-ethane ratios, and isotope reversal
properties (U.S. EPA. 2014f). The Pennsylvania Department of Environmental Protection (PA DEP] cited at
least two operators for failure to prevent gas migration at wells within the study area. Evidence cited by the
state included isotopic comparison of gas samples from drinking water wells, water bodies, and gas wells;
inadequate cement jobs; and sustained casing pressure (although, under Pennsylvania law, oil or gas
operators can be cited if they cannot disprove the contamination was caused by their well using pre-drilling
samples) (Llewellyn et al.. 20151A separate study (Ingraffea et al.. 20141 showed that wells in this area had
higher incidences of mechanical integrity problems relative to wells in other parts of Pennsylvania. While the
study did not definitively show that stray gas was linked to construction problems, it does imply that there
may be more difficulties in constructing wells in this area. In the Little Creek Field in the Raton Basin, the
source of methane was identified as the Vermejo coalbeds. While the nature of the migration pathway is
unknown, modeling suggests that it could have occurred along natural rock features in the area and/or along
a gas production well fU.S. EPA. 2015kl Because the production wells were shut in shortly after the incident
began, the wells could not be inspected to determine whether a mechanical integrity failure in the wellbore
was a likely cause of the migration.1
These two case studies illustrate the considerations involved with understanding stray gas migration and the
difficulty in determining sources and migration pathways. To more conclusively determine sources and
migration pathways, studies in which data are collected on mechanical integrity and hydrocarbon gas (e.g.,
methane, ethane) concentrations both before and after hydraulic fracturing operations, in addition to the
types of data summarized above, would be needed.
In the Wattenberg Field in Colorado, Li etal. (2016a) investigated the concentration of various ions
in water from an uncontaminated aquifer, an aquifer containing thermogenic methane, and
produced water from oil and gas wells to understand the transport of aqueous- and gas-phase
fluids at the site. The results indicated that the methane that was contaminating water wells was
not transported with aqueous phase fluids; the authors suggested that this can provide evidence for
migration mechanisms, because certain pathways (e.g., migration from improperly sealed well
1 Shutting in a well refers to sealing off a well by either closing the valves at the wellhead, a downhole safety valve, or a
blowout preventer.
6-25
-------
Chapter 6 - Well Injection
casings) could potentially result in gas-phase but not aqueous-phase migration. See Text Box 6-4 for
another example of an investigation into the occurrence of stray gas in drinking water wells.
Text Box 6-4. Parker County, Texas.
Peer-reviewed studies have been conducted within the Barnett Shale area, which includes Parker County,
Texas. These include sampling studies of private water well composition, noble gas content, and isotopic
signatures of natural gases, as well as analysis of existing water sample data. Disagreement exists about the
origin of the increased natural gas in private well water.
One suggested possibility is that production casing annuli could serve as a migration pathway for natural gas
from formations located between the Barnett and the Trinity to reach overlying intervals (including the
Trinity aquifer) fDarrah et al„ 20141 However, using measurements of hydrocarbon and noble gas isotopes,
Wen etal. (2016) suggests the source of methane in the Trinity aquifer water wells is directly from the
underlying Strawn Formation and not from pathways associated with the gas production wells although the
timing of methane entry into the Strawn is not known.
6.2.2.2 Pathways Related to Cement
Fluid movement can result from inadequate well design or construction (e.g., cement loss or other
problems that arise in cementing of wells) or degradation of the cement over time (e.g., corrosion
or the formation of microannuli), which may, if undetected and not repaired, cause the cement to
succumb to the stresses exerted during hydraulic fracturing.1 The well cement must be able to
withstand the subsurface conditions and the stresses encountered during hydraulic fracturing
operations. This section presents data and information that can help indicate that pathways within
the cement are present or allowing fluid movement.
Uncemented zones can allow fluids or brines to move into drinking water resources. If a fluid-
containing zone is left uncemented, the open annulus between the formation and casing can act as a
pathway for migration of that fluid. Fluids can enter the wellbore along any uncemented section of
the wellbore if a sufficient pressure gradient is present. Once the fluids have entered the wellbore,
they can travel up along the entire uncemented length of the wellbore as shown in Pathway 2 of
Figure 6-4.
As mentioned in Section 6.2.2.1, Fleckenstein etal. (2015) found uncemented gas zones to be a
significant factor in barrier failures in wells in the Wattenberg basin in Colorado. A report on the
Pavillion field by AME (2016) identified a similar set of risk factors for fluid migration including:
uncemented production casing, shallow surface casing, and the presence of both an intermediate
pressurized gas zone and a permeable groundwater zone encountered in the same production
wellbore.
Because of their low density and buoyancy, gaseous fluids such as methane will migrate up the
wellbore if an uncemented wellbore is exposed to a gas-containing formation. Gas may then be able
1 Microannuli are very small openings that form between the cement and its surroundings and that may serve as
pathways for fluid migration to drinking water resources.
6-26
-------
Chapter 6 - Well Injection
to enter other formations (including drinking water resources) if the wellbore is uncemented and
the pressure in the annulus is sufficient to force fluid into the surrounding formation fWatson and
Bachu. 2009: Harrison. 19851. The rate at which the gas can move will depend on the difference in
pressure between the annulus and the formation fWoitanowicz. 20081. See Chapter 10 for a
discussion of practices, such as well testing, that can decrease the frequency of such gas migration
that could impact drinking water quality.
In several cases, poor or failed cement has been linked to stray gas migration (Text Box 6-3). A
Canadian study found that uncemented portions of casing were the most significant contributors to
gas migration (Watson and Bachu. 20091. The same study also found that 57% of all casing leaks
occurred in uncemented segments. In the study by Darrah etal. f20141 fSection 6.2.2.1), using
isotopic data, four clusters of gas contamination were linked to poor cementing. In three clusters in
the Marcellus and one in the Barnett, gas found in drinking water wells had isotopic signatures
consistent with intermediate formations overlying the producing zone. This suggests that gas
migrated from the intermediate units along the well annulus, along uncemented portions of the
wellbore, or through channels or microannuli.
Cementing of the surface casing is the primary aspect of well construction intended to protect
drinking water resources. Most states require the surface casing to be set and cemented from the
level of the lowermost drinking water resource to the surface (GWPC. 20141. Most wells—including
those used in hydraulic fracturing operations—have such cementing in place. Among the wells
represented in the Well File Review, surface casing was found to be fully cemented in 93% of wells.
Of these, an estimated 55% of wells (12,600 wells) were cemented to below the operator-reported
protected groundwater resource; in an additional 28% of wells (6,400 wells), the operator-
reported protected groundwater resources were fully covered by the next cemented casing
string.12 3 A portion of the annular space between the casing and the operator-reported protected
groundwater resources was uncemented in at least 3% of wells (600 wells) (U.S. EPA. 2015nl.4
Improper placement of cement can lead to defects in external mechanical integrity. For example, an
improper cement job can be the result of loss of cement during placement into a formation with
1 In the Well File Review, protected groundwater resources were as reported by well operators. For most wells
represented in the Well File Review, protected groundwater resources were identified based on state or federal
authorization documents. Other data sources used by well operators included aquifer maps, data from offset production
wells, open hole log interpretations by operators, operator experience, online databases, and references to a general
requirement by the oil and gas agency.
2 The research that the EPA reviewed used various terms to describe subsurface water resources that are
used/potentially used for drinking water. Where another term is relevant to describing the author's research, we use that
term; for the purpose of this assessment, all of these terms are considered to fall within the assessment's definition of
"drinking water resources." See Chapter 2 for additional information on the definition of a drinking water resource.
3 6,400 wells (95% confidence interval: 500 - 12,300 wells].
4 600 wells (95% confidence interval: 10 - 1,800 wells]. The well files representing an estimated 8% of wells in the Well
File Review did not have sufficient data to determine whether the operator-reported protected groundwater resource
was uncemented or cemented. In these cases, there was ambiguity either in the depth of the base or the top of the
operator-reported protected groundwater resource. An additional 6% of wells represented had surface casing set below
the reported protected groundwater resource depth, but because the protected groundwater depth was based on a
nearby water well depth, the true base of the protected groundwater resource may be deeper, leaving uncertainty as to
whether the surface casing in these wells is set deeper than the base of the protected groundwater resource.
6-27
-------
Chapter 6 - Well Injection
high porosity or fractures, causing a lack of adequate cement across a water- or brine-bearing zone.
Additionally, failure to use cement that is compatible with the anticipated subsurface conditions,
failure to remove drilling fluids from the wellbore, and improper centralization of the casing in the
wellbore can all lead to the formation of channels (i.e., small connected voids) in the cement during
the cementing process (McDaniel et al.. 2014: Sabins. 1990). If the channels are small and isolated,
they may not lead to fluid migration. However, if they are long and connected, extending across
multiple formations, or connecting to other existing channels or fractures, they can present a
pathway for fluid migration. Figure 6-4 shows a variety of pathways for fluid migration that are
possible from failed cement jobs.
One example of how hydraulic fracturing of a well with insufficient and improperly placed cement
led to contamination occurred in Bainbridge Township, Ohio. This incident was well studied by the
Ohio Department of Natural Resources fODNR. 20081 and by an expert panel fBair etal.. 20101. The
level of detail available for this case is not typically found in studies of such events but was collected
because of the severity of the impacts and the resulting legal action. The English #1 well was drilled
to a depth of 3,900 ft (1,200 m) below ground surface (bgs) in October 2007 with the producing
formation located between 3,600 and 3,900 ft (1,100 and 1,200 m) bgs. Overlying the producing
formation were several uneconomic formations containing over-pressured gas (i.e., gas at
pressures higher than the hydrostatic pressure exerted by the fluids within the well).1 The original
cement design required the cement to be placed 700 - 800 ft (210 - 240 m) above the producing
formation to seal off these areas. During cementing, however, both the spacer fluid and cement
were lost in the subsurface, and the cement did not reach the intended height2 Despite the lack of
sufficient cement, the operator proceeded with hydraulic fracturing.
During the hydraulic fracturing operation in November 2007, about 840 gal (3,200 L) of fluid
flowed up the annulus and out of the well. When the fluid began flowing out of the annulus, the
operator immediately ceased operations and shut in the well; this caused the pressure in the
wellbore to increase. About a month later, there was an explosion in a nearby house where methane
had entered from an abandoned and unplugged drinking water well connected to the cellar (Bair et
al.. 2010). In addition to the explosion, the over-pressured gas entering the aquifer resulted in the
contamination of 26 private drinking water wells with methane. The wells, some of which had
histories of elevated methane prior to the incident, were taken off-line. By 2010, all of the well
owners had been connected to a public water supply fTomastik and Bair. 20101.
Contamination at the Bainbridge Township site was the result of inadequate cement. The ODNR
determined that failure to cement the over-pressured gas formations, proceeding with the
hydraulic fracturing operation without adequate cement, and the extended period during which the
well was shut in all contributed to the contamination of the aquifer with stray gas (ODNR. 2008).
Cement logs found the cement top was at 3,640 ft (1,110 m) bgs, leaving the uneconomic gas-
producing formations and a portion of the production zone uncemented. The surface casing was
253 ft (77 m) deep and cemented to the surface. Hydraulic fracturing fluids flowing out of the
1 Hydrostatic pressure is the pressure exerted by a column of fluid at a given depth. Here, it refers to the pressure exerted
by a column of drilling mud or cement on the formation at a particular depth.
2 Spacer fluid is a fluid pumped before the cement to clean drilling mud out of the wellbore.
6-28
-------
Chapter 6 - Well Injection
annulus provided an indication that hydraulic fracturing had created a path from the producing
formation to the well annulus in addition to the uncemented gas zones. Because the well was shut
in, the pressure in the annulus could not be relieved, and the gas eventually traveled through
natural fractures surrounding the wellbore into local drinking water aquifers (during the time the
well was shut in, natural gas seeped into the well annulus and pressure built up from an initial
pressure of 90 psi (0.6 MPa) to 360 psi (2.5 MPa)). From the aquifer, the gas moved into drinking
water wells and from one of those wells into a cellar, resulting in the explosive accumulation of gas.
The Well File Review found that 3% of all hydraulic fracturing jobs (800 jobs) reported a
mechanical integrity failure that allowed fluid to enter an annular space (U.S. EPA. 2016c).1 The
mechanical integrity failures generally resulted in hydraulic fracturing fluid entering the annular
space between the casing and formation or between two casings, and were generally noted by
increases in annular pressure or fluid bubbling to the surface. Other possible mechanisms for the
failures include casing leaks, cement failure, and fractures extending above the height of the
cement (See Section 6.3.2.2 for additional information on fracture overgrowth.) While failures
were noted, these do not necessarily indicate there was movement of fluid into a drinking water
resource. In most cases, when problems occurred, the hydraulic fracturing operation was stopped
and operators addressed the cause of the failure before hydraulic fracturing operations resumed;
however, in 0.5% of the hydraulic fracturing jobs (100 jobs) with identified failures, there was no
additional barrier between the annular space with fluid and protected drinking water resources.2
While it could not definitively be determined whether fluid movement into the protected drinking
water resource occurred, in these cases, all of the protective barriers intended to prevent such fluid
migration failed, leaving the groundwater resource vulnerable to contamination.
While limited literature is available on construction (including cementing) flaws in hydraulically
fractured wells, several studies have examined construction flaws in oil and gas wells in general.
One study that examined reported drinking water contamination incidents in Texas identified 10
incidents related to drilling and construction activities among 250,000 oil and gas wells (Kell.
2011). The study noted that many of the contamination incidents were associated with wells that
were constructed before Texas revised its regulations on cementing in 1969 (it is not clear how old
the wells were at the time the contamination occurred). Because this study relied on reported
incidents, it is possible that other wells exhibited mechanical integrity issues but did not result in
contamination of a drinking water well or were not reported. Therefore, this should be considered
a low-end estimate of the number of mechanical integrity issues that could be tied directly to
drilling and construction activities. It is important to note that the 10 contamination incidents
identified were not associated with wells that were hydraulically fractured (Kell. 2011).
Several investigators have studied violations information from the PA DEP online violation
database to evaluate the rates of and possible factors contributing to mechanical integrity
problems, including those related to cement The results of these studies are summarized in Table
6-2.
1800 jobs (95% confidence interval: 10 - 1,700 jobs].
2100 jobs (95% confidence interval: 10 - 300 jobs].
6-29
-------
Chapter 6 - Well Injection
Table 6-2. Results of studies of PA DEP violation data that examined mechanical integrity
failure rates.
Study
Violations
investigated
Wells
studied
Data
timeframe
Key findings3
Considine et al.
(2012)
Violations resulting
in environmental
damage
3,533
2008-2011
Of 845 environmental damage incidents
(which resulted in 1,144 violations),
approximately 10% were related to casing
or cement problems. The overall violation
rate dropped from 52.9% of all wells in
2008 to 20.8% of all wells in 2011.
Davies et al.
(2014)
Failure of one of the
barriers preventing
fluid migration
8,030
2005-2013
Approximately 5% of wells received this
type of violation. The incident rate
increased to 6.3% when failures noted on
forms, but not resulting in violations, were
included.
Ingraffea et al.
(2014)
Violations and
inspection records
indicating structural
integrity loss
3,391
2000-2012
Wells in unconventional reservoirs
experienced a rate of structural integrity
loss of 6.2%, while the rate for
conventional wells was 1%.
Vidic et al.
(2013)
Construction
violations related to
casing or cement
6,466
2008-2013
Approximately 3.4% of wells received this
type of violation.
Olawovin et al.
(2013)
All violations
2,001
2008-2010
Analysis of 2,601 violations from 65
operators based on weighted risks found
that potentially risky violations increased
342% over the study period, while total
violations increased 110%.
Brantlev et al.
(2014)
Violations related
to well construction
issues
7,234
2005 - 2013
Over the period studied, a total of 3.4% of
well operators received violations for
construction issues. Violations in any given
year ranged from 0.6% to 10.8%. Also,
0.24% of wells were cited for methane
migration.
a While all of these studies used the same database, their results vary because they studied different timeframes and used
different definitions of what violations constituted a mechanical integrity problem or failure.
Because a significant portion of Pennsylvania's recent oil and gas activity is in the Marcellus Shale,
many of the wells in these studies were most likely used for hydraulic fracturing. For example,
Ingraffea et al. (2014) found that approximately 16% of the oil and gas wells drilled in the state
between 2000 and 2012 were completed in unconventional reservoirs, and nearly all of these wells
were used for hydraulic fracturing. Wells drilled in unconventional reservoirs experienced higher
rates of structural integrity loss, as defined by the authors, than conventional wells drilled during
the same time period (Ingraffea etal.. 2014). The authors did not compare rates of structural
6-30
-------
Chapter 6 - Well Injection
integrity loss in conventional wells that were and were not hydraulically fractured; they assumed
that unconventional wells were hydraulically fractured and conventional wells were not.
Violation rates resulting in environmental damage among all Pennsylvania wells dropped from
52.9% in 2008 to 20.8% in 2011 (Considine etal.. 20121. and the drop may be due to a number of
factors. Violations related to failure of cement or other well components represented a minority of
all well violations (i.e., among wells that were and were not hydraulically fractured). Of 845 events
that caused environmental damage, including but not limited to contamination of drinking water
resources, Considine etal. (2012) found that about 10% (85 events) were related to casing and
cement problems. The rest of the incidents were related to site restoration and spills; the violations
noted are confined to those incidents that caused environmental damage (i.e., the analysis excluded
construction flaws that did not have adverse environmental effects). In addition, two wells (0.06%)
were found to have contributed to methane migration into drinking water. Ingraffea et al. f20141
identified a significant increase in mechanical integrity problems such as casing leaks, sustained
casing pressure, and insufficient cement from 2009 to 2011, rising from 5% to 6% of all newly
drilled oil and gas wells, followed by a decrease beginning in 2012 to about 2% of all wells, a
reduction of approximately 100 violations among 3,000 wells from 2011 to 2012. The rise in
mechanical integrity problems between 2009 and 2011 coincided with an increase in the number of
wells in unconventional reservoirs.
While all of the studies shown in the table used the same database, their results vary, not only
because of the different timeframes studied, but also because they used different definitions of what
violations constituted a mechanical integrity problem or failure. For example, Considine et al.
(2012) considered all events resulting in environmental damage—including effects such as
erosion—and found a relatively high violation rate. Davies etal. (2014) and Ingraffea et al. (2014)
investigated violations related to mechanical integrity, while Vidic etal. f20131 looked only at
mechanical integrity violations resulting in fluid migration out of the wellbore; these more specific
studies found relatively lower violation rates. Olawovin etal. (2013) performed a statistical
analysis that weighted violations based on risk and found that the most risky violations included
those involving pits, erosion, waste disposal, and blowout preventers.
Another source of information on contamination caused by wells is positive determination letters
(PDLs) issued by the PA DEP. PDLs are issued in response to a complaint when the state determines
that contamination did occur in proximity to oil and gas activities. The PDLs take into account the
impact, timing, mechanical integrity, and formation permeability; liability is presumed for wells
within a given distance if the oil and gas operator cannot refute that they caused the contamination,
based on pre-drilling sampling (Brantley etal.. 2014).1 Brantley etal. (2014) examined these PDLs,
and concluded that, between 2008 and 2012, the water supplies of approximately seven properties
were impacted; depending on the assumptions used to determine how many unconventional gas
wells affected a single property; this equates to a rate of 0.12 to 1.1% of the 6,061 wells begun in
that timeframe. While these oil and gas wells were linked to contamination of wells and springs, the
1 Under Pennsylvania's Oil and Gas Act, operators of oil or gas wells are presumed liable if water supplies within 1,000 ft
(305 m] were impacted within 6 months of drilling, unless the claim is rebutted by the operator; this was expanded to
2,500 ft (762 m] and 12 months in 2012.
6-31
-------
Chapter 6 - Well Injection
mechanisms for the impacts (including whether fluids may have been spilled at the surface or if
there was a pathway through the well or through the subsurface rock formation to the drinking
water resource) were not described by Brantley et al. (2014).
While the studies discussed above present possible explanations for higher violation incidences in
unconventional wells that are likely to be hydraulically fractured, it should be noted that other
explanations not specific to hydraulic fracturing are also possible. These could include different
inspection protocols and different formation types.
Cementing in horizontal wells, which are commonly hydraulically fractured, presents challenges
that can contribute to higher rates of mechanical integrity issues. The observation by Ingraffea et al.
f20141 that wells drilled in unconventional reservoirs (which are horizontal in Pennsylvania)
experience higher rates of structural integrity loss than conventional wells is supported by
conclusions of Sabins (1990). who noted that horizontal wells have more cementing problems
because they are more difficult to center properly and can be subject to settling of solids on the
bottom of the wellbore. Cementing in horizontal wells presents challenges that can contribute to
higher rates of mechanical integrity issues.
Thermal and cyclic stresses caused by intermittent operation also can stress cement (King and
King. 2013: Ali etal.. 20091. Increased pressures and cyclic stresses associated with hydraulic
fracturing operations can contribute to cement integrity losses and, if undetected, small mechanical
integrity problems can lead to larger ones. Temperature differences between the (typically
warmer) subsurface environment and the (typically cooler) injected fluids, followed by contact
with the (typically warmer) produced water, can lead to contraction of the well materials (both
casing and cement), which introduces additional stresses. Similar temperature changes may occur
when multiple fracturing stages are performed. Because the casing and cement have different
mechanical properties, they may respond differently to these stress cycles and debond.
Several studies illustrate the effects of cyclic stresses. Dusseault et al. (2000) indicate that wells that
have undergone several cycles of thermal or pressure changes will almost always show some
debonding between cement and casing. Another laboratory study by DeAndrade etal. (2015) found
that cycling temperatures between 61°F and 151°F (16°C and 66°C) at 35 bar pressure (2.5 MPa)
led to the formation of cracks in cement across both shale and sandstone formations. Cement
damage was more significant in sandstone formations and worsened with each thermal cycle. A
similar study by Roy etal. (2016) at ambient pressure did not find any cracks larger than 200
microns with temperature fluctuation between -40°F and 158°F (-40°C and 70°C), although
numerical modeling of the same scenario predicted that cracks up to 1 to 10 microns would form,
which would not have been detected by the methods used. Microannuli formed by this debonding
can serve as pathways for gas migration, in particular because the lighter density of gas provides a
larger driving force for migration through the microannuli than for heavier liquids.1 One laboratory
study indicated that microannuli on the order of 0.01 in (0.25 mm) could increase effective cement
permeability from 1 nD (1 x 10-21 m2) in good quality cement up to 1 mD (1 x 10-15 m2) fBachu and
Bennion. 2009). This six-order magnitude increase in permeability shows that even small
1 Microannuli can also form due to an inadequate cement job, e.g., poor mud removal or improper cement placement rate.
6-32
-------
Chapter 6 - Well Injection
microannuli can significantly increase the potential for flow through the cement. Typically, these
microannuli form at the interface between the casing and cement or between the cement and
formation. Debonding and formation of microannuli can occur through intermittent operation,
pressure tests, and workover operations fDusseaultetal.. 20001.1 While a small area of debonding
may not lead to fluid migration, the microannuli in the cement resulting from the debonding can
serve as initiation points for fracture propagation if re-pressurized gas enters the microannulus
fDusseault etal.. 20001.
A number of modeling studies have indicated that fractures can propagate upwards from existing
defects in cement or areas with poorer bonding (Kim etal.. 2016: Roy etal.. 2016: De Andrade etal..
20151. Feng etal. f20151 showed that fractures in cement tended to propagate upwards along the
wellbore instead of radially. Modeling studies have also shown that cements with lower Young's
modulus tend to propagate fractures more slowly than stiffer cements fKim etal.. 2016: Feng etal..
20151.2
The Council of Canadian Academies f20141 found that the repetitive pressure surges occurring
during the hydraulic fracturing process would make maintaining an intact cement seal more of a
challenge in these wells. Wang and Dahi Taleghani (2014) performed a modeling study, which
concluded that hydraulic fracturing pressures could initiate annular cracks in cement Another
study of well data indicated that cement failure rates are higher in intermediate casings compared
to other casings fMcDaniel et al.. 20141. The failures occurred after drilling and completion of wells,
and the authors surmised that the cement failures were most likely due to cyclic pressure stresses
caused by drilling. Theoretically, similar cyclic pressure events could also be experienced in the
production casing during multiple stages of hydraulic fracturing. Mechanical stresses associated
with well operation or workovers and pressure tests also may lead to small cracks in the cement,
which may provide migration pathways for fluid.
Corrosion can lead to cement failure. Cement can fail to maintain integrity as a result of degradation
of the cement after the cement is set Cement degradation can result from attack by corrosive brines
or chemicals such as sulfates, sulfides, and carbon dioxide that exist in formation fluids fRenpu.
20111. These chemicals can alter the chemical structure of the cement, resulting in increased
permeability or reduced strength and leading to loss of cement integrity over time. Additives or
specialty cements exist that can decrease cement susceptibility to specific chemicals.
6.2.2.3 Well Age
Hydraulic fracturing within older (legacy) wells has the potential to impact drinking water
resources, either due to inadequate design and construction or degradation of the well components
over time that afford pathways for the unintended migration of fluids. While new wells can be
specifically designed to withstand the stresses associated with hydraulic fracturing operations,
1A workover refers to any maintenance activity performed on a well that involves ceasing operations and removing the
wellhead. Depending on the purpose of the workover and the tools used, workovers may induce pressure changes in the
well.
2 Young's modulus, a ratio of stress to strain, is a measure ofthe rigidity ofa material.
6-33
-------
Chapter 6 - Well Injection
older wells, which are sometimes used in hydraulic fracturing operations, may not have been
designed to the same specifications, and their reuse for this purpose could be a concern.
Aging and extended use of a well contribute to casing corrosion and degradation, and the potential
for fluid migration related to compromised casing tends to be higher in older wells. For example,
exposure to corrosive chemicals such as hydrogen sulfide, carbonic acid, and brines can accelerate
corrosion (Renpu. 20111. Aiani and Kelkar (20121 studied wells in Oklahoma and found a
correlation between well age and mechanical integrity issues. Specifically, in wells spaced between
1,000 and 2,000 ft (300 and 600 m) from a well being fractured, the likelihood of impact on the well
(defined in the study as a loss of gas production or increase in water production) rose from
approximately 20% to 60% as the well's age increased from 200 days to over 600 days. Age was
also found to be a factor in mechanical integrity problems in a study of wells drilled offshore in the
Gulf of Mexico (Brufatto etal.. 2003).
The Well File Review (U.S. EPA. 2016c. 2015n) provides evidence that fracturing does occur in
older wells, including re-entering existing wells to fracture them for the first time or re-fracturing
in wells that have been previously fractured. The Well File Review found that the median age of
wells being initially fractured was 45 days, with well ages at time of fracturing ranging from 8 days
to nearly 51 years. While 64% of the wells studied in the Well File Review were fractured within 6
months of the well spud date, the median age for wells being re-fractured was 6 years.12 An
estimated 11% of fracture jobs studied in the Well File Review were re-completions in a different
zone than the original fracture job and 8% were re-fractures in the same zone as the original
fracture job.3 4
The Well File Review also found that well component failures appeared to occur more frequently in
older wells that were being re-completed or re-fractured.5 The failure rate in hydraulic fracturing
jobs involving re-completions and re-fractures was 6%, compared to 2% for hydraulic fracturing
jobs in wells that had not been previously fractured.6 7 While the confidence levels overlap, there is
an indication that re-fractured and re-completed wells are more likely to suffer a failure of one or
more components during hydraulic fracturing operations.
Frac strings, which are specialized pieces of casing inserted inside the production casing, can be
used to protect older casing during fracturing. However, the effect of hydraulic fracturing on the
cement on the production casing in older wells is unknown. One study on re-fracturing of wells
noted that the mechanical integrity of the well was a key factor in determining the success or failure
of the fracture treatment (Vincent. 2011). The Well File Review (U.S. EPA. 2016c) found that
1 Spudding refers to starting the well drilling process by removing rock, dirt, and other sedimentary material with the drill
bit.
2 64% of wells (95% confidence interval: 48 - 77% of wells].
311%) of jobs (95%o confidence interval: 5 - 23% of jobs].
4 8%> of jobs (95%o confidence interval: 5 - 12% of jobs].
5 The Well File Review defines a failure as a defect in a well component that allows fluid to flow into an annular space.
6 6%> failure rate (95%> confidence interval: 2 - 19%> failure rate].
7 2%o failure rate (95%> confidence interval: 0.5 - 8%o failure rate].
6-34
-------
Chapter 6 - Well Injection
failures occurred more frequently in completions using frac strings, with failures occurring 20% of
the time, compared to failures occurring 0.9% of the time when a frac string was not used.12
Note that there are also potential issues related to where these older wells are sited. For example,
some wells could be in areas with naturally occurring subsurface faults or fractures that could not
be detected or fully characterized with the technologies available at the time of construction. It is
also possible that, in areas of historic petroleum exploration, old abandoned wells can be present
which may have been improperly plugged or have degraded over time.3 These wells could serve as
pathways for fluid migration if they are located within the fracture network of the well; see Section
6.3.2.
6.2.2.4 Sustained Casing Pressure
Sustained casing pressure illustrates how the issues related to casing and cement discussed in the
preceding sections can work together and be difficult to differentiate.4 It is an indicator that
pathways within the well related to the well's casing, cement, or both allowed fluid movement to
occur. Sustained casing pressure can result from casing leaks, uncemented intervals, microannuli,
or some combination of the three, which can be an indication that a well has lost mechanical
integrity. Sustained casing pressure can be observed when an annulus (either the annulus between
the tubing and production casing or between any two casings) is exposed to a source of nearly
continuous elevated pressure. Goodwin and Crook f!9921 found that sudden increases in sustained
casing pressure occurred in wells that were exposed to high temperatures and pressures.
Subsequent logging of these wells showed that the high temperatures and pressures led to shearing
of the cement/casing interface and a total loss of the cement bond. Alv etal. (20151 demonstrated
methods using a combination of chemical analysis, isotopic analysis, well logs, and drilling records
to identify the most likely source of fluids causing sustained casing pressure.
Sustained casing pressure occurs more frequently in older wells and horizontal or deviated wells.
One study found that sustained casing pressure becomes a greater concern as a well ages. Sustained
casing pressure was found in less than 10% of wells that were less than a year old, but was present
in up to 50% of 15-year-old wells fBrufatto etal.. 20031. While these wells may not have been
hydraulically fractured, the study demonstrates that older wells can exhibit more mechanical
integrity problems. Fleckenstein et al. (20151 also found that older wells exhibited more barrier
failures, including sustained casing pressure. They reported that 3.53% of the wells in the study
with under-pressured intermediate gas zones developed sustained casing pressure, although it is
likely the sustained casing pressure was due to poor well design (i.e., under older standards) rather
1 20% failure rate (95% confidence interval: 10 - 36% failure rate].
2 0.9%) failure rate (95%> confidence interval: 0.8 - 1.0% failure rate].
3 An abandoned well refers to a well that is no longer being used, either because it is not economically producing or it
cannot be used because of its poor condition.
4 Sustained casing pressure is pressure in any well annulus that is measurable at the wellhead and rebuilds after it is bled
down, not caused solely by temperature fluctuations or imposed by the operator fSkierven etal.. 2011], If the pressure is
relieved by venting natural gas from the annulus to the atmosphere, it will build up again once the annulus is closed (i.e.,
the pressure is sustained]. The return of pressure indicates that there is a small leak in a casing or through uncemented or
poorly cemented intervals that exposes the annulus to a pressured source of gas. It is possible to have pressure in more
than one of the annuli.
6-35
-------
Chapter 6 - Well Injection
than well age. Watson and Bachu (20091 found that a higher portion of deviated wells had
sustained casing pressure compared to vertical wells. Increased pressures and cyclic stresses fSved
and Cutler. 20101 during hydraulic fracturing and difficulty in cementing horizontal wells fSabins.
19901 also can lead to increased instances of sustained casing pressure fMuehlenbachs etal.. 2012:
Rowe and Muehlenbachs. 19991.
Sustained casing pressure can be a concern for several reasons. If the pressures are allowed to build
up to above the burst pressure of the exterior casing or the collapse pressure of the interior casing,
the casing may fail. Increased pressure can also cause gas or liquid to enter lower-pressured
formations that are exposed to the annulus either through leaks or uncemented sections.
Laboratory experiments by Harrison f!9851 demonstrated that over-pressurized gas in the annulus
could cause rapid movement of gas into drinking water resources if a permeable pathway exists
between the annulus and the groundwater. Over-pressurization of the annulus is commonly
relieved by venting the annulus to the atmosphere; however, this does not address the underlying
problem in the well and can result in additional releases of methane to the atmosphere.
One example of an area where sustained casing pressure is common is Alberta, Canada, where 14%
of the wells drilled since 1971 experienced serious sustained casing flow. This was defined in a
study by Tackson and Dussealt (2014) as more than 10,594 ft3 (300 m3)/day at pressures higher
than 0.48 psi/ft (11 kPa/m) of depth times the depth of the surface casing. Another study in the
same area found gas in nearby drinking water wells had a composition consistent with biogenic
methane mixing with methane from nearby coalbed methane and deeper natural gas fields fTillev
and Muehlenbachs. 20121.
In a few cases, sustained casing pressure in wells that have been hydraulically fractured may have
been linked to drinking water contamination, although it is challenging to definitively determine
the actual cause. In one study in northeastern Pennsylvania, methane to ethane ratios and isotopic
signatures were used to investigate stray gas migration into domestic drinking water (U.S. EPA.
2014fl. Composition of the gas in the water wells was consistent with that of the gas found in
nearby gas wells with sustained casing pressures; other possible sources of the gas could not be
ruled out. Several gas wells in the study area were cited by the PA DEP for having elevated
sustained casing annulus pressures. One such case included four well pads with two wells drilled on
each pad in southeastern Bradford County. The wells, drilled between September 2009 and May
2010, were 6,890 to 7,546 ft (2,100 to 2,300 m) deep and had surface casing to 984 ft (300 m). The
casing below the surface casing was uncemented. All four wells experienced sustained casing
pressure, with pressures ranging from 483 to 909 psi (3.3 to 6.3 MPa). Methane appeared in three
nearby domestic drinking water wells in July 2010. Investigation into the cause of the methane
contamination identified the drilled gas wells with sustained casing pressure as the most likely
cause. The likely path was over-pressured gas from intermediate zones above the Marcellus Shale
entering the uncemented well annulus and traveling up the annulus and along bedding planes
which intersected the well annulus.1 The determination was based on multiple lines of evidence,
including: no methane present in a pre-drill sample, increases in methane after the wells had been
1A bedding plane is the surface that separates two layers of stratified rocks.
6-36
-------
Chapter 6 - Well Injection
drilled, similar isotopic composition of the gas in the domestic wells and the gas in the annular
space of the gas wells, and the presence of bedding planes which intersected the uncemented
portion of the gas wells leading upwards toward the domestic wells (Llewellyn etal.. 2015).
Adequate well design, detection (i.e., through annulus pressure monitoring), and repair of sustained
casing pressure reduce the potential for fluid movement (See Chapter 10 for additional discussion
of practices that can reduce the frequency or severity of impacts to drinking water quality.) Watson
and Bachu f20091 found that regulations requiring monitoring and repair of sustained casing vent
flow or sustained casing pressure had a positive effect on lowering leak rates. The authors also
found injection wells initially designed for the higher pressures associated with injection (vs.
production) experienced sustained casing pressure less often than those that were retrofitted
fWatson and Bachu. 20091. As mentioned above, Fleckenstein etal. f20151 found that placing the
surface casing below all potential sources of drinking water and cementing intermediate gas zones
significantly reduced sustained casing pressure.
Another study in Mamm Creek, Colorado, obtained similar results. The Mamm Creek field is in an
area where lost cement and shallow, gas-containing formations are common. All the wells in the
formation were hydraulically fractured (S.S. Papadopulos & Associates. 2008). A number of wells in
the area have experienced sustained casing pressure, and methane has been found in several
drinking water wells along with seeps into local creeks and ponds. In one well, drilled in January
2004, four pressured gas zones were encountered during drilling and there was a lost cement
incident, which resulted in the cement top being more than 4,000 ft (1,000 m) lower than originally
intended. Due to high bradenhead pressure (661 psi, or 4.6 MPa), cement remediation efforts were
implemented (Crescent. 2011: COGCC. 2004).1 The operator of this well was later cited by the
Colorado Oil and Gas Conservation Commission (COGCC) for causing natural gas and benzene to
seep into a nearby creek. The proposed route of contamination was contaminants flowing up the
well annulus and then along a fault The proposed contamination route appeared to be validated
because, once remedial cementing was performed on the well, methane and benzene levels in the
creek began to drop (Science Based Solutions LLC. 2014). In response to the incident, the state
instituted requirements to identify and cement above the top of the highest gas-producing
formation in the area and to monitor casing pressures after cementing.
A study in the Woodford Shale in Oklahoma examined how various cement design factors affected
sustained casing pressure fLandrv etal.. 20151. The study focused on wells in the Cana-Woodford
basin, a very deep basin at 11,000 to 15,000 ft (3,400 to 4,600 m) below ground surface, where the
depth, long laterals, fracture gradients, and low permeability of the formations in the basin make
cementing a challenge. One operator had seven test wells in the basin, of which six exhibited
sustained casing pressure, usually after hydraulic fracturing operations. In early designs, the
operator had not been using centralizers on the horizontal sections of the well, because they
increased the frequency of stuck pipe. However, improvements in centralizer design allowed the
operator to use centralizers more frequently on later well designs, and the operator tried several
different techniques to address the sustained casing pressure problems, with varying results:
1 Bradenhead pressure is pressure between two casings in an oil and gas well.
6-37
-------
Chapter 6 - Well Injection
• In three of the wells, the operator used three different techniques: a conventional cement
job with a water-based drilling mud and single slurry design; oil-based mud with single
slurry design; and a foamed cement to cement the vertical portion of the well from the
kickoff point up with conventional water-based cement on the lateral. All three of these
wells experienced sustained casing pressure after hydraulic fracturing operations.
• In a fourth well, in 2013, the operator used centralizers, with three centralizers per every
two casing joints along the lateral and one centralizer per joint in the vertical section. The
design also involved an enhanced spacer fluid to remove drilling mud and a self-healing
cement in the upper portion of the well. While some channeling was detected in this well,
the channels were not connected and did not lead to sustained casing pressure.
The operator constructed an additional 21 wells using the same technique as was performed in the
fourth well, and 20 did not show any sustained casing pressure after fracturing. This study shows
the importance of cement design factors, such as casing centralization and mud removal, in
preventing sustained casing pressure.
Not every well that shows positive pressure in the annulus poses a potential problem. Sustained
pressure is only a problem when it exceeds the ability of the wellbore to contain it or when it
indicates leaks in the cement or casing fTIPRO. 20121. A variety of management options are
available for managing such pressure including venting, remedial cementing, and use of kill fluids in
the annulus fTIPRO. 20121.1 While venting may be a common method to address sustained casing
pressure, it does not address the underlying mechanical integrity failure and is only a temporary
solution. Furthermore, venting releases fluids at the wellhead which, if gaseous, can contribute to
increased atmospheric emissions, or if liquid, potential spills on the surface.
6.3 Fluid Migration Associated with Induced Fractures within Subsurface
Formations
This section discusses potential pathways for fluid movement associated with induced fractures
and subsurface formations (outside of the well system described in Section 6.2). It examines the
potential for fluid migration into drinking water resources by evaluating the development of
migration pathways within subsurface formations, the flow of injected and formation fluids, and
important factors that affect these processes.
Fluid movement requires both a physical pathway (e.g., via the interconnected pores within a
permeable rock matrix or via a fracture in the rock) and a driving force.2 In subsurface formations,
fluid movement is driven by the existence of a hydraulic gradient, which depends on elevation and
pressure and is influenced by fluid density, composition, and temperature (Pinder and Celia. 2006).
1A kill fluid is a weighted fluid with a density that is sufficient to overcome the formation pressure and prevent fluids
from flowing up the wellbore.
2 Permeability (i.e., intrinsic or absolute permeability] of formations describes the ability of water to move through the
formation matrix, and it depends on the rock's grain size and the connectedness of the void spaces between the grains.
Where multiple phases of fluids exist in the pore space, the flow of fluids also depends on relative permeabilities.
6-38
-------
Chapter 6 - Well Injection
In the context of hydraulic fracturing, two key factors govern fluid migration during and after the
hydraulic fracturing event:
• Pressure differentials in the reservoir, which are influenced both by initial subsurface
conditions and by the pressures created by injection and production regimes. Specific
factors that may influence pressure differentials include structural or topographic features,
over-pressure in the shale reservoir, or a temporary increase in pressure as a result of fluid
injection during hydraulic fracturing (Birdsell et al.. 2015a).
• Buoyancy, which is driven by density differences among and between gases and liquids.
Fluid migration can occur when these density differences exist in the presence of a pathway
fPinder and Gray. 20081.
During hydraulic fracturing, pressurized fluids leaving the well create fractures within the
production zone and then enter the formation through these newly created (induced) fractures.
Unintended fluid migration can result from this fracturing process. Migration pathways to drinking
water resources could develop as a result of changes in the subsurface flow or pressure regime
associated with hydraulic fracturing; via fractures that extend beyond the intended formation or
that intersect existing natural faults or fractures; and via fractures that intersect offset wells or
other artificial structures flackson et al.. 2013dl. These subsurface pathways may facilitate the
migration of fluids by themselves or in conjunction with the well-based pathways described in
Section 6.2. Fluids potentially available for migration include both fluids injected into the well
(including leakoff) and fluids already present in the formation (including brine or natural gas).1
The potential for subsurface fluid migration into drinking water resources can be evaluated during
two different time periods (Kim and Moridis. 2015):
1. Following the initiation of fractures in the reservoir¦, prior to any oil or gas production. The
injected fluid, pressurizing the formation, flows through the fractures and the fractures
grow into the reservoir. Fluid leaks off into the formation, allowing the fractures to close
except where they are held open by the proppant (Adachi et al.. 2007). Fractures will
generally continue to propagate until the fluid lost to leakoff is equal to the fluid injection
rate (King and Durham. 2015).
2. During the production periodafter fracturing is completed and pressure in the fractures is
reduced. At this time, fluids (including oil/gas and produced water) flow from the reservoir
into the well. As fluids are withdrawn from the formation, pore pressure decreases; as a
result, the effective stress applied to fractures increases and (in the absence of proppant)
fractures will close fAvbar etal.. 20151.
Note that these two time periods vary in duration. As described in Chapter 3, the first period of
fracture creation and propagation (i.e., the hydraulic fracturing itself) is a relatively short-term
process, typically lasting 2 to 10 days, depending on the number of stages in the fracture treatment
1 Leakoff is the fraction of the injected fluid that infiltrates into the formation and is not recovered (i.e., it "leaks off and
does not return through the well to the surface] during production fEconomides et al.. 20071. Fluids that leak off and are
not recovered are sometimes referred to as "lost" fluids.
6-39
-------
Chapter 6 - Well Injection
design. On the other hand, operation of the well for production covers a substantially longer period
(depending on many factors such as the amount of hydrocarbons in place and economic
considerations), and can be as long as 40 or 60 years in onshore tight gas reservoirs fRoss and King.
20071.
The following discussion of potential subsurface fluid migration into drinking water resources
focuses primarily on the physical movement of fluids and the factors affecting this movement.
Section 6.3.1 describes the basic principles of subsurface fracture creation, geometry, and
propagation, to provide context for the discussion of potential fluid migration pathways in Section
6.3.2. Geochemical and biogeochemical reactions among hydraulic fracturing fluids, formation
fluids, subsurface microbes, and rock formations are another important component of subsurface
fluid migration and transport. See Chapter 7 for a discussion of the processes that affect pore fluid
biogeochemistry and influence the chemical and microbial composition of produced water.
6.3.1 Overview of Subsurface Fracture Growth
Fracture initiation and growth is a highly complex process due to the heterogeneous nature of the
subsurface environment. As shown in Figure 6-5, fracture formation is controlled by the three in
situ principal compressive stresses: the vertical stress, the maximum horizontal stress, and the
minimum horizontal stress. During hydraulic fracturing, pressurized fluid injection creates high
pore pressures around the well. Fractures form when this pressure exceeds the local least principal
stress and the tensile strength of the rock (Zoback. 2010: Fjaer etal.. 2008).
Fractures propagate (increase in length) in the direction of the maximum principal stress; they are
tensile fractures that open in the direction of least resistance and then propagate in the plane of the
greatest and intermediate stresses fNolen-Hoeksema. 20131. Deep in the subsurface, the maximum
principal stress is generally in the vertical direction, because the overburden (the weight of
overlying rock) is the largest single stress. Therefore, in deep formations, fracture orientation is
expected to be vertical. This is the scenario illustrated in Figure 6-5. At shallower depths, where the
rock is subjected to less pressure from the overburden, more fracture propagation is expected to be
in the horizontal direction. Using tiltmeter data from over 10,000 fractures in various North
American shale reservoirs, Fisher and Warpinski (2012) found that induced fractures deeper than
about 4,000 ft (1,000 m) are primarily vertical (see below for more information on tiltmeters).
Between approximately 4,000 and 2,000 ft (1,000 and 600 m), they observed that fracture
complexity increases, and fractures shallower than about 2,000 ft (600 m) are primarily (though
not entirely) horizontal.1 However, local geologic conditions can cause fracture orientations to
deviate from these general trends (Ryan etal.. 2015). Horizontal fracturing can also occur in deeper
1 Fracture complexity is the ratio of horizontal-to-vertical fracture volume distribution, as defined hv Fisherand
Warpinski (2012). Fracture complexity is higher in fractures with a larger horizontal component. For the reasons
explained above, this is more likely to occur at shallower depths. However, even in shallow zones, fractures are unlikely to
be completely horizontal. As noted by Fisher and Warpinski, "All of the fractures do not necessarily turn horizontal; they
might have significant vertical and horizontal components with more of a T-shaped geometry." In the Fisherand
Warpinski data set, the maximum horizontal component of the fractures is approximately 70%.
6-40
-------
Chapter 6 - Well Injection
settings in some less-common reservoir environments where the principal stresses have been
altered by salt intrusions or similar types of geologic activity flones and Britt. 20091.
Figure 6-5. Hydraulic fracture planes (represented as ovals), with respect to the principal
subsurface compressive stresses: Sv (the vertical stress), Sh (the maximum horizontal stress),
and Sh (the minimum horizontal stress).
In addition to the principal subsurface stresses, a variety of factors and processes affect the
complex process of fracture creation, propagation, geometry, and containment.1 Computational
modeling techniques have been developed to simulate fracture creation and propagation and to
provide a better understanding of this complex process (Kim and Moridis. 2013).2 Modeling
hydraulic fracturing in shale or tight gas reservoirs requires integrating the physics of both flow
and geomechanics to account for fluid flow, fracture propagation, and dynamic changes in pore
volume and permeability. Some important flow and geomechanical parameters included in these
1 Fracture geometry refers to characteristics of the fracture such as height and aperture (width).
2 There are different kinds of mathematical models. Analytical models have a closed-form solution and therefore are
relatively simple to solve. In contrast, computational models (also called numerical models] require more extensive
computational resources and are used to study the behavior of complex systems.
6-41
-------
Chapter 6 - Well Injection
types of advanced models are: permeability, porosity, Young's modulus, Poisson's ratio, and tensile
strength, as well as heterogeneities associated with these parameters.1
Based on modeling and laboratory experiments (e.g., by Khanna and Kotousov. 2016: Li etal..
2016c: Li etal.. 2016b: de Pater. 2015: Kim and Moridis. 2015: Lee etal.. 2015: Narasimhan etal..
2015: Smith and Montgomery. 2015: Wang and Rahman. 2015: Kim and Moridis. 20131. below are
some of the factors that have been noted in the literature as influencing fracture growth:
• Geologic properties of the production zone such as rock type and composition, permeability,
thickness, and the presence of pre-existing natural fractures;
• The presence, composition, and properties of the liquids and gases trapped in pore spaces;
• Geomechanical properties, including tensile strength, Young's modulus, and the pressure at
which the rock will fracture;
• Characteristics of the interface (boundary) between adjacent rock layers; and
• Operational characteristics, including injection rate and pressure, the properties of the
hydraulic fracturing fluids, and fracture spacing.
Some modeling investigations have indicated that the vertical propagation of fractures (due to
tensile failure) may be limited by shear failure, which increases the permeability of the formation
and allows more fluid to leak off into the rock. These findings demonstrate that elevated pore
pressure can cause shear failure, thus further affecting matrix permeability, flow regimes, and
leakoff fPaneshv. 20091.
It is important to note that, while computational modeling is a useful tool to understand complex
systems, modeling has limitations and associated uncertainties. All models rely on assumptions and
simplifications, and there is, as stated by Ryan etal. (2015). "currently no single numerical
approach that simultaneously includes the most important thermo-hydromechanical and chemical
processes which occur during the migration of gas and fluids along faults and leaky wellbores."
Uncertainties in selecting values for input parameters and potentially inadequate field data for
model verification limit the reliability of model predictions.
In addition to their use in research applications, analytical and numerical modeling approaches are
used to design hydraulic fracturing treatments and predict the extent of fractured areas (Adachiet
al.. 2007). Specifically, modeling techniques are used to assess the treatment's sensitivity to critical
parameters such as injection rate, treatment volumes, fluid viscosity, and leakoff. Existing models
range from simpler (typically two-dimensional) theoretical models to computationally more
complicated three-dimensional models.
Monitoring of hydraulic fracturing operations can also provide insights into fracture development.
Monitoring techniques involve both operational monitoring methods and "external" methods not
1 As described in Section 6.2.2.2, Young's modulus, a ratio of stress to strain, is a measure of the rigidity of a material.
Poisson's ratio is a ratio of transverse-to-axial (or latitudinal-to-longitudinal] strain, and it characterizes how a material is
deformed under pressure. See Zoback ("20101 for more information on the geomechanical properties of reservoir rocks.
6-42
-------
Chapter 6 - Well Injection
directly related to the production operation. Operational monitoring refers to the monitoring of
pressure and flow rate, along with related parameters such as fluid density and additive
concentrations, using surface equipment and/or downhole sensors fEberhard. 20111. This
monitoring is conducted to ensure the operation is proceeding as planned and to determine if
operational parameters need to be adjusted. Interpretation of pressure data can be used to better
understand fracture behavior (Kim and Wang. 20141. For example, pressure data from previous
hydraulic fracturing operations can indicate whether a geologic barrier to fracture growth exists
and whether the barrier has been penetrated, or whether fractures have intersected with natural
fractures or faults (API. 2015). Anomalies in operational monitoring data can also indicate whether
an unexpected event has occurred, such as communication with another well (Section 6.3.2.3).
As described in Chapter 4, the volume of fluid injected is typically monitored and tracked to provide
information on the volume and extent of fractures created fFlewelling et al.. 20131. However,
numerical investigations have found that reservoir gas flows into the fractures immediately after
they open from hydraulic fracturing, and injection pressurizes both gas and water within the
fracture to induce further fracture propagation fKim and Moridis. 20151. Therefore, the fracture
volume can be larger than the injected fluid volume. As a result, simple estimation of fracture
volume based on the amount of injected fluid may underestimate fracture growth, and additional
information (e.g., from geophysical monitoring techniques) is needed to accurately predict the
extent of induced fractures.
External monitoring technologies can also be used to collect data on fracture characteristics and
extent during hydraulic fracturing and/or production. These monitoring methods can be divided
into near-wellbore and far-field techniques. Near-wellbore techniques include the use of tracers,
temperature logs, video logs, and caliper logs that measure conditions in and immediately around
the wellbore fHolditch. 20071. However, near-wellbore techniques and logs only provide
information for, at most, a distance of two to three wellbore diameters from the well and are,
therefore, not suited for tracking fractures for their entire length (Holditch. 2007).
Far-field methods, such as microseismic monitoring or tiltmeters, are used if the intent is to
estimate fracture growth and height across the entire fractured reservoir area. Microseismic
monitoring involves placing geophones in a position to detect the very small amounts of seismic
energy generated during subsurface fracturing (Warpinski. 2009).1 Monitoring these microseismic
events gives an idea of the location and size of the fracture network, as well as the orientation and
complexity of fracturing fFisher and Warpinski. 20121. Using the results of microseismic
monitoring in conjunction with other information, such as time-lapse, multicomponent seismic data
(collected with surface surveys), can provide additional information for understanding fracture
complexity and the interaction between natural and induced fractures (D'Amico and Davis. 2015).
The Well File Review fU.S. EPA. 2016cl found that microseismic monitoring was conducted at 0.5%
(100) of the hydraulic fracturing jobs studied.2 Tiltmeters, which measure extremely small
deformations in the earth, can be used to determine the direction and volume of the fractures and,
1 Typical microseismic events associated with hydraulic fracturing have a magnitude on the order of -2.5 (negative two
and half) ("Warpinski. 20091
2100 jobs (95% confidence interval: 40 - 300 jobs].
6-43
-------
Chapter 6 - Well Injection
within certain distances from the well, to estimate their dimensions (Lecampion et al.. 20051. Other
monitoring techniques, such as seismic surveys, can also be used to gather information about the
subsurface environment. For example, Vinal and Davis f20151 demonstrated the use of time-lapse
multi-component seismic surveys to monitor changes in the overburden due to hydraulic
fracturing. Chapter 10 provides additional discussion of factors and practices, such as site
monitoring, that can reduce the frequency or severity of impacts to drinking water quality.
6.3.2 Migration of Fluids through Pathways Related to Fractures/Formations
As described above, subsurface migration of fluids requires a pathway, induced or natural, with
enough permeability to allow fluids to flow, as well as a hydraulic gradient physically driving the
movement The following subsections describe and evaluate potential pathways for the migration
of hydraulic fracturing fluids, hydrocarbons, or other fluids from producing formations to drinking
water resources. They also present cases where the existence of these pathways has been
documented. The potential subsurface migration pathways are categorized as follows:
(1) migration out of the production zone through pore space in the rock, (2) migration due to
fracture overgrowth out of the production zone, (3) migration via fractures intersecting offset wells
or other artificial structures, and (4) migration via fractures intersecting other geologic features,
such as permeable faults or pre-existing natural fractures. Although these four potential pathways
are discussed separately here, they may act in combination with each other or in combination with
pathways along the well (as discussed in Section 6.2) to affect drinking water resources.
The possibility of fluid migration between a hydrocarbon-bearing formation and a drinking water
resource can be related to the vertical distance between these formations (Reagan etal.. 2015:
Tackson etal.. 2013d). In general, as the separation distance between the production zone and a
drinking water aquifer decreases, the likelihood of upward migration of hydraulic fracturing to
drinking water aquifers increases fBirdsell etal.. 2015al. The separation distance between
hydraulically fractured producing zones and drinking water resources (and these formations' depth
from the surface) varies substantially among shale gas plays, coalbed methane plays, and other
areas where hydraulic fracturing takes place in the United States (Figure 6-6 and Table 6-3). Many
hydraulic fracturing operations target deep shale zones such as the Marcellus or Haynesville/
Bossier, where the vertical distance between the top of the shale formation and the base of drinking
water resources may be 1 mi (1.6 km) or greater. This is reflected in the Well File Review, in which
approximately half of the wells were estimated to have 5,000 ft (2,000 m) or more of measured
distance along the wellbore between the point of the shallowest hydraulic fracturing and the
operator-reported base of the protected groundwater resource fU.S. EPA. 2015^^ Similarly, in a
review of FracFocus data from over 40,000 wells across the United States, Tackson etal. f20151
found that the median depth of wells used for hydraulic fracturing was 8,180 ft (2,490 m) and the
mean depth was 8,290 ft (2,530 m).
1 In the Well File Review, 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. True vertical separation
distances were not reported in the Well File Review. Measured distance along a well is equal to the true vertical distance
only in straight, vertical wells. Otherwise, the true vertical distance is less than the measured distance.
6-44
-------
Chapter 6 - Well Injection
Drinking Water Resource
No Vertical
| Separation
= .
v Drinking Water Resource
and Targeted Rock Formation
Targeted Rock Formation
Figure 6-6. Vertical distances in the subsurface separating drinking water resources and
hydraulic fracturing depths.
Separation
Distance in
Measured Depth
Vertical
Separation
Distance
However, as shown in Table 6-3, some hydraulic fracturing operations occur at shallower depths or
in closer proximity to drinking water resources. For example, both the Antrim and the New Albany
plays are relatively shallow, with distances of 100 to 1,900 ft (31 to 580 m) between the producing
formation and the base of drinking water resources. In the Jackson etal. (2015) review of
FracFocus data, 16% of wells reviewed were within 1 mi (1.6 km) of the surface and 3% were
within 2,000 ft (600 m) of the surface.1 The distribution of the more shallow hydraulically fractured
wells varied nationally but was concentrated in Texas, California, Arkansas, and Wyoming. For
example, in California and Arkansas, 88% and 85% of hydraulically fractured wells, respecti vely,
were within about 5,000 ft (2,000 m) of the surface. Overall, the Well File Review found a higher
proportion of relatively shallow wells—the data in the Well File Review indicated that 20% of wells
used for hydraulic fracturing (an estimated 4,600 wells) had less than 2,000 ft (600 m) between the
shallowest point of the fractures and the base of protected groundwater resources (U.S. EPA.
2015 rt ).2 This is likely because the Well File Review results are more representative of hydraulic
fracturing operations across the country; lacksonetal. f2015) acknowledge that their analysis
1 lackson et al. f 2015^1 use true vertical depth data from FracFocus; this represents the depth of the well but not
necessarily the depth of the fractures. The depth of the fractures may be shallower than the true vertical depth of the well,
though lackson etal. f20151 note that most states do not require operators to submit information on the true vertical
depth to the top of the fractures.
2 4,600 wells (95% confidence interval: 900 - 8,300 wells). The Well File Review defines this separation distance as the
measured depth of the point of shallowest hydraulic fracturing in the well, minus the depth of the operator-reported
protected groundwater resource.
6-45
-------
Chapter 6 - Well Injection
underestimates the occurrence of relatively shallow hydraulic fracturing for states in which
FracFocus reporting is not required.
Table 6-3. Comparing the approximate depth and thickness of selected U.S. shale gas plays
and coalbed methane basins.
Shale data are reported in GWPC and ALL Consulting (2009) and NETL (2013); coalbed methane data are reported
in ALL Consulting (2004) and U.S. EPA (2004a). See Chapter 3 for information on the locations of these basins,
plays, and formations.
Basin/play/
formation3
Approx. depth
(ft [m] below surface)
Approx. net
thickness (ft [m])
Distance between top of
production zone and base
of treatable water (ft [m])b
Shale plays
Antrim
600 to 2,200
[200 to 670]
70 to 120
[20 to 37]
300 to 1,900
[90 to 580]
Barnett
6,500 to 8,500
[2,000 to 2,600]
100 to 600
[30 to 200]
5,300 to 7,300
[1,600 to 2,200]
Eagle Ford
4,000 to 12,000
[1,000 to 3,700]
250
[76]
2,800 to 10,800
[850 to 3,290]
Fayetteville
1,000 to 7,000
[300 to 2,000]
20 to 200
[6 to 60]
500 to 6,500
[200 to 2,000]
Haynesville-Bossier
10,500 to 13,500
[3,200 to 4,120]
200 to 300
[60 to 90]
10,100 to 13,100
[3,080 to 3,990]
Marcellus
4,000 to 8,500
[1,000 to 2,600]
50 to 200
[20 to 60]
2,125 to 7,650
[648 to 2,330]
New Albany
500 to 2,000
[200 to 600]
50 to 100
[20 to 30]
100 to 1,600
[30 to 490]
Woodford
6,000 to 11,000
[2,000 to 3,400]
120 to 220
[37 to 67]
5,600 to 10,600
[1,700 to 3,230]
Coalbed methane basins
Black Warrior
(Upper Pottsville)
0 to 3,500
[0 to 1,100]
< 1 to > 70
[< 1 to > 20]
As little as zero0
Powder River
(Fort Union)
450 to >6,500
[140 to >2,000]
75
[23]
As little as zero0
Raton (Vermejo
and Raton)
< 500 to > 4,100
[< 200 to > 1,300]
10 to >140
[3 to >43]
As little as zero0
San Juan (Fruitland)
550 to 4,000
[170 to 1,000]
20 to 80
[6 to 20]
As little as zero0
a For coalbed methane, values are given for the specific coal units noted in parentheses.
b The base of treatable water is defined at the state level; the information in the table is based on depth data from state oil and
gas agencies and state geological survey data.
c Formation fluids in producing formations meet the salinity threshold that is used in some definitions of a drinking water
resource in at least some areas of the basin. See the discussion after Text Box 6-5 for more information about this definition.
6-46
-------
Chapter 6 - Well Injection
In coalbed methane plays, which are typically shallower than shale gas plays, vertical separation
distances can be even smaller. In the Raton Basin of southern Colorado and northern New Mexico,
approximately 10% of coalbed methane wells have less than 675 ft (206 m) of separation between
the gas wells' perforated intervals and the depth of local water wells. In certain areas of the basin,
this distance is less than 100 ft (31 m) (Watts. 2006). In California, nearly half of the hydraulic
fracturing has occurred at depths less than about 900 ft (300 m) (CCST. 2015b). with hundreds of
wells in the San Joaquin Valley between 150 ft (46 m) and 2,000 ft (600 m) deep flackson etal..
20151.
Some hydraulic fracturing operations are conducted within formations containing drinking water
resources (Table 6-3). One example of hydraulic fracturing taking place within a geologic formation
that is also used as a drinking water source is in the Wind River Basin in Wyoming (Digiulio and
Tackson, 2016; WYOGCC, 2014b; Wright etal., 2012). Vertical gas wells in this area target the lower
Wind River Formation and the underlying Fort Union Formation, which consist of interbedded
layers of sandstones, siltstones, and mudstones. The Wind River Formation also serves as the
principal source of domestic, municipal, and agricultural water in this rural area. There are no
laterally continuous confining layers of shale in the basin to prevent upward movement of fluids.
While flow in the basin generally tends to be downward, local areas of upward flow have been
documented ("Digiulio and Tackson. 20161. Assessing the relative depths of drinking water resources
and hydraulic fracturing operations near Pavillion, Wyoming, Digiulio and Tackson (2016) found
that approximately 50% of fracture jobs were within 1,969 ft (600 m) of the deepest domestic
drinking water well in the area, and that 10% were within 820 ft (250 m) (Digiulio and Tackson.
2016). Among the wells evaluated by DiGiulio and Jackson, the shallowest fracturing occurred at
1,057 ft (322 m) below ground surface, which is comparable to depths targeted for drinking water
withdrawal in the formation. See Text Box 6-5 for more information on Pavillion, Wyoming.
Text Box 6-5. Pavillion, Wyoming.
The Pavillion gas field is located east of the town of Pavillion, Wyoming. In addition to gas production, the
field is also home to rural residences that rely on approximately 40 private wells to supply drinking water.
The oldest known domestic water well in the field dates to 1934 fAME. 20161. Gas production in the field
began in 1960 and, by the 2000s, it had grown to producing from at least 180 wells. Most of these gas wells
were drilled since 1990, and approximately 140 to 145 were not plugged as of mid-2016 fAME. 2016: Digiulio
and lackson. 20161.
In the Pavillion gas field the same geologic formation that is used to produce hydrocarbons supplies the area's
drinking water fDigiulio and lackson. 20161. Water wells draw from the Wind River Formation, and gas is
extracted from both the Wind River Formation and the underlying Fort Union Formation. The Wind River
Formation contains variably permeable strata with lenses of relatively higher permeability rock enriched
with natural gas. Water quality is typically freshest nearer the surface, and there is no rock formation acting
as a natural barrier to separate the drinking water from hydrocarbons fDigiulio and lackson. 20161. There is
approximately 200 ft (60 m) vertical distance separating the deepest domestic well in the field from the
shallowest hydraulic fracturing, although there is approximately 2.5 mi (4 km) lateral distance between them
fAME. 2016: Digiulio and lackson. 20161.
(Text Box 6-5 is continued on the following page.)
6-47
-------
Chapter 6 - Well Injection
Text Box 6-5 (continued). Pavillion, Wyoming.
Gas Well
Monitoring Well Water Well Approximate depth
below ground (ft)
Surface casing length
220-2,300 ft
Uncemented
bradenhead annulus
Wellbore
Casing
1
Cement
1
Perforation
A
Sand pack
Natural gas
&
Water
Production casing
length ~6,000 ft
—=£r\
—O-t
'—O"
-30-750
775-980
Wind River Formation
Salinity range: ~600-6,000 mg/L TDS
Hydraulic fracturing depths:
typically 1,500 - 6,000 feet
"3,400
--C--
Fort Union Formation
Salinity range: "*1,000-5,000 mg/L TDS
Not to scale
"6,200
Following complaints by area residents about changes to their water quality in the mid-2000s, state and
federal agencies began a series of investigations, centering on various aspects of the site and supporting
differing conclusions about the source and mechanism of the water quality changes fAME. 20161
Twenty-five pits that were used to dispose of drill cuttings, drilling mud, and spent drilling fluids near some
of the water wells were also investigated as a potential source of the groundwater contamination. Based on
these evaluations, soil and/or groundwater remediation was performed at approximately six of the pits, no
further action was recommended at approximately twelve pits, and the remaining pits are receiving further
investigation fAME. 2016).
Samples collected from two monitoring wells at depths between those of the drinking water and active
intervals in gas production wells show elevated pH, unexpectedly high potassium values, and several organic
constituents, including natural gas, alcohols, phenols, glycols, and benzene, toluene, ethylbenzene, and
xylenes (BTEX] fDigiulio and lackson. 20161. The potential source of chemicals in these two monitoring wells
include formation water, contaminants remaining after well construction fAME. 2016) and hydraulic
fracturing and other oil and gas activities fDigiulio and lackson. 20161
Water samples collected from domestic wells contain dissolved methane and some contain high sodium and
sulfate concentrations. Organic chemicals have also been detected in some domestic wells fAME. 2016:
Digiulio and lackson. 20161. These same investigators suspect that pit proximity explains the origin of organic
chemicals. In addition, natural gases from intermediate depths not hydraulically fractured are likely moving
along some gas wellbores, potentially into zones used for drinking water fAME. 20161.
(Text Box 6-5 is continued on the following page.)
6-48
-------
Chapter 6 - Well Injection
Text Box 6-5 (continued). Pavillion, Wyoming.
Of about 40 production wells at which pressure was measured on the bradenhead annulus between the
production and surface casings, about 25% exhibited sustained casing pressure consistent with an ongoing
source of gas and/or liquid. Gas samples collected from bradenhead annuli, production tubing and casing, and
water wells indicate that the samples have similar gas compositions. This suggests a common origin, which is
consistent with long-term migration from a deeper source CAME, 2016; WYOGCC, 2014b").
Production wells may be the source of gas migration, and groundwater immediately around some of the
disposal pits has been affected fAME. 20161 However, the investigative reports conclude that identifying the
precise source(s) of the water quality issues is challenging due to the lack of comprehensive pre-drilling
water quality and other baseline monitoring, the unique hydrogeologic setting, and the difficulty of
identifying specific geologic or well pathways.
In other cases, hydraulic fracturing takes place in formations that are not currently being used as
sources of drinking water, but that meet the salinity threshold that is used in some definitions of
drinking water resources.1 This occurs in low-salinity coal-bearing formations in the Raton Basin of
Colorado (U.S. EPA. 2015k). the San Juan Basin of Colorado and New Mexico (U.S. EPA. 2004a). the
Powder River Basin of Montana and Wyoming (as described in Chapter 7), and in several other
coalbed methane plays. Hydraulic fracturing in these regions occurs in formations characterized by
total dissolved solids (TDS) values substantially lower than the 10,000 mg/L TDS value used in the
federal definition of an underground source of drinking water.2 Across various basins, coalbed
methane operations have been reported to occur in formations with 300 to 3,000 mg/L TDS and at
depths as shallow as 350 ft (110 m) fU.S. EPA. 2004al In one field in Alberta, Canada, there is
evidence that fracturing in the same formation as a drinking water resource (in combination with
mechanical integrity problems; see Section 6.2.2.4) led to gas migration into water wells fTillev and
Muehlenbachs. 2012).
California is another area where hydraulic fracturing occurs in shallow zones with low-salinity
groundwater. A study by the California Council on Science and Technology fCCST. 2015bl found
that 3% of the hydraulic fracturing in the state occurred within 2,000 ft (600 m) of the surface. In
California's San Joaquin Valley, hydraulic fracturing appears to have been conducted in formations
with a TDS of less than 1,500 mg/L (CCST. 2014). Another study in California examined the TDS
values of water samples taken during oil and gas activities and found that 15% to 19% of the oil and
1 For the purposes of this discussion, the federal definition of an underground source of drinking water is used. Pursuant
to 40 CFR 144.3, an underground source of drinking water is "an aquifer or its portion which supplies any public water
system; or which contains a sufficient quantity of groundwater to supply a public water system; and currently supplies
drinking water for human consumption; or contains fewer than 10,000 mg/L TDS; and which is not an exempted aquifer."
This definition is used by the EPA's Underground Injection Control Program, which regulates injection wells (but not
hydrocarbon production wells].
2 This salinity threshold is used as a point of comparison only. While the definition of an underground source of drinking
water is not exactly the same as the definition of a drinking water resource (and many states have their own definitions of
protected drinking water zones], the former provides a useful frame of reference when considering the ability of an
aquifer to potentially serve as a source of drinking water.
6-49
-------
Chapter 6 - Well Injection
gas activities in Kern County, California, occurred within zones containing water with less than
3,000 mg/L TDS fKang and Tackson. 20161.1
The overall frequency at which hydraulic fracturing occurs in formations that meet the definition of
drinking water resources across the United States is uncertain. Some information, however, that
provides insights on the occurrence and geographic distribution of this practice is available.
According to the Well File Review, an estimated 0.4% (90) of the 23,200 wells represented in that
study had perforations used for hydraulic fracturing that were placed shallower than the base of
the protected groundwater resources reported by well operators (U.S. EPA. 2 015n).2 Additional
information is available from a database of produced water composition data maintained by the
U.S. Geological Survey (USGS). The USGS produced water database contains results from analyses of
samples of produced water, including (among other data) samples collected from more than 8,500
oil and gas production wells in unconventional formations (coalbed methane, shale gas, tight gas,
and tight oil) within the contiguous United States.3 Just over 5,000 of these samples, which were
obtained from wells located in 37 states, reported TDS concentrations. Because the database does
not track whether samples were from wells that were hydraulically fractured, the EPA selected
samples from wells that were more likely to have been hydraulically fractured by restricting
samples to those collected in 1950 or later and to those that were collected from wells producing
from tight gas, tight oil, shale gas, or coalbed methane formations.4 This yielded 1,650 samples from
wells located in Alabama, Colorado, North Dakota, Utah, and Wyoming, with TDS concentrations
ranging from approximately 90 mg/L to 300,000 mg/L.5 Of the 1,650 samples, approximately 1,200
(from wells in Alabama, Colorado, Utah, and Wyoming) reported TDS concentrations at or below
10,000 mg/L, indicating that hydraulic fracturing there may have occurred within formations that
meet the salinity threshold that is used in some definitions of a drinking water resource. This
analysis, in conjunction with the result from the Well File Review, suggests that the overall
frequency of this occurrence is relatively low, but is concentrated in particular areas of the country.
6.3.2.1 Flow of Fluids Out of the Production Zone
One potential pathway for fluid migration out of the production formation into drinking water
resources is advective or dispersive flow of injected or displaced fluids through the formation
matrix. In this scenario, fluids (such as those "lost" to leakoff, which are not recovered during
1 Kern County accounts for 85 percent of the hydraulic fracturing that occurs in California fCCST. 2015bl
2 90 wells (95% confidence interval: 10 - 300 wells].
3 The EPA used the USGS Produced Water Geochemical Database Version 2.1 (USGS database v 2.1] for this analysis
fhttp: / /energv.cr.usgs.gov/prov/prodwat /I The database is comprised of produced water samples compiled by the USGS
from 25 individual databases, publications, or reports.
4 See Chapter 3, Text Box 3-1, which describes how commercial hydraulic fracturing began in the late 1940s.
5 For this analysis, the EPA assumed that produced water samples collected in 1950 or later from shale gas, tight oil, and
tight gas wells were from wells that had been hydraulically fractured. To estimate which coal bed methane wells had been
hydraulically fractured, the EPA matched API numbers from coal bed methane wells in the USGS database v 2.1 to the
same API numbers in the commercial database Drillinglnfo, in which hydraulically fractured wells had been identified by
the EPA using the assumptions described in Section 3.4. Wells with seemingly inaccurate (i.e., less than 12 digit] API
numbers were also excluded. Only coalbed methane wells from the USGS database v 2.1 that matched API numbers in the
Drillinglnfo database were retained for this analysis.
6-50
-------
Chapter 6 - Well Injection
production) would flow through the pore spaces of rock formations, moving from the production
zone into other formations. In deep, low-permeability shale and tight gas settings and where
induced fractures are contained within the production zone, flow through the production formation
has generally been considered an unlikely pathway for migration into drinking water resources
(lackson et al.. 2013d).
Leakoff into shale gas formations can be as high as 90% or more of the injected volume (Table 7-2).
The actual amount of leakoff depends on multiple factors, including the amount of injected fluid, the
concentration of different components in the fracture fluid, the hydraulic properties of the
reservoir (e.g., permeability), the composition of the formation matrix, the capillary pressure near
the fracture faces, and the period of time the well is shut in following hydraulic fracturing before
the start of production fKim etal.. 2014: Byrnes. 2011I1'2 Researchers generally agree that the
subsequent flow of this "lost" leakoff fluid is controlled or limited by processes such as imbibition
by capillary forces and adsorption onto clay minerals (Duttaetal.. 2014: Dehghanpour etal.. 2013:
Dehghanpour etal.. 2012: Rovchaudhuri et al.. 2011) and osmotic forces (Zhou. 2016: Wang and
Rahman. 2015: Engelder etal.. 20141.3*4 It has been suggested that these processes can sequester
the fluids in the producing formations permanently or for geologic time scales fEngelder etal..
2014: Engelder. 2012: Byrnes. 2011). Birdsell et al. (2015b) made quantitative estimates of the
amount of fluid that could be imbibed in shale formations. Their results indicate that between 15%
and 95% of injected fluid volumes may be imbibed in shale gas systems, while amounts are lower in
shale oil systems (3% to 27% of injected volumes). In modeling investigations, O'Mallev etal.
f20151 found that it is likely that most hydraulic fracturing fluid that does not flow back is stored in
rock pore spaces (i.e., having displaced the gas that was present there) and not fractures. The
amount that can be stored in fractures is highly dependent on the effective interconnected pore
lengths.
If the injected fluid is not sequestered in the immediate vicinity of the fracture network, migration
into drinking water resources would likely require a substantial upward hydraulic gradient (e.g.,
due to the pressures introduced during injection for hydraulic fracturing), particularly for brine
that is denser than the groundwater in the overlying formations (Flewelling and Sharma. 2014). In
the presence of natural gas, buoyancy of the less dense gas could potentially provide an upward flux
(Vengosh etal.. 2014). However, Flewelling and Sharma (2014) indicated that pressure
1 Relative permeability is a dimensionless property allowing for the comparison of the different abilities of fluids to flow
in multiphase settings. If a single fluid is present, its relative permeability is equal to 1, but the presence of multiple fluids
generally inhibits flow and decreases the relative permeability fSchlumberger. 20141
2 Shutting in the well after fracturing allows fluids to move farther into the formation, resulting in a higher gas relative
permeability near the fracture surface and improved gas production fBertoncello etal.. 2014).
3 Imbibition is the displacement of a nonwetting fluid (i.e., gas] by a wetting fluid (typically water]. The terms wetting or
nonwetting refer to the preferential attraction of a fluid to the surface. In typical reservoirs, water preferentially wets the
surface, and gas is nonwetting. Capillary forces arise from the differential attraction between immiscible fluids and solid
surfaces; these are the forces responsible for capillary rise in small-diameter tubes and porous materials. These
definitions are adapted from Pake f19781
4 The contrast in water activity between brine and fresh water generates very substantial osmotic pressure differences
that will drive fluids into the shale matrix. The osmosis process requires a semi-permeable membrane and a
concentration gradient to allow the solvent to pass through it. The clay in the shale formation can provide a function
similar to a membrane ("Zhou. 20161
6-51
-------
Chapter 6 - Well Injection
perturbations due to hydraulic fracturing operations are localized to the immediate vicinity of the
fractures, due to the very low permeabilities of shale formations; this means that hydraulic
fracturing operations are unlikely to generate sufficient pressure to drive fluids into shallow
drinking water zones. Some natural conditions could also create an upward hydraulic gradient in
the absence of any effects from hydraulic fracturing. However, these natural mechanisms have been
found to cause very low flow rates over very long distances, yielding extremely small vertical fluxes
in sedimentary basins. These translate to some estimated travel times of 100,000 to 100,000,000
years across a 328 ft (100 m) thick layer with about 0.01 nD (1 x 10-23 m2) permeability fFlewelling
and Sharma. 2014). In an area of the Permian Basin with over-pressured source rocks, Engle etal.
(2016) concluded that chemical, isotopic, and pressure data suggest that there is little potential for
vertical fluid migration to shallow zones in the absence of pathways such as improperly abandoned
wells (Section 6.3.2.3).
To account for the combined effect of capillary imbibition, well operation, and buoyancy in upward
fluid migration, Birdsell etal. (2015a) conducted a numerical analysis over five phases of activity at
a hypothetical Marcellus-like hydraulic fracturing site: a pre-drilling steady state, the injection of
fluids, a shut-in period, production, and the continued migration of hydraulic fracturing fluids after
the well is plugged and abandoned. They quantified how much hydraulic fracturing fluid flows back
up the well after fracturing, how much reaches overlying aquifers, and how much is permanently
sequestered by capillary imbibition (which is treated as a sink term). Their results affirmed that,
without a pathway such as a permeable fault or leaky wellbore, it is very unlikely that hydraulic
fracturing fluid from a deep shale could reach an overlying aquifer. However, the study did indicate
that upward migration on the order of 328 ft (100 m) could occur through relatively low-
permeability overburden, even if no discrete, permeable pathway exists.
6.3.2.2 Fracture Overgrowth out of the Production Zone
Fractures extending out of the intended production zone into another formation, or into an
unintended zone within the same formation, could provide a potential fluid migration pathway into
drinking water resources (Tackson etal.. 2013d). This migration could occur either through the
fractures themselves or in connection with other permeable subsurface features or formations
(Figure 6-7). Such "out-of-zone fracturing" is undesirable from a production standpoint and may
occur as a result of inadequate reservoir characterization or fracture treatment design (Eisner etal..
2006). Some researchers have noted that fractures growing out of the targeted production zone
could potentially contact other formations, such as higher conductivity sandstones or conventional
hydrocarbon reservoirs, which may create an additional pathway for migration into a drinking
water resource f Reagan etal.. 20151. In addition, fractures growing out of the production zone
could potentially intercept natural, preexisting fractures (discussed in Section 6.3.2.4) or active or
abandoned wells near the well where hydraulic fracturing is performed (discussed in Section
6.3.2.3).
6-52
-------
Chapter 6 - Well Injection
Water Well
Production Well
Note: Figure not to scale
Figure 6-7. Conceptualized depiction of potential pathways for fluid movement out of the
production zone: (a) induced fracture overgrowth into over- or underlying formations; (b)
induced fractures intersecting natural fractures; and (c) induced fractures intersecting a
permeable fault.
The fracture's geometry (Section 6.3.1] affects its potential to extend beyond the intended zone and
serve as a pathway to drinking water resources. Vertical heights of fractures created during
hydraulic fracturing operations have been measured in several U.S. shale plays, including the
Barnett, Woodford, Marcellus, and Eagle Ford, using microseismic monitoring and tiltmeters
(Fisher and Warpinski. 2012). These data indicate typical fracture heights extending from tens to
hundreds of feet1Davies etal. (2012) analyzed this data set and found that the maximum fracture
height was 1,929 ft (588 m) and that 1% of the fractures had a height greater than 1,148 ft (350 m).
This may raise some questions about fractures being contained within the producing formation, as
some Marcellus fractures were found to extend vertically for at least 1,500 ft (460 m), while the
maximum thickness of the formation is generally 350 ft (110 m) or less (MCOR. 2012). However,
the majority of fractures within the Marcellus were found to have heights less than 328 ft (100 m),
suggesting limited possibilities for fracture overgrowth exceeding the separation between shale
reservoirs and shallow aquifers (Davies etal.. 2012). This is consistent with modeling results found
by Kim and Moridis (2015) and others, as described below. Where the producing formation is not
1 As described in Section 6.3.1, microseismic data represent the small amounts of seismic energy generated during
subsurface fracturing. The Fisher and Warpinski dataset includes the top and bottom depths of mapped fracture
treatments in the four shale plays mentioned, giving the maximum propagation length.
6-53
-------
Chapter 6 - Well Injection
continuous horizontally, the lateral extent of fractures may also become important For example, in
the Fisher and Warpinski T20121 data set, fractures were found to extend to horizontal lengths
greater than 1,000 ft (300 m).
Results of National Energy Technology Laboratory (NETL) research in Greene County,
Pennsylvania, are generally consistent with those reported in the Fisher and Warpinski T20121 data
set Microseismic monitoring was used at six horizontal Marcellus Shale wells to identify the
maximum upward extent of brittle deformation (i.e., rock breakage) caused by hydraulic fracturing
(Hammack etal.. 2014). At three of the six wells, fractures extending between 1,000 and 1,900 ft
(300 and 580 m) above the Marcellus Shale were identified. Overall, approximately 40% of the
microseismic events occurred above the Tully Limestone, the formation overlying the Marcellus
Shale. The microseismic data suggest that fracture propagation occurs above the Tully Limestone,
which is sometimes referred to as an upper barrier to hydraulic fracture growth fHammack etal..
2014). However, all microseismic events were at least 5,000 ft (2,000 m) below drinking water
aquifers, as the Marcellus Shale is one of the deepest shale plays (Table 6-3), and no impacts to
drinking water resources or another local gas-producing interval were identified. See Text Box 6-6
for more information on the Greene County site.
Text Box 6-6. Monitoring at the Greene County, Pennsylvania, Hydraulic Fracturing Test Site.
Monitoring performed at the Marcellus Shale test site in Greene County, Pennsylvania, evaluated fracture
height growth and zonal isolation during and after hydraulic fracturing operations fHammack et al.. 2014).
The site has six horizontally drilled wells and two vertical wells that were completed into the Marcellus Shale.
Pre-fracturing studies of the site included a 3D seismic survey to identify faults, pressure measurements, and
baseline sampling for isotopes; drilling logs were also run. Hydraulic fracturing occurred April 24 to May 6,
2012, and June 4 to 11, 2012. Monitoring at the site included the following:
• Microseismic monitoring was conducted during four of the six hydraulic fracturing jobs on the site,
using geophones placed in the two vertical Marcellus Shale wells. These data were used to monitor
fracture height growth above the six horizontal Marcellus Shale wells during hydraulic fracturing.
• Pressure and production data were collected from a set of existing vertical gas wells completed in
Upper Devonian/Lower Mississippian zones 3,800 to 6,100 ft (1,200 to 1,900 m) above the Marcellus.
Data were collected during and after the hydraulic fracturing jobs and used to identify any
communication between the fractured areas and the Upper Devonian/Lower Mississippian rocks.
• Chemical and isotopic analyses were conducted on gas and water produced from the Upper
Devonian/Lower Mississippian wells. Samples were analyzed for stable isotope signatures of hydrogen,
carbon, and strontium and for the presence of perfluorocarbon tracers used in 10 stages of one of the
hydraulic fracturing jobs to identify possible gas or fluid migration to overlying zones fSharma et al..
2014a: Sharma et al.. 2014b).
As of September 2014, no evidence was found of gas or brine migration from the Marcellus Shale fHammack
et al.. 20141. although longer-term monitoring is necessary to confirm that no impacts to overlying zones
have occurred fZhang et al.. 2014al.
Similarly, in Dunn County, North Dakota, there is evidence suggestive of out-of-zone fracturing in
the Bakken Shale fU.S. EPA. 2015il. At the Killdeer site (Section 6.2.2.1), hydraulic fracturing fluids
6-54
-------
Chapter 6 - Well Injection
and produced water were released during a rupture of the casing at the Franchuk 44-20 SWH well.
Water quality characteristics at two monitoring wells located immediately downgradient of the
Franchuk well reflected a mixing of local Kill deer Aquifer water with deep formation brine. Ion and
isotope ratios used for brine fingerprinting suggest that Madison Group formations (which directly
overlie the Bakken in the Williston Basin) were the source of the brine observed in the Killdeer
Aquifer, and the authors concluded that this provides evidence for out-of-zone fracturing (U.S. EPA,
2015i). Industry experience also indicates that out-of-zone fracturing could be fairly common in the
Bakken and that produced water from many Bakken wells has Madison Group chemical signatures
("Arkadakskiy and Rostron, 2013; Arkadakskiy and Rostron, 2012; Peterman etal., 2012],
Fracture growth from a deep formation to a near-surface aquifer is generally considered to be
limited by layered geological environments and other physical constraints ("Fisher and Warpinski,
2012; Daneshy, 2009"). For example, differences in in-situ stresses in layers above and below the
production zone can restrict fracture height growth in sedimentary basins ("Fisher and Warpinski,
2012], High-permeability layers near hydrocarbon-producing zones can reduce fracture growth by
acting as a "thief zone" into which fluids can migrate, or by inducing a large compressive stress that
acts on the fracture fde Pater and Dong, 2009, as cited in Fisher and Warpinski, 2012"). Although
thief zones may prevent fractures from reaching shallower formations or growing to extreme
vertical lengths, they do allow fluids to migrate out of the production zone into receiving
formations, which could (depending on site-specific conditions) potentially contain drinking water
resources. A volumetric argument has also been used to discuss limits of vertical fracture growth;
that is, the volumes of fluid needed to sustain fracture growth beyond a certain height would be
unrealistic (Fisher and Warpinski, 2012). However, as described in Section 6.3.1, fracture volume
can be greater than the volume of injected fluid due to the effects of pressurized water combined
with the effects of gas during injection (Kim and Moridis, 2015). Nevertheless, some numerical
investigations suggest that, unless unrealistically high pressures and injection rates are applied to
an extremely weak and homogeneous formation that extends up to the near surface, hydraulic
fracturing generally induces stable and finite fracture growth in a Marcellus-type environment and
fractures are unlikely to extend into drinking water resources (Kim and Moridis, 2015).
Modeling studies have identified other factors that can affect the containment of fractures within
the producing formation. As discussed above, additional numerical analysis of fracture propagation
during hydraulic fracturing has demonstrated that contrasts in the geomechanical properties of
rock formations can affect fracture height containment (Gu and Siebrits, 2008) and that geological
layers present within shale gas reservoirs can limit vertical fracture propagation (Kim and Moridis,
2015). In another modeling study, Myshakin etal. (2015) applied a multi-layered geologic model to
study whether fracture growth can extend upward through overlying strata and reach drinking
water resources in a Marcellus Shale-type environment Most fractures were predicted either to
extend upward to the overlying layer (about 46%) or to remain in the Marcellus Shale (about 34%).
About 20% of the fractures were predicted to extend further upward into or above the overlying
limestone. These model results are consistent with microseismic events observed above the Tully
Limestone in Greene County, Pennsylvania (Hammack et al., 2014), where the fracture heights
ranged from 0 to 700 ft (0 to 200 m) and most of the fractures terminated less than 100 ft (31 m)
above the top of the Marcellus.
6-55
-------
Chapter 6 - Well Injection
If fractures were to propagate from the production zone to drinking water resources, other factors
would need to be in place for fluid migration to occur. Using a numerical simulation, Reagan et al.
f20151 investigated potential short-term migration of gas and water between a shale or tight gas
formation and a shallower groundwater unit, assuming that a permeable pathway already exists
between the two formations. Note that, for the purposes of this study, the pathway was assumed to
be pre-existing, and Reagan etal. (20151 did not model the hydraulic fracturing process itself.
The subsurface system evaluated in the Reagan etal. f20151 modeling investigation included a
horizontal well used for hydraulic fracturing and gas production, a connecting pathway between
the producing formation and the aquifer, and a shallow vertical water well in the aquifer (Figure
6-7). The parameters and scenarios used in the study are shown in Table 6-4; two vertical
separation distances between the producing formation and the aquifer were investigated, along
with a range of production zone permeabilities and other variables used to describe four
production scenarios. The horizontal well was assigned a constant bottomhole pressure of half the
initial pressure of the target reservoir, not accounting for any over-pressurization from hydraulic
fracturing. (As noted in Section 6.3.2.1, over-pressurization during hydraulic fracturing can create
an additional driving force for upward migration.) In the simulation, migration was assessed
immediately after hydraulic fracturing and for up to a 2-year simulation period representing the
production stage.
Results of this modeling investigation indicate a generally downward water flow within the
connecting fracture (i.e., flow from the aquifer through the connecting fracture into the
hydraulically induced fractures in the production zone) with some upward migration of gas
(Reagan etal.. 2015). In certain simulated cases, gas breakthrough (the appearance of gas at the
base of the drinking water aquifer) was also observed. The key parameter affecting migration of gas
into the aquifer was the production regime, particularly whether gas production (which drives
migration toward the production well) was occurring in the reservoir. Simulations that included a
producing gas well showed only a few instances of breakthrough, while simulations without gas
production (i.e., that assumed the well was shut-in) tended to result in breakthrough. When gas
breakthrough did occur, the breakthrough times ranged from minutes to 20 days. However, in all
cases, the gas escape was limited in duration and scope, because the amount of gas available for
immediate migration toward the shallow aquifer was limited to that initially stored in the
hydraulically induced fractures after the stimulation process and prior to production. These
simulations indicate that the target reservoir may not be able to replenish the gas that was
available for migration prior to production.
Based on the results of the Reagan etal. (2015) modeling study, gas production from the reservoir
appears likely to mitigate gas migration, both by reducing the amount of available gas and
depressurizing the induced fractures (which counters the buoyancy of any gas that may escape
from the production zone into the connecting fracture). Production at the gas well also creates
pressure gradients that drive a downward flow of water from the aquifer via the fracture into the
producing formation, increasing the amount of water produced at the gas well. Furthermore, the
effective permeability of the connecting feature is reduced during water (downward) and gas
(upward) counter-flow within the fracture, further retarding the upward movement of gas or
6-56
-------
Chapter 6 - Well Injection
allowing gas to dissolve into the downward flow. However, Reagan etal. (20151 did find an
increased potential for gas release from the producing formation in cases where there is no gas
production following hydraulic fracturing. The potential for gas migration during shut-in periods
following hydraulic fracturing and prior to production may be more significant, especially when
out-of-zone fractures are formed. Without the effects of production, gas can rise via buoyancy, with
any downward-flowing water from the aquifer displacing the upward-flowing gas.
Reagan etal. f20151 also found that the permeability of a connecting fault or fracture may be an
important factor affecting the potential upward migration of gas (although not as significant as the
production regime). For the cases where gas escaped from the production zone, the maximum
volume of migrating gas depended upon the permeability of the connecting feature: the higher the
permeability, the larger the volume. The modeling results also showed that lower permeabilities
delay the downward flow of water from the aquifer, allowing the trace amount of gas that entered
into the fracture early in the modeled period to reach the aquifer, which was otherwise predicted to
dissolve in the water flowing downward in the feature. Similarly, the permeabilities of the target
reservoir, fracture volume, and the separation distance were found to affect gas migration, because
they affected the initial amount of gas stored in the hydraulically induced fractures. In contrast, the
permeability of the drinking water aquifer was not found to be a significant factor in the
assessment.
Table 6-4. Modeling parameters and scenarios investigated by Reagan et al. (2015).
This table illustrates the range of parameters included in the Reagan et al. (2015) modeling study. See Figure 6-7,
Figure 6-8, and Figure 6-9 for conceptualized illustrations of these scenarios.
Model parameter or variable
Values investigated in model scenarios
All scenarios
Lateral distance from connecting feature to water well
328 ft (100 m)
Vertical separation distance between producing
formation and drinking water aquifer
656 ft (200 m);
2,625 ft (800 m)
Producing formation permeability range
1 nD (lx 10"21 m2);
100 nD (lx 1019 m2);
1 HD (1 x 1018 m2)
Drinking water aquifer permeability
0.1 D (lx 10 13 m2);
1 D (1 x 1012 m2)
Initial conditions
Hydrostatic
Production well bottom hole pressure
Half of the initial pressure of the producing formation
(not accounting for over-pressurization
from hydraulic fracturing)
Production regime
Production at both the water well and the gas well;
Production at only the water well;
Production at only the gas well;
No production
6-57
-------
Chapter 6 - Well Injection
Model parameter or variable
Values investigated in model scenarios
Fracture pathway scenarios
Connecting feature permeability
1 D (1 x 1012 m2);
10 D (1 x 10 11 m2);
1,000 D (1 x 10 9 m2)
Offset well pathway scenarios
Lateral distance from production well to offset well
33 ft (10 m)
Cement permeability of offset well
1 HD (1 x 10 18 m2);
lmDjlx 1015 m2);
1 D (1 x 1012 m2);
1,000 D (1 x 10 9 m2)
6.3.2.3 Migration via Fractures Intersecting with Offset Wells and Other Artificial Structures
Another potential pathway for fluid migration is one in which hydraulic fracturing fluids or
displaced subsurface fluids move through newly created fractures into an offset well or its fracture
network, resulting in a process called well communication flackson et al.. 2013dl. The offset well
can be an abandoned (i.e., plugged), inactive, or actively producing well. In addition, if the offset
well has also been used for hydraulic fracturing, the fracture networks of the two wells might
intersect. The situation where hydraulic fractures propagate to (and inject fluid into and/or cause
pressure increases in) other existing wells or hydraulic fractures is referred to as a "frac hit" and is
known to occur in areas with a high density of wells (Tackson etal.. 2013a).
Frac hits can be more common in unconventional production settings compared to conventional
production settings, due to the closer/denser well spacing fKing and Valencia. 20161. Figure 6-8
provides a schematic to illustrate fractures that intercept an offset well, and Figure 6-9 depicts (in a
simplified illustration) how the fracture networks of two such wells might intersect. This can be a
particular concern in shallower formations, where the local least principal stress is vertical
(resulting in more horizontal fracture propagation), and in situations where there are drinking
water wells in the same formation as wells used for hydraulic fracturing.
Instances of well communication have been known to occur and are described in well records and
the oil and gas literature. For example, an analysis of operator data collected by the New Mexico Oil
Conservation Division (NM OCD) in 2013-2014 identified 120 instances of well communication in
the San Juan Basin between 2007 and 2013 fVaidyanathan. 20141. In some cases, well
communication incidents have led to documented production and/or environmental problems. A
study in the Barnett Shale noted two cases of well communication, one with a well 1,100 ft (340 m)
away and the other with a well 2,500 ft (760 m) away from the initiating well; ultimately, one of the
offset wells had to be re-fractured because the well communication halted production (Craig etal..
20121. In some cases, the fluids that intersect the offset well flow up the wellbore and spill onto the
surface. In its report Review of State and Industry Spill Data: Characterization of Hydraulic
Fracturing-Related Spills, the EPA (2015m) recorded 10 incidents in which fluid spills were
6-58
-------
Chapter 6 - Well Injection
attributed to well communication events (see Text Box 5-10 for more information on this effort).1
The Well File Review fU.S. EPA. 2016cl reports that 1% of the wells (an estimated 280 wells]
represented in the study reported a frac hit, where the hydraulic fracturing operation documented
in the Well File Review led to communication with a nearby well.2 (It was not possible to determine
whether fluids reached protected groundwater resources during these frac hits based on
information in the well files.) While the subsurface effects of frac hits have not been extensively
studied, these cases demonstrate the possibility of fluid migration via communication with other
wells and/or their fracture networks. More generally, well communication events can indicate
fracture behavior that was not intended by the treatment design.
Water Well
Offset Well
Production Well
Figure 6-8. Induced fractures intersecting an offset well (in a production zone, as shown, or in
overlying formations into which fracture growth may have occurred).
This image shows a conceptualized depiction of potential pathways for fluid movement out of the production
zone (not to scale).
1 These spills are represented by line numbers 163,236,265,271,286,287, 375, 376, 377, and 380 in Appendix B of U.S.
EPA f2015ml
2 280 wells (95% confidence interval: 240 - 320 wells).
6-59
-------
Chapter 6 - Well Injection
thousands of feet between wells
Figure 6-9. Well communication (a frac hit).
This image shows a conceptualized depiction of the fractures of a newly fractured well (Well A) intersecting the
existing fracture network created during a previous hydraulic fracturing operation in an offset well (Well B).
Evidence of this interaction may be observed in the offset well as a pressure change, lost production, and/or
introduction of new fluids. Depending on the condition of the offset well, this can result in fluid being spilled onto
the surface, rupturing of cement and/or casing and hydraulic fracturing fluid leaking into subsurface formations,
and/or fluid flowing out through existing flaws in the casing and/or cement, (Figure is not to scale.)
A well communication event is usually observed at the offset well as a pressure spike, due to the
elevated pressure from the originating well, or as an unexpected drop in the production rate (Lawal
et al.. 2014: Tackson et al.. 2013al. Aiani and Kelkar f20121 performed an analysis of frac hits in the
Woodford Shale in Oklahoma, studying 179 wells over a 5-year period. The authors used fracturing
records from the newly completed wells and compared them to production records from
surrounding wells. The authors assumed that sudden changes in production of gas or water
coinciding with fracturing at a nearby well were caused by communication between the two wells,
and increased water production at the surrounding wells was assumed to be caused by hydraulic
fracturing fluid flowing into these offset wells. The results of the Oklahoma study showed that
24 wells had decreased gas production or increased water production within 60 days of the initial
gas production at the nearby fractured well. A total of 38 wells experienced decreased gas or
increased water production up to a distance of 7,920 ft (2,410 m), which the study authors defined
as the distance between the midpoints of the laterals; 10 wells saw increased water production
6-60
-------
Chapter 6 - Well Injection
from as far away as 8,422 ft (2,567 m). In addition, one well showed a slight increase in gas
production rather than a decrease.1
Other studies of well communication events have relied on similar information. In the NM OCD
operator data set, the typical means of detecting a well communication event was through pressure
changes at the offset well, production lost at the offset well, and/or fluids found in the offset well. In
some instances, well operators determined that a well was producing fluid from two different
formations, while in one instance, the operator identified a potential well communication event due
to an increase in production from the offset well (Vaidvanathan. 2014). In another study, Tackson et
al. (2013a) found that the decrease in production due to well communication events was much
greater in lower permeability reservoirs. The authors note an example where two wells 1,000 ft
(300 m) apart communicated, reducing production in the offset well by 64%. These results indicate
that the subsurface interactions of well networks or complex hydraulics driven by each well at a
densely populated (with respect to wells) area are important factors to consider for the design of
hydraulic fracturing treatments and other aspects of oil and gas production.
The key factor affecting the likelihood of a well communication event and the impact of a frac hit is
the location of the offset well relative to the well where hydraulic fracturing was conducted (Ajani
and Kelkar. 2012). In the Aiani and Kelkar (2012) analysis, the likelihood of a communication event
was less than 10% in wells more than 4,000 ft (1,000 m) apart, but rose to nearly 50% in wells less
than 1,000 ft (300 m) apart. Well communication was also much more likely with wells drilled from
the same pad. The affected wells were found to be in the direction of maximum horizontal stress in
the field, which correlates with the expected direction of fracture propagation. Modeling work by
Mvshakin etal. (2015) is generally consistent with these results, indicating that the risk of fluid
movement through pre-existing wellbores or open faults is negligible unless hydraulic fractures are
located very close to these features.2
Statistical modeling by Montague and Pinder (2015) investigated the probability that a hypothetical
new well used for hydraulic fracturing within the area of New York underlain by the Marcellus
Shale would intersect an existing wellbore. The results indicated that this probability would be
from 0 to 3.45%. The model incorporated the depth of the hypothetical new well, the vertical
growth of induced fractures, and the depth and locations of existing nearby wells. The model also
assumed that the existing wells are vertical and fracture growth is not impacted by nearby wells or
existing fractures. However, the authors concluded that the inclusion of horizontal wells within the
data set could increase the chance of intersection with induced fractures.
Well communication may be more likely to occur where there is less resistance to fracture growth.
Such conditions may be related to existing production operations (e.g., where previous
hydrocarbon extraction has reduced the pore pressure, changed stress fields, or affected existing
fracture networks) or the existence of high-permeability rock units fIackson et al.. 2 013 al. As Aiani
and Kelkar (2012) found in the Woodford Shale, one of the deepest major shale plays (Table 6-3),
induced fractures tend to enter portions of the reservoir that have already been fractured as
1 The numbers of wells cited in the study reflect separate analyses, and the numbers cited are not additive.
2 In the Mvshakin etal. ("20151 paper, the authors do not quantify or explain what is meant by "very close."
6-61
-------
Chapter 6 - Well Injection
opposed to entering previously unfractured rocks, ultimately causing interference in offset wells.
Mukheriee etal. f20001 described this tendency for asymmetric fracture growth toward depleted
areas in low-permeability gas reservoirs due to pore pressure depletion from production at offset
wells. The authors note that pore pressure gradients in depleted zones would affect the subsurface
stresses. Therefore, depending on the location of the new well with respect to depleted zone(s) and
the orientation of the existing induced fractures, the newly created fracture can be asymmetric,
with only one wing of the fracture extending into the depleted area and developing significant
length and conductivity fMukheriee etal.. 20001. The extent to which the depleted area affects
fracturing depends on factors such as cumulative production, pore volume, hydrocarbon saturation,
effective permeability, and the original reservoir or pore pressure (Mukherjee etal.. 2000).
Similarly, high-permeability rock types acting as thief zones may also cause preferential fracturing
due to a higher leakoff rate into these layers flackson et al.. 2013al.
In addition to location, the potential for impact on a drinking water resource also depends on the
condition of the offset well. (See Section 6.2 for information on the mechanical integrity of well
components.) In their analysis, Aiani and Kelkar f20121 found a correlation between well
communication and well age: older wells were more likely to be affected. If the cement in the
annulus between the casing and the formation is intact and the well components can withstand the
stress exerted by the pressure of the fluid, nothing more than an increase in pressure and extra
production of fluids would occur during a well communication event. However, if the offset well is
not able to withstand the pressure of the hydraulic fracturing fluid, well components could fail
(Figure 6-4), allowing fluid to migrate out of the well.
The highest pressures most hydraulic fracturing wells will face during their life spans occur during
the process of fracturing (Section 3.3). In some cases, temporary equipment is installed in wells
during fracturing to protect the well against the increased pressure. Therefore, many producing
wells may not be designed to withstand pressures typical of hydraulic fracturing (Enform. 2013)
and can experience problems when fracturing occurs in nearby wells. Depending on the location of
the weakest point in the offset well, this could result in fluid being spilled onto the surface;
rupturing of cement and/or casing and hydraulic fracturing fluid leaking into subsurface
formations; and/or fluid flowing out through existing flaws in the casing and/or cement (See
Chapters 5 and 7 for additional information on how such spills can affect drinking water resources.)
For example, a documented well communication event near Innisfail, Alberta, Canada (Text Box
6-7) occurred when several well components failed, because they were not rated to handle the
increased pressure caused by the well communication (ERCB. 2012). In addition, if the fractures
were to intersect an uncemented portion of the wellbore, the fluids could potentially migrate into
formations that are uncemented along the wellbore.
In older wells near a hydraulic fracturing operation, plugs and cement can degrade over time; in
some cases, abandoned wells may never have been plugged properly. Before the 1950s, most well
plugging efforts were focused on preventing water from the surface from entering oil fields. As a
result, many wells from that period were abandoned with little or no cement (NPC. 2011b). This
can be a significant issue in areas with legacy (i.e., historic) oil and gas exploration and when wells
are re-entered and hydraulically fractured (or re-fractured) to increase production in a reservoir. In
6-62
-------
Chapter 6 - Well Injection
one study, 18 of 29 plugged and abandoned wells in Quebec were found to show signs of leakage
fCouncil of Canadian Academies. 20141. Similarly, a PA DEP report cited three cases where
migration of natural gas had been caused by well communication events with old, abandoned wells,
including one case where private drinking water wells were affected fPA DEP. 2009cl. In Tioga
County, Pennsylvania, following hydraulic fracturing of a shale gas well, an abandoned well nearby
produced a 30 ft (9 m) geyser of brine and gas for more than a week fDilmore etal.. 20151.
Text Box 6-7. Well Communication at a Horizontal Well near Innisfail, Alberta, Canada.
In most cases, well communication during fracturing results in a pressure surge accompanied by a drop in gas
production and additional flow of produced water or hydraulic fracturing fluid at an offset well. However, if
the offset well is not capable of withstanding the high pressures of fracturing, more significant damage can
occur.
In January 2012, fracturing at a horizontal well near Innisfail in Alberta, Canada, caused a surface spill of
fracturing and formation fluids at a nearby operating vertical oil well. According to the investigation report by
the Alberta Energy Resources Conservation Board (ERCB. 20121. pressure began rising at the vertical well
less than two hours after fracturing ended at the horizontal well.
Several components of the vertical well facility—including surface piping, discharge hoses, fuel gas lines, and
the pressure relief valve associated with compression at the well—were not rated to handle the increased
pressure and failed. Ultimately, the spill released, in addition to gas, an estimated 19,816 gal (75,012 L:i] of
hydraulic fracturing fluid, brine, and oil covering an area of approximately 656 ft by 738 ft (200 m by 225 m].
The ERCB determined that the lateral of the horizontal well passed within 423 ft (129 m] of the vertical well
at a depth of approximately 6,070 ft (1,850 m] below the surface in the same formation. The operating
company had estimated a fracture half-length of 262 to 295 ft (80 to 90 m] based on a general fracture model
for the field.1 While there were no regulatory requirements for spacing hydraulic fracturing operations in
place at the time, the 423 ft (129 m] distance was out of compliance with the company's internal policy to
space fractures from adjacent wells at least 1.5 times the predicted half-length. The company also did not
notify the operators of the vertical well of the hydraulic fracturing operations. The incident prompted the
ERCB to issue Bulletin 2012-02—Hydraulic Fracturing: Interwellbore Communication between Energy Wells,
which outlines expectations for avoiding well communication events and preventing adverse effects on offset
wells.
Various studies estimate the number of abandoned wells in the United States to be significant The
Interstate Oil and Gas Compact Commission flOGCC. 20081 estimates that over one million wells
were drilled in the United States prior to the enactment of state oil and gas regulations, and the
status and location of many of these wells are unknown. A recent estimate of wells completed
before the adoption of statewide well abandonment criteria in 1957 in Pennsylvania placed the
range at 305,000 to 390,000 wells in the state, with more than 176,000 of those wells likely
abandoned pre-1957 fDilmore et al.. 20151. As of 2000, PA DEP's well plugging program reported
that it had documented 44,700 wells that had been plugged and 8,000 that were in need of
plugging, and approximately 184,000 additional wells with an unknown location and status (PA
1 The fracture half-length is the radial distance from a wellbore to the outer tip of a fracture propagated from that well
fSchlumberger. 20141
6-63
-------
Chapter 6 - Well Injection
DEP. 20001. A similar evaluation from New York State found that the number of unplugged wells
was growing in the state despite an active well plugging program fBishop. 20131. In the Midwest,
Sminchak et al. f20141 examined two areas of historical oil and gas exploration as part of an
investigation of potential carbon dioxide sequestration sites. They found that a 4.3 mi by 4.3 mi (6.9
km by 6.9 km) square area in Michigan contained 22 abandoned oil and gas wells and a 9.3 mi by
9.3 mi (15.0 km by 15.0 km) square area in Ohio contained 359 abandoned oil and gas wells.
Various state programs exist to plug identified orphaned wells, but they face the challenge of
identifying and addressing a large number of wells.1 In some cases, remote sensing technologies can
be used to identify wells for which no records exist. For example, an NETL study in Pennsylvania
found that helicopter-based high-resolution magnetic surveys can be used to accurately locate wells
with steel casing; wells with no steel casing exhibit weak or no magnetic anomaly and are not
detected by such surveys fVeloski et al.. 20151. Chapter 10 includes a discussion of factors and
practices, including those related to active and abandoned wells near hydraulic fracturing
operations, that can reduce the frequency of impacts to drinking water quality.
The Reagan etal. f20151 numerical modeling study included an assessment of migration via an
offset well as part of its investigation of potential fluid migration from a producing formation into a
shallower groundwater unit (Section 6.3.2.2). For the offset well pathway, it was assumed that the
hydraulically induced fractures intercepted an older offset well with deteriorated components.
(This assessment can also be applicable to cases where potential migration may occur via the
production well-related pathways discussed in Section 6.2) The highest permeability value tested
for the connecting feature represented a case with an open wellbore. A key assumption for this
investigation was that the offset well was already directly connected to a permeable feature in the
reservoir or within the overburden.
Similar to the cases for permeable faults or fractures discussed in Section 6.3.2.2, the study
investigated the effect of multiple well- and formation-related variables on potential fluid migration
(Table 6-4). Based on the simulation results, an offset well pathway can have a greater potential for
gas release from the production zone into a shallower groundwater unit than the fracture pathway
discussed in Section 6.3.2.2 fReagan etal.. 20151. This difference is primarily due to the total pore
volume of the connecting pathway within the offset well; if the offset well pathway has a
significantly lower pore volume compared to the fracture pathway, this would reduce possible gas
storage in the connecting feature and increase the speed of buoyancy-dependent migration.
However, as with the fracture scenario, the gas available for migration in this case is still limited to
the gas that is initially stored in the hydraulically induced fractures. Accordingly, any incidents of
gas breakthrough in the model results were limited in both duration and magnitude.
In their modeling study, Reagan etal. f20151 found that production at the gas well (the well used
for hydraulic fracturing) also affects the potential upward migration of gas and its arrival times at
the drinking water formation due to its effect on the driving forces (e.g., pressure gradient). Similar
to the fracture cases described in Section 6.3.2.2, production in the target reservoir appears to
mitigate upward gas migration, both by reducing the amount of gas that might otherwise be
1 An orphaned well is an inactive oil or gas well with no known (or financially solvent] owner.
6-64
-------
Chapter 6 - Well Injection
available for upward migration and by creating a pressure gradient toward the production well.
Only scenarios without the mitigating feature of gas production result in upward migration into the
aquifer. This assessment also found a generally downward water flow within the connecting well
pathway, which is more pronounced when the production well is operating and there is de-
pressurization within the fractures. The producing formation and aquifer permeabilities appear not
to be significant factors for upward gas migration via this pathway. Instead, Reagan etal. (2015)
found the permeability of the connecting well to be the key factor affecting the migration of gas to
the aquifer and the water well. Very low permeabilities (less than 1 mD, or 1 x 10"15 m2) for the
connecting well lead to no migration of gas into the aquifer regardless of the vertical separation
distance, whereas larger permeabilities presented a greater potential for gas breakthrough.
Brownlow et al. f20161 also modeled communication with an abandoned well. The modeling
exercise was based on operator data from the Eagle Ford Shale. Two types of cases were modeled:
cases with an open (unplugged) abandoned well (which the authors note are known to occur in
Texas) and cases with an abandoned well that was converted into a water well after the lower
portion of the well had been filled with drilling mud (a practice allowed in Texas until 1967). The
modeling results indicated that fluid could potentially migrate up both types of abandoned wells,
with relatively greater flow rates in open abandoned wells and in abandoned wells closer to the
well used for hydraulic fracturing. Similar to the Reagan etal. (2015) study, the production regime
was also a key factor; when production and flowback were included in the simulation, they were
found to inhibit upward migration. Modeled flow rates through the mud-filled well were
comparable to those found by Reagan etal. f20151 with higher flows predicted through the open
well.
A similar study was conducted by Nowamooz etal. (2015). who modeled a hypothetical well in the
Utica Shale in Quebec. They assumed a 7.9 in (200 mm) wellbore with an approximately 2 in (51
mm) annulus space filled with intact cement. The researchers varied the permeability of the cement
from 1 [iD (1 x 10"19 m2) to 1 mD (1 x 10-15 m2). The results indicated that, atthe highest
permeability of 1 mD, a flow of methane of 1.02 x 10 2 ft3/day (2.9 x 10 4 m3/d) was possible. This
was two orders of magnitude higher than the flow when the cement permeability was 1 [J.D
(1 x 10"19 m2). The wellbore permeabilities used by Nowamooz etal. f20151 appear to be consistent
with actual permeabilities observed in the field, which can vary widely. For example, a study of 31
abandoned oil and gas wells in Pennsylvania found effective permeability values along the
wellbores in the range of 10~6 to 102 mD (1 x 10~21 to 1 x 10~13 m2) fKangetal.. 20151.
In the same way that fractures can propagate to intersect offset wells, they can also potentially
intersect other artificial subsurface structures including mine shafts or solution mining sites. No
known incidents of this type of migration have been documented. However, the Bureau of Land
Management (BLM) has identified over 48,000 abandoned mines in the United States and is adding
new mines to its inventory every year fBLM. 20151. In addition, the Well File Review identified an
estimated 800 cases where wells used for hydraulic fracturing were drilled through mining voids,
and an additional 90 cases of drilling through gas storage zones or wastewater disposal zones (U.S.
6-65
-------
Chapter 6 - Well Injection
EPA. 2015n).1'2 The analysis suggests emplacing cement within such zones can be challenging,
which, in turn, could lead to a loss of zonal isolation (as described in Section 6.2) and create a
pathway for fluid migration.
6.3.2.4 Migration via Fractures Intersecting Geologic Features
Potential fluid migration via natural, permeable fault or fracture zones in conjunction with
hydraulic fracturing has been recognized as a potential contamination hazard for several decades
("Harrison, 1983). Natural fracture systems have a strong influence on the success of a fracture
treatment, and the topic has been studied extensively from the perspective of optimizing treatment
design (e.g., Dahi Taleghani and Olson. 2011; Wengetal., 2011; Vulgamore etal., 2007). While
porous flow in unfractured shale or tight sand formations is assumed to be negligible due to very
low formation permeabilities (as discussed in Section 6.3.2.1), the presence of small natural
fractures known as "microfractures" within tight sand or shale formations is widely recognized, and
these fractures affect fluid flow and production strategies. Naturally occurring permeable faults
and larger-scale fractures within or between formations can potentially allow for more significant
flow pathways out of the production zone (lackson etal., 2013d). Figure 6-7 illustrates the concept
of induced fractures intersecting with permeable faults or fractures extending out of the target
reservoir.3
The specific effects of natural fractures on fluid migration, and the mechanisms by which these
effects occur, are not completely understood. While naturally occurring microfractures can impact
the growth of induced fractures (e.g., by affecting the tensile strength of a shale layer), studies
based on modeling and monitoring data generally do not indicate that they contribute to fracture
growth in a way that could affect the frequency or severity of impacts. Microfractures could affect
fluid flow patterns near the induced fractures by increasing the effective contact area. Conversely,
these microfractures could act as capillary traps for the hydraulic fracturing fluid during treatment
(contributing to fluid leakoff) and potentially hinder hydrocarbon flow due to lower gas relative
permeabilities (Dahi Taleghani etal.. 2013). Ryan etal. (2015) suggested that some natural fracture
processes/patterns (such as the presence of two subvertical fracture sets) can contribute to
upward gas migration, while others (such as small fracture sets with low connectivity that are
confined to individual geologic layers) can preclude it
In some areas, larger-scale geologic features may affect potential fluid flow pathways. As discussed
in Text Box 6-3, baseline measurements taken before shale gas development show evidence of
thermogenic methane in some shallow aquifers, suggesting that, in some cases, natural subsurface
pathways exist and might allow for naturally occurring migration of gas over geologic time
(Robertson et al.. 2012). There is also evidence demonstrating that gas undergoes mixing in
1800 wells (95% confidence interval: 10 - 1,900 wells].
2 90 wells (95% confidence interval: 50 - 100 wells].
3 Faults and fractures can exhibit a range of permeabilities. For example, permeable (also referred to as "transmissive" or
"conductive"] faults or fault segments have enough permeability to allow fluids to flow along or across them, while others
are relatively impermeable and can serve as barriers to flow. These differences in permeability are associated with
geologic conditions such as rock type, depth, and stress regime. Generally, when researchers refer to the potential for
migration via natural geologic features, it is assumed that these features are sufficiently permeable to serve as a pathway.
6-66
-------
Chapter 6 - Well Injection
subsurface pathways (Baldassare etal.. 2014: Molofskv etal.. 2013: Fountain and Tacobi. 20001.
Warner etal. f20121 compared recent sampling results to data published in the 1980s and found
geochemical evidence for migration of fluids through natural pathways between deep underlying
formations and shallow aquifers—pathways that the authors suggest could lead to contamination
from hydraulic fracturing activities. In northeastern Pennsylvania, there is evidence that brine from
deep saline formations has migrated into shallow aquifers over geologic time, preferentially
following certain geologic structures fLlewellvn. 20141. However, this depends on local geologic
characteristics and does not appear to happen in all locations; for example, in the Monongahela
River Basin in West Virginia, shallow groundwater samples did not show evidence of mixing with
deep brines (Boothrovd etal.. 2016). As described in Chapter 7, karst features (created by the
dissolution of soluble rock) can also serve as a potential pathway of fluid movement on a faster time
scale.
Monitoring data show that the presence of natural faults and fractures can affect both the height
and width of induced hydraulic fractures. When faults are present, relatively larger microseismic
responses are seen and larger fracture growth can occur, as described below. Rutledge and Phillips
f20031 suggested that, for a hydraulic fracturing operation in East Texas, pressurizing existing
fractures (rather than creating new hydraulic fractures) was the primary process that controlled
enhanced permeability and fracture network conductivity at the site. Salehi and Ciezobka (2013)
used microseismic data to investigate the effects of natural fractures in the Marcellus Shale and
concluded that fracture treatments are more efficient in areas with clusters or "swarms" of small
natural fractures, while areas without these fracture swarms require more thorough stimulation.
These microseismic data show that swarms of natural fractures within a shale formation can result
in a fracture network with a larger width-to-height ratio (i.e., a shorter and wider network) than
would be expected in a zone with a low degree of natural fracturing.
A few studies have used monitoring data to specifically investigate the effect of natural faults and
fractures on the vertical extent of induced fractures. A statistical analysis of microseismic data by
Shapiro etal. (2011) found that fault rupture (movement along a fault) from hydraulic fracturing is
limited by the extent of the stimulated rock volume and is unlikely to extend beyond the fracture
network. However, as demonstrated by microseismic data presented by Vulgamore etal. f20071. in
some settings, the fracture network—and, in this case, the possibility of fault rupture—could
extend laterally for thousands of feet In the Fisher and Warpinski f20121 data set (Section 6.3.2.2),
the greatest fracture heights occurred when the hydraulic fractures intersected pre-existing faults.
Similarly, Hammacketal. (2014) reported that fracture growth seen above the Marcellus Shale is
consistent with the inferred extent of pre-existing faults at the Greene County, Pennsylvania,
research site (Section 6.3.2.2 and Text Box 6-6). The authors suggested that clusters of
microseismic events may have occurred where preexisting small faults or natural fractures were
present above the Marcellus Shale. Vinal (2015) used time-lapse multi-component seismic
monitoring to monitor the overburden of the Montney Shale during a hydraulic fracturing
operation in Alberta, Canada. The researchers found increases in the anisotropy in the overburden,
which they interpreted as fractures being propagated along natural faults out of the shale and into
the overburden. At a site in Ohio, Skoumal etal. (2015) found that hydraulic fracturing induced a
rupture along a pre-existing fault approximately 0.6 mi (1 km) from the hydraulic fracturing
6-67
-------
Chapter 6 - Well Injection
operation. Using a new monitoring method known as tomographic fracturing imaging, Lacazette
and Geiser ("20131 also found vertical hydraulic fracturing fluid movement from a production well
into a natural fracture network for distances of up to 0.6 mi (1.0 km). However, Davies etal. ("2013)
questioned whether this technique actually measures hydraulic fracturing fluid movement.
Modeling studies have also investigated whether hydraulic fracturing operations are likely to
reactivate faults and create a potential fluid migration pathway into shallow aquifers. Results from
one study suggest that, under specific circumstances, interaction with a permeable fault could
result in fluid migration to the surface but only on relatively long (ca. 1,000 year) time scales
("Gassiatetal., 2013). These findings have been disputed in the literature due to certain suggested
limitations of the study, including the model setup, assumptions, and calibration; unrealistic fault
representation; lack of constraints on fluid overpressure; and exclusion of the capillary imbibition
effect ("Birdsell etal., 2015b; Flewelling and Sharma, 2015). In response to these critiques, the
authors stated that their work was a parametric study in which the model geometry, parameter,
and boundary conditions were defined based on data collected from multiple shale gas basins, and
the objective of the study was not to calibrate results to a specific site ("Lefebvre etal., 2015). Other
researchers reject the notion that open, permeable faults coexist with hydrocarbon accumulation
(Fie welling et al., 2013). However, it is unclear whether the existence of faults in low permeability
reservoirs affects the accumulation of hydrocarbons because, under natural conditions, the flow of
gas may be limited due to capillary tension.
Like the other pathways discussed in this section, other conditions in addition to the physical
presence of a permeable fault or fracture would need to exist for fluid migration to a drinking water
resource to occur. The modeling study conducted by Reagan etal. ("2015) and discussed in Section
6.3.2.2 indicates that, if such a permeable feature exists, the transport of gas and fluid flow would
strongly depend upon the production regime and, to a lesser degree, the features' permeability and
the separation between the reservoir and the aquifer. In addition, the pressure distribution within
the reservoir (e.g., over-pressurized vs. hydrostatic conditions) will affect the fluid flow through
fractures/faults. As a result, the presence of multiple geologic and well-related factors can increase
the potential for fluid migration into drinking water resources. For example, in the Mamm Creek
area of Colorado (Section 6.2.2.4), mechanical integrity and drilling-related problems likely acted in
concert with natural fracture systems to result in a gas seep into surface water and shallow
groundwater (Crescent, 2011). A similar situation occurred in southeastern Bradford County,
Pennsylvania (discussed in Section 6.2), where natural fractures intersected an uncemented casing
annulus and allowed gas to flow from the annulus into nearby domestic wells and a stream
(Llewellyn etal., 2015).
Other modeling studies investigating the potential of fluid migration related to existing faults and
fractures have given mixed results. Pfuntetal. (2016) performed modeling based on conditions in
the North German Basin, i.e., deep geological settings where undisturbed cap rocks are present
between the fractured formation and shallow aquifers. Their modeling indicated that the hydraulic
fracturing fluid did not reach the near-surface area either during hydraulic fracturing operations or
in the long-term in the presence of highly permeable pathways (fault zones, fractures). Like
6-68
-------
Chapter 6 - Well Injection
previous modeling studies, the authors found that the injection pressure and permeability of the
connecting fault are two important factors that control upward fluid migration.
Rutqvist et al. (2013) found that, while somewhat larger microseismic events are possible in the
presence of faults, repeated events and a seismic slip would amount to a total rupture length of
164 ft (50 m) or less along a fault, not far enough to allow fluid migration between a deep gas
reservoir (approximately 6,562 ft or 2,000 m deep) and a shallow aquifer. A follow-up study using
more sophisticated three-dimensional modeling techniques also found that deep hydraulic
fracturing is unlikely to create a direct flow path into a shallow aquifer, even when hydraulic
fracturing fluid is injected directly into a fault (Rutqvist et al.. 20151. Similarly, a modeling study
that investigated potential fluid migration from hydraulic fracturing in Germany found potential
vertical fluid migration up to 164 ft (50 m) in a scenario with high fault zone permeability, although
the authors note this is likely an overestimate because their goal was to "assess an upper margin of
the risk" associated with fluid transport (Lange et al.. 2013). More generally, results from Rutqvist
etal. (2013) indicate that fracturing along an initially impermeable fault (as is expected in a shale
gas formation) would result in numerous small microseismic events that act to prevent larger
events from occurring (and, therefore, prevent the creation of more extensive potential pathways).
Schwartz (2015) modeled methane flow through a hypothetical permeable fault at a well in
Germany. Methane flow was modeled through a permeable leakage zone that was 0.1 ft by 13 ft
(0.03 m by 4 m) with an assumed permeability in the range of approximately 100 D to of 10,000 D
(1 x 1010 m2 to 1 x 10"8 m2). The model indicated that methane could reach a drinking water aquifer
approximately 2,953 ft (900 m) above the gas zone in about a half a day and reach a maximum flow
after two days. According to the model results, methane entering the aquifer led to an increase in
pH, the release of negatively charged constituents such as chromium, and the adsorption of
positively charged ions such as arsenic. Decreasing the permeability of the leakage zone by a factor
of 100 increased the travel time by a factor of four. In another study, Mvshakin etal. (2015)
modeled brine migration through a natural and induced fracture network. Their results indicated
that the main pathway for vertical migration of hydraulic fracturing fluid to overlying layers is
through the induced fractures, and not the natural fractures. The location of hydraulic fractures
relative to each other affects the extent of brine migration into overburden layers; compared to
single fractures separated by large distances, closely spaced fractures were associated with higher
pressures in—and, consequently, more brine migration into—overlying layers.
6.4 Synthesis
In the injection stage of the hydraulic fracturing water cycle, operators inject hydraulic fracturing
fluids into a well under pressure that is high enough to fracture the production zone. These fluids
flow through the well and then out into the surrounding formation, where they create fractures in
the rock, allowing hydrocarbons to flow through the fractures, to the well, and then up the
production string.
The production well and the surrounding geologic features function as a system that is often
designed with multiple elements that can isolate hydrocarbon-bearing zones and water-bearing
zones, including groundwater resources, from each other. This physical isolation optimizes oil and
6-69
-------
Chapter 6 - Well Injection
gas production and can protect drinking water resources via isolation within the well (by the casing
and cement) and/or through the presence of multiple layers of subsurface rock between the target
formations where hydraulic fracturing occurs and drinking water aquifers.
6.4.1 Summary of Findings
In this chapter, we consider impacts to drinking water resources to occur if hydraulic fracturing
fluids or other subsurface fluids affected by hydraulic fracturing enter and adversely impact the
quality of groundwater resources. Potential pathways for fluid movement to drinking water
resources may be linked to one or more components of the well and/or features of the subsurface
geologic system. If present, these potential pathways can, in combination with the high pressures
under which fluids are injected and pressure changes within the subsurface due to hydraulic
fracturing, result in the subsurface movement of fluids to drinking water resources.
The potential for these pathways to exist or form has been investigated through modeling studies
that simulate subsurface responses to hydraulic fracturing, and demonstrated via case studies and
other monitoring efforts. In addition, the development of some of these pathways—and fluid
movement along them—has been documented. It is important to note that, if multiple barriers
afforded by the well design and the presence of subsurface rock formations are present, the
development of a pathway within this system does not necessarily result in an impact on a drinking
water resource.
6.4.1.1 Fluid Movement via the Well
A production well undergoing hydraulic fracturing is subject to higher stresses during the relatively
brief hydraulic fracturing phase than during any other period of activity in the life of the well. If the
well cannot withstand the stresses experienced during hydraulic fracturing operations, pathways
associated with the casing and cement can form that can result in the unintended movement of
fluids into the surrounding environment (Section 6.2).
Multiple barriers within the well, including casing, cement, and a completion assembly can, if
present, isolate hydrocarbon-bearing formations from drinking water resources located at a
different depth. However, inadequate construction, defects in or degradation of the casing or
cement, and/or the absence of redundancies such as multiple layers of casing and proper
emplacement of cement can allow fluid movement into drinking water resources. Various studies of
wells in the Marcellus Shale showed failure rates between 3 and 10%, depending on the type of
failure studied (contamination of drinking water resources may or may not have occurred at these
wells). The EPA's Well File Review found that 3% of all hydraulic fracturing jobs involved a
downhole mechanical integrity failure, which generally resulted in hydraulic fracturing fluid
entering the annular space between the casing and formation or between two casing strings.
Ensuring proper well design and mechanical integrity—particularly proper cement placement and
quality—are important actions for preventing unintended fluid migration along the wellbore. While
not all of the mechanical integrity failures described above resulted in fluid movement to—or
contamination of—a drinking water resource, aspects of well design that lead to increased failure
6-70
-------
Chapter 6 - Well Injection
rates have the potential to increase the frequency or severity of impacts to drinking water quality
associated with hydraulic fracturing operations.
6.4.1.2 Fluid Movement within Subsurface Geologic Formations
Potential subsurface pathways for fluid migration to drinking water resources include flow of fluids
out of the production zone into formations above or below it, fractures extending out of the
production zone or into other induced fracture networks, intersections of fractures with abandoned
or active wells, and hydraulically induced fractures intersecting with faults or natural fractures
(Section 6.3).
Vertical separation between the zone where hydraulic fracturing operations occur and drinking
water resources reduces the potential for fluid migration to impact the quality of drinking water
resources. However, not all hydraulic fracturing operations are characterized by large vertical
distances between the production zone and drinking water resources. In coalbed methane plays,
which are typically shallower than shale gas plays, these separation distances can be smaller than
in other types of formations. Also, in certain areas, hydraulic fracturing is known to take place in
formations containing water that meets the salinity threshold that is used in some definitions of a
drinking water resource.
Lateral separation between wells undergoing hydraulic fracturing and other wells (including active
and abandoned wells) also reduces the potential for fluid migration to impact drinking water
resources. While some operators design fracturing treatments to communicate with the fractures of
another well and optimize oil and gas production, unintended communication between two wells or
their fracture systems can lead to spills in an offset well, which is an indicator of hydraulic
fracturing treatments extending beyond their planned design. These well communication incidents,
or "frac hits," have been reported in New Mexico, Oklahoma, and a few other locations. Surface
spills from well communication incidents have also been documented. Based on the available
information, frac hits most commonly occur on multi-well pads and when wells are spaced less than
1,100 ft (340 m) apart, but they have been observed at wells up to 8,422 ft (2,567 m) away from a
well undergoing hydraulic fracturing.
6.4.1.3 Impacts to Drinking Water Resources
We identified some example cases in the literature where the pathways associated with hydraulic
fracturing resulted in an impact on the quality of drinking water resources.
One of these cases took place in Bainbridge Township, Ohio, in 2007. Failure to cement
over-pressured formations through which a production well passed—and proceeding with the
hydraulic fracturing operation without adequate cement and an extended period during which the
well was shut in—led to a buildup of natural gas within the well annulus and high pressures within
the well. This ultimately resulted in movement of gas from the production zone into local drinking
water aquifers (Section 6.2.2.2). Twenty-six domestic drinking water wells were taken off-line and
the houses were connected to a public water system after the incident due to elevated methane
levels.
6-71
-------
Chapter 6 - Well Injection
Casings at a production well near Kill deer, North Dakota, ruptured in 2010 following a pressure
spike during hydraulic fracturing, allowing fluids to escape to the surface. Brine and tert-butyl
alcohol were detected in two nearby monitoring wells. Following an analysis of potential sources,
the only source consistent with the conditions observed in the two impacted water wells was the
well that ruptured during hydraulic fracturing. There is also evidence that out-of-zone fracturing
occurred at the well (Sections 6.2.2.1 and 6.3.2.2).
There are other cases where contamination of or changes to the quality of drinking water resources
near hydraulic fracturing operations were identified. Hydraulic fracturing remains a potential
contributing cause in these cases. For example:
• Migration of stray gas into drinking water resources involves many potential routes,
including poorly constructed casing and naturally existing or induced fractures in
subsurface formations. Multiple pathways for fluid movement may have worked in concert
in northeastern Pennsylvania (possibly due to cement issues or sustained casing pressure),
the Raton Basin in Colorado (where fluid migration may have occurred along natural rock
features or faulty well seals), and the Wattenberg field in Colorado (where the surface
casing depth and the presence of uncemented gas zones are major factors in determining
the likelihood of mechanical integrity failures and contamination). While the sources of
methane identified in drinking water wells in each study area could be determined with
varying degrees of certainty, attempts to definitively identify the pathways of migration
have generally been inconclusive (Text Box 6-3).
• At the East Mamm Creek drilling area in Colorado, inadequate placement of cement allowed
the migration of methane through natural faults and fractures in the area. This case
illustrates how construction issues, sustained casing pressure, and the presence of natural
faults and fractures, in conjunction with elevated pressures associated with hydraulic
fracturing, can work together to create a pathway for fluids to migrate toward drinking
water resources (Sections 6.2.2.2 and 6.3.2.4).
Additionally, there are places in the subsurface where oil and gas resources and drinking water
resources co-exist in the same formation. Evidence we examined indicates that some hydraulic
fracturing for oil and gas occurs within formations where the groundwater has a salinity of less
than 10,000 mg/L TDS. By definition, this results in the introduction of hydraulic fracturing fluids
into formations that meet both the Safe Drinking Water Act's salinity-based definition of an
underground source of drinking water and the broader definition of a drinking water resource
developed for this assessment. According to the data we examined, these formations are generally
in the western United States, e.g., near Pavillion, Wyoming. Hydraulic fracturing in a drinking water
resource may be of concern in the short-term (where people are currently using these zones as a
drinking water supply) or the long-term (if drought or other conditions necessitate the future use
of these zones for drinking water).
There are other cases in which production wells associated with hydraulic fracturing are alleged to
have caused contamination of drinking water resources. Data limitations in most of those cases
6-72
-------
Chapter 6 - Well Injection
(including the unavailability of information in litigation settlements resulting in sealed documents)
make it difficult to assess whether or not hydraulic fracturing was a cause of the contamination.
6.4.2 Factors Affecting Frequency or Severity of Impacts
The multiple barriers within the hydraulic fracturing well and the presence of subsurface low-
permeability geologic formations between the production zone and drinking water resources
isolate fluids from drinking water resources. Because of this, any factors that affect the integrity of
the system comprised of the well and the surrounding geology have the potential to affect the
frequency or severity of impacts on drinking water quality. The primary factors that can affect the
frequency or severity of impacts are: (1) the construction and condition of the well that is being
hydraulically fractured, (2) the amount of vertical separation between the production zone and
formations that contain drinking water resources, and (3) the location, depth, and condition of
nearby wells or natural faults or fractures.
The presence and condition of the well's casing and cement are key factors that affect the frequency
or severity of impacts to drinking water resources. Even in wells where there is substantial vertical
separation (e.g., thousands of feet), defects in the well can, in theory, allow fluid movement over
significant vertical distance. For example, fully cemented surface casing that extends through the
base of drinking water resources is a key protective component of the well. Risk evaluation studies
of a limited number of injection wells show that, if the surface casing is not set deeper than the
bottom of the drinking water resource, the risk of aquifer contamination increases a thousand-fold.
A review of wells that were hydraulically fractured in the Wattenberg field in Colorado showed that
wells with fewer casing and cementing barriers across gas-bearing zones exhibited higher rates of
failures. Most, but not all, wells used in hydraulic fracturing operations have fully cemented surface
casing.
The absence of or defects in casing or cement can be the result of inadequate design or
construction, including fewer layers of protective casing or when cement is incomplete (i.e., not
present across all oil-gas- or water-bearing formations), of inadequate quality, or improperly
emplaced. Wells that were constructed pursuant to older, less stringent requirements have a
greater likelihood of exhibiting mechanical integrity problems associated with inadequate design
and/or construction.
Deviated and horizontal wells may exhibit more casing and cement problems compared to vertical
wells. Some (but not all) studies have shown that sustained casing pressure—a buildup of pressure
within the well annulus that can indicate the presence of leaks—occurs more frequently in deviated
and horizontal wells compared to vertical wells. Cement integrity problems can arise as a result of
challenges in centering the casing and placing the cement in these wells. Absent efforts to ensure
the emplacement of sufficient cement that is of adequate integrity, the increased use of these wells
in hydraulic fracturing operations has the potential to increase the frequency at which associated
cementing problems occur. This, in turn, has the potential to increase the frequency of impacts to
the quality of drinking water resources.
6-73
-------
Chapter 6 - Well Injection
Even in optimally designed wells, degradation of the casing and cement as they age or due to the
cumulative effects of formation or operational stresses exerted on the well over time (e.g., cyclic
stresses in multi-stage fractures) can impact the mechanical integrity of the well and affect the
frequency of impacts to drinking water quality. Older wells exhibit more mechanical integrity
problems compared to newer wells when hydraulically fractured or re-fractured. If mechanical
integrity issues exist but are not detected and subsequently addressed, hydraulic fracturing fluids
or other fluids can move into drinking water resources and the concentrations of contaminants in
those drinking water resources—and therefore the severity of the impact—can increase.
In areas where there is little or no vertical separation between the production zone and drinking
water resources, there is a greater potential to increase the frequency or severity of impacts to
drinking water quality. For example, when the vertical separation is relatively small and other
subsurface pathways (e.g., artificial penetrations) are present, the potential for these pathways to
provide a more direct link between the production zone and a drinking water resource is greater
than if there is a large separation. As described above, there are places where hydraulic fracturing
operations occur in formations meeting the salinity threshold that is used in some definitions of a
drinking water resource. The practice of injecting hydraulic fracturing fluids into a formation that
also contains a drinking water resource can affect the quality of that water, because it is likely some
of that fluid remains in the formation following hydraulic fracturing. The properties (e.g., chemical
composition, toxicity, etc.) of hydraulic fracturing fluids or naturally occurring fluids that migrate to
drinking water resources can affect the severity of the impact on the quality of those resources (see
Chapter 9 for more information on the chemicals used in hydraulic fracturing fluids).
Where the separation between the production zone and drinking water resources is small, and
where natural or induced fractures that transect the layers between these formations are present,
there is a potential for increased frequency of impacts to drinking water quality via induced or
natural fractures or faults. (Impacts via well-related pathways can also be a concern in these
situations, as described above.)
Research shows that fractures created during hydraulic fracturing can extend out of the production
zone, and that the vertical component of fracture growth is generally greater in deeper formations
than shallow formations. Out-of-zone fracturing could be a concern in deeper formations if there is
little vertical separation between the production zone and a deep drinking water resource and
fractures propagate to unintended vertical heights. If out-of-zone fracturing is not detected (e.g., via
monitoring) and subsequently addressed, the impacts to the quality of drinking water resources
associated with fluid movement via these induced fractures have the potential to become more
severe.
Regardless of the extent of the vertical separation between the production zone and drinking water
resources, the presence of active or abandoned wells near hydraulic fracturing operations can
increase the potential for hydraulic fracturing fluids to move to drinking water resources. For
example, a deficiency in the construction of a nearby well (or degradation of the well's
components), can provide a pathway for movement of hydraulic fracturing fluids, methane, or
brines that might affect drinking water quality. If the fractures intersect an uncemented portion of a
6-74
-------
Chapter 6 - Well Injection
nearby wellbore, the fluids can potentially migrate along that wellbore into any formations where
the well is not cemented.
The frequency of impacts to the quality of drinking water resources may increase where wells are
densely spaced (particularly in shallow hydraulic fracturing operations where more fracture
propagation is expected to be in the horizontal direction). The frequency of impacts may also be
higher in mature oil and gas fields that pre-date the use of construction/plugging methods that can
withstand the stresses associated with hydraulic fracturing operations. In these mature fields, wells
tend to be older so degradation is a concern, and the location or condition of abandoned wells may
not be documented. Based on the information presented in this chapter, the increased use of
hydraulic fracturing in horizontal wells and in multiple wells on a single pad can increase the
likelihood that these pathways could develop. This, in turn, could increase the frequency at which
impacts on drinking water quality occur.
See Chapter 10 for a discussion of factors and practices that can reduce the frequency or severity of
impacts to drinking water quality.
6.4.3 Uncertainties
Generally, less is known about the occurrence of (or potential for) impacts of injection-related
pathways in the subsurface than for other components of the hydraulic fracturing water cycle,
which tend to be easier to observe and measure. Furthermore, while there is a large amount of
information available on production wells in general, there is little information that is both specific
to hydraulic fracturing operations and readily accessible across the states to form a national
picture.
6.4.3.1 Limited Availability of Information Specific to Hydraulic Fracturing Operations
There is extensive information available on the design goals for hydraulically fractured oil and gas
wells (i.e., to address the stresses imposed by high-pressure, high-volume injection), including from
industry-developed best practices documents. Additionally, many studies have documented how
production wells have historically been constructed, how they perform, and the rates at which they
experience problems that can lead to pathways for fluid movement. However, because of possible
differences in well construction and operational practices, it is unknown how historical well
performance studies apply to wells used in hydraulic fracturing operations.
Because wells that have been hydraulically fractured must withstand many of the same downhole
stresses as other production wells, we consider studies of the pathways for impacts to drinking
water quality in production wells to be relevant to identifying the potential pathways relevant to
hydraulic fracturing operations. However, without specific data on the as-built construction of wells
used in hydraulic fracturing operations, we cannot definitively state whether these wells are
consistently constructed to withstand the stresses they may encounter.
There is also, in general, very limited information available on the monitoring and performance of
wells used in hydraulic fracturing operations. Published information is sparse regarding
mechanical integrity tests (MITs) performed during and after hydraulic fracturing, the frequency at
6-75
-------
Chapter 6 - Well Injection
which mechanical integrity issues arise in wells used for hydraulic fracturing, and the degree and
speed with which identified issues are addressed. There is also little information available
regarding MIT results for the original hydraulic fracturing event in wells built for that purpose, for
wells that are later re-fractured, or for existing, older wells not initially constructed for hydraulic
fracturing but repurposed for that use.
These limitations on hydraulic fracturing-specific information make it difficult to provide definitive
estimates of the rate at which wells used in hydraulic fracturing operations experience the types of
mechanical integrity problems that can contribute to the movement of hydraulic fracturing fluids or
other fluids to drinking water resources.
There is also a limited number of peer-reviewed published studies based on groundwater sampling
that provide evidence to assess whether formation brines, hydraulic fracturing fluids, or gas move
in unintended ways through the subsurface during and after hydraulic fracturing. Subsurface
monitoring data (i.e., data that characterize the presence, migration, or transformation of fluids
within subsurface formations related to hydraulic fracturing operations) are scarce relative to the
tens of thousands of oil and gas wells that are estimated to be hydraulically fractured across the
country each year (see Chapter 3 for more information on the occurrence of hydraulic fracturing in
the United States).
Information on fluid movement within the subsurface and the extent of fractures that develop
during hydraulic fracturing operations is also limited. For example, limited information is available
in the published literature on how flow regimes or other subsurface processes change at sites
where hydraulic fracturing is conducted. Instead, much of the available research, and therefore the
literature, addresses how hydraulic fracturing and other production technologies perform to
optimize hydrocarbon production. In addition, much of the published data on fracture propagation
are for shale formations, and no large-scale data sets on fracture growth in other unconventional
formations exist or are publicly available.
These limitations on hydraulic fracturing-specific information make it difficult to provide definitive
estimates of the rate at which wells used in hydraulic fracturing operations experience the types of
mechanical integrity problems that can contribute to unintended fluid movement.
6.4.3.2 Limited Systematic, Accessible Data on Well Performance or Subsurface Movement
While the oil and gas industry generates a large amount of information on well performance as part
of operations, most of this is proprietary, or otherwise not readily available to the public in a
compiled or summary manner. Therefore, no national or readily accessible way exists to evaluate
the design and performance of individual wells or wells in a region, particularly in the context of
local geology or the presence of other wells and/or hydraulic fracturing operations. Many states
have large amounts of operator-submitted data, but information about construction practices or the
performance of individual wells is typically not in a searchable or aggregated form that would
enable assessments of well performance under varying settings, conditions, or timeframes.
Although it is collected in some cases, there is no collection, reporting, or publishing of baseline
(pre-drilling and/or pre-fracturing) and post-fracturing monitoring data on a national basis that
6-76
-------
Chapter 6 - Well Injection
could indicate the presence or absence of hydraulic fracturing-related fluids in shallow zones and
whether or not migration of those fluids has occurred. (See Chapter 10 for additional discussion of
data limitations.) Ideally, data from groundwater monitoring are needed to complement theories
and modeling on potential pathways and fluid migration.
While some of the types of impacts described above can occur quickly (i.e., on the scale of days or
weeks, as with mechanical integrity problems or well communication events), other impacts (e.g., in
slow-moving, deep groundwater) may be detectable only on much longer timescales. Without
comprehensive collection and review of information about how hydraulic fracturing operations
perform, fluid movement could occur without early detection, which could, in turn, increase the
severity of any resultant impacts to drinking water quality. For example, testing the mechanical
integrity of wells, monitoring the extent of the fractures that form, and conducting pre- and post-
hydraulic fracturing water quality monitoring can detect fluid movement (or the potential for fluid
movement) and provide opportunities to mitigate or minimize the severity of impacts associated
with unforeseen events.
The limited amount of available information also hinders our ability to evaluate how frequently
drinking water impacts are occurring, the probability that these impacts occur, or to what extent
they are tied to specific well construction, operation, and maintenance practices. This also
significantly limits our ability to evaluate the aggregate potential for hydraulic fracturing
operations to affect drinking water resources or to identify the potential cause of drinking water
contamination in areas where hydraulic fracturing occurs. The absence of this information greatly
limits the ability to make quantitative statements about the frequency or severity of these impacts.
6.4.4 Conclusions
The production well and the surrounding geologic features function as a system that provides
multiple barriers that can isolate hydrocarbon-bearing zones and water-bearing zones, including
drinking water resources. Because of this, factors affecting the integrity of any of these barriers
have the potential to adversely affect the quality of drinking water resources.
We have identified a number of pathways by which hydraulic fracturing fluids can reach and affect
the quality of drinking water resources. These pathways include migration via inadequate casing
and/or cement in the hydraulic fracturing well, fluid movement in the subsurface via fractures
extending out of the target zone, or vertical fluid movement via other natural or artificial structures.
The primary factors affecting the frequency or severity of impacts to drinking water quality
associated with hydraulic fracturing operations include the condition of the casing and cement of
the production well and their placement relative to drinking water resources, the extent of the
vertical separation between the production zone and drinking water resources, and the presence
and condition of offset wells or natural faults or fractures near the hydraulic fracturing operation.
There is evidence that, in some cases highlighted in the literature, these pathways have formed and
the quality of drinking water resources has been impacted. We do not know the frequency of such
impacts associated with the injection stage of the hydraulic fracturing water cycle, however. This is
related to the following: the subsurface environment is geologically complex, the relevant
6-77
-------
Chapter 6 - Well Injection
production processes cannot be directly observed, and publicly available data that can support an
evaluation of the impacts of hydraulic fracturing on the quality of drinking water resources is, in
general, very limited.
6-78
-------
Chapter 7 - Produced Water Handling
Chapter 7. Produced Water Handling
Abstract
Produced water is a byproduct of hydrocarbon production and flows to the surface through the
production well, along with oil and gas. Operators must store and dispose of (or in some cases treat)
large amounts of non-potable produced water, either on site or off site, and spills or releases of
produced water have the potential to impact drinking water resources. Unlike produced water from
conventional oil and gas production, produced water generated following hydraulic fracturing initially
contains returned hydraulic fracturing fluids. Much of the hydraulic fracturing fluid remains below
ground; the median amount of fluid returned to the surface is 30% or less. Up to several million gallons
of water can be produced from each well, with production generally decreasing with time.
Produced water contains several classes of constituents: salts, metals, radioactive materials, dissolved
organic compounds, and hydraulic fracturing chemicals and their transformation products (the result of
reactions of these chemicals in the subsurface). The concentrations of these constituents change with
time, as the initially returning hydraulic fracturing fluid blends with formation water. Typically, this
means that the produced water becomes more saline with time. Produced water composition and
volume vary from well to well, both among different formations and within formations. A large number
of organic compounds have been identified in produced water, many of which are naturally occurring
petroleum hydrocarbons; some are known hydraulic fracturing chemicals. Only a few transformation
products have been identified, and they include chlorinated organics.
Spills and releases of produced water with a variety of causes have been documented at different steps
in the production process. The causes include human error, equipment or container failure (for instance,
pipeline, tank or storage pit leaks), accidents, and storms. Unauthorized discharges may account for
some releases as well. An estimated half of the spills are less than 1,000 gal (3,800 L). A small number of
much larger spills has been documented, including a spill of 2.9 million gal (11 million L). Both short-
and long-term impacts to soil, groundwater, and surface from spills have occurred. For many spills,
however, the impacts are unknown. The potential of spills of produced water to affect drinking water
resources depends upon the release volume, duration, and composition, as well as watershed and water
body characteristics.
Data are lacking to characterize the severity and frequency of impacts on a nationwide scale. Suspected
local-scale impacts often require an extensive multiple lines-of-evidence investigation to determine
their cause. Further, when investigations do take place, the lack of baseline water quality data can make
it difficult to determine the cause and severity of the impact. In such cases, additional data are necessary
to determine the full extent of the impact of releases of produced water.
7-1
-------
Chapter 7 - Produced Water Handling
This page is intentionally left blank.
7-2
-------
Chapter 7 - Produced Water Handling
7. Produced Water Handling
7.1 Introduction
Water is a byproduct of oil and gas production. After the hydraulic fracturing of the formation is
completed, the injection pressure is reduced, and a possible inactive period where the well is "shut
in" is completed, water is allowed to flow back from the well to prepare for oil or gas production.1
This return-flow water may contain chemicals injected as part of the hydraulic fracturing fluid,
chemicals naturally occurring in the formation, or the products of reactions that take place in the
formation. Initially this water, sometimes called flowback, is mostly hydraulic fracturing fluid, but
as time goes on, water chemistry becomes more similar to water associated with the formation. For
formations containing saline water (brine), the salinity of the returned water increases as time
passes as the result of increased contact time between the hydraulic fracturing fluid and the
formation and inclusion of an increased portion of formation water. For this assessment, and
consistent with industry practice, the term produced water is used to refer to any water flowing
from the oil or gas well.
Produced water is piped directly to an injection well or stored and accumulated at the surface for
eventual management by injection into disposal wells, transport to wastewater treatment plants,
reuse, or in some cases, placement in evaporation pits or permitted direct discharge. See Text Box
ES-11 and Section 8.4 for discussion of these management practices.
Produced water spills and releases can occur due to several causes, including events associated
with pipelines, transportation, blowouts, and storage. Impacts to drinking water resources can
occur if this released water enters surface water bodies or reaches groundwater. Such impacts may
result in the water becoming unfit for consumption, either through obvious taste and odor
considerations or the constituents in the water exceeding hazard levels (Chapter 9). Once released
to the environment, transport of chemical constituents depends on the characteristics of the:
• Spill (volume, duration, concentration);
• Fluid (density as influenced by salinity);
• Chemicals (volatility, sorption, solubility); and
• Site-specific environmental characteristics (surface topography and location of surface
water bodies, the type of the soil and aquifer materials, layering and heterogeneity of
rocks, and the presence of dissolved oxygen and other factors needed to support
biodegradation, and the presence of inorganic species that affect metal transport).
This chapter provides characterization of produced water and also provides background
information for the coverage of wastewater disposal and reuse in Chapter 8. Chapter 7 addresses
the characteristics of produced water including per-well generation of produced water. Chapter 8
considers management of this water, now called wastewater, at an aggregate level, and thus
1 There can be no shut-in period at all or it can last several weeks fStepan etal.. 20101
7-3
-------
Chapter 7 - Produced Water Handling
discusses state, regional, and national estimates of treatment volumes. While Chapter 7 considers
impacts from several types of unintentional releases, Chapter 8 focuses on impacts that are
associated with wastewater management practices. One specific issue, leakage from pits and
impoundments, is introduced in Chapter 7 as one of several avenues for accidental releases, with a
more detailed exploration of the use of pits in wastewater management presented in Chapter 8.
Chapter 7 begins with a review of definitions for flowback and produced water in Section 7.1.1.
Definitions are followed by a discussion in Section 7.2 of water volumes per well, first presenting
data on the volume and percent of hydraulic fracturing fluid returned to the surface and then
presenting data on the volume of water returned during production. These data all represent the
response of individual wells. Because of the need to have aggregated volumes for estimating
wastewater treatment loadings, estimates of total volumes are given in Section 8.2.
Chapter 7 continues with discussion of the chemical composition of produced water (Section 7.3).
Because the composition of produced water is only known through analysis of samples, laboratory
methods and their limitations are described in Section 7.3.1. Time-dependent changes in
composition are discussed via three specific examples in Section 7.3.3, followed by discussion of
five types of constituents: salts, metals, radioactive materials, organics, and known hydraulic
fracturing additives in Section 7.3.4. The chemical and geological processes controlling the chemical
composition of produced water are described in Appendix E. Spatial and temporal trends in the
composition of produced water are illustrated with examples from the literature and data compiled
for this report (Section 7.3.5).
The potential for impacts on drinking water resources of produced water releases and spills are
described based on reported spill incidents (Section 7.4), and examples of spills from specific
sources and data compilation studies are given in Section 7.4.2. The potential for impacts is
described using contaminant transport principles in Section 7.6. The chapter concludes with a
discussion of uncertainties and knowledge gaps, factors that influence the severity of impacts, and
major findings (Section 7.7).
7.1.1 Definitions
Multiple definitions exist for the terms flowback and produced water. Appendix Section E.l gives
examples of definitions used by different organizations. These differing definitions reflect differing
usage of the terms among various groups and that produced water reflects the continuously
varying mixture between returning injection fluid and formation water. The majority of produced
water definitions are fundamentally similar. The following definition is used in this report for
produced water: any type of water that flows from the subsurface through oil and gas wells to the
surface as a by-product of oil and gas production. Thus produced water can variously refer to
returned hydraulic fracturing fluid, formation water alone, or a mixture of the two.
The term flowback has two major meanings. First is the process used to prepare the well for
production by allowing excess liquids and proppant to return to the surface. The second use of the
term is to refer to fluids predominantly containing hydraulic fracturing fluid that return to the
surface. Because formation water can contact and mix with injection fluids, the distinction between
returning hydraulic fracturing fluid and formation water is not clear. Definitions of flowback are
7-4
-------
Chapter 7 - Produced Water Handling
operational in the sense that they include some characteristic of the oil and gas operation (i.e.,
fluids returning within 30 days). These reflect that during the early phases of operation, a higher
concentration of chemical additives is expected and later, water is characteristic of the formation.
Because we use existing literature in our review, we do not introduce a preferred definition of
flowback, and describe all water flowing from the well as produced water.
7.2 Volume of Hydraulic Fracturing Flowback and Produced Water
Veil f20151 estimated that, in 2012, all types (i.e., from conventional and unconventional
reservoirs) of U.S. onshore and offshore oil and gas production generated 8.90 x 1011 gal (3.37 x
1012 L) of produced water. More details and state-level estimates are given in Section 8.2. This
section presents information on flowback and produced water volume over various time scales, and
where possible, on a per-well and per-formation basis, because characteristics and volume of
flowback and produced water vary by well, formation, and time.
The amount of produced water from a well varies and depends on several factors, including
production, formation, and operational factors. Production factors include the amount of fluid
injected, the type of hydrocarbon produced (gas or liquid), and the location within the formation.
Formation factors include the formation pressure, the interaction between the formation and
injected fluid (capillary forces), and reactions within the reservoir. Operational factors include the
volume of the fractured production zone that includes the length of well segments and the height
and width of the fractures. Certain types of problems also influence water production, including
possible loss of mechanical integrity and subsurface communication between wells, both of which
can result in an unexpected increase in water production (U.S. GAP. 2012: Byrnes. 2011: DOE.
2011a: GWPC and ALL Consulting. 2009: Reynolds and Kiker. 20031.
The processes that allow gas and liquids to flow are related to the conditions along the faces of
fractures. Byrnes (2011) conceptualized fluid flow across the fracture face as being composed of
three phases. The first is characterized by forced imbibition of fluid into the reservoir and occurs
during and immediately following fracture stimulation.1 Second is fluid redistribution within the
reservoir rock, due to capillary forces. Estimates have shown that 50% or more of fracturing fluid
could be captured within the Marcellus shale if imbibition drives water 2 to 6 in (5 to 15 cm) into
the formation fEngelder. 2012: Byrnes. 2011: He. 20111. In the last phase, water flows out of the
formation when the well is opened and pressure is reduced in the wellbore and fractures. The
purpose of this phase is to recover as much of the injected fluid as possible (Byrnes. 2011) to allow
higher oil or gas flow rates. The length of the last phase and, consequently, the amount of water
removed, depends on factors such as the amount of injected fluid, the permeability and relative
permeability of the reservoir, capillary pressure properties of the reservoir rock, and the pressure
near the fracture faces.2 The well can be shut in for varying time periods depending on operator
scheduling, surface facility construction and connection thereto, or other reasons.
1 The displacement of a non-wet fluid (i.e., gas] by a wet fluid (typically water]. Adapted from Pake f19781.
2 When multiple fluids (water, oil, gas] occupy portions of the pore space, the permeability to each fluid depends on the
fraction of the pore space occupied by the fluid and the fluid's properties. As defined by Pake f 19781. when this effective
permeability is normalized by the absolute permeability, the resulting relationship is known as the relative permeability.
7-5
-------
Chapter 7 - Produced Water Handling
7.2.1 Flowback of Injected Hydraulic Fracturing Fluid
The amount of water produced by wells within the first few days following fracturing varies from
formation to formation. Wells in the Mississippi Lime and Permian Basin can produce 1 million gal
(3.8 million L) in the first 10 days of production. Wells in the Barnett, Eagle Ford, Granite Wash,
Cleveland/Tonkawa Sand, Niobrara, Marcellus, and Utica Shales can produce 300,000 to 1 million
gal (1.14 to 3.78 million L) within the first 10 days. Haynesville wells produce less, about 250,000
gal (950,000 L) (Mantell, 2013"). Data show that the rate of water produced during the flowback
period decreases as time passes (Ziemkiewicz etal., 2014; Hansen etal., 2013; Hayes, 2009").
It is not possible to specify precisely the amount of injected fluids that return in the flowback,
because there is not a clear distinction between flowback and produced water, and the indicators
(e.g., salinity and radioactivity, to name two) are not routinely monitored (GWPC and ALL
Consulting, 2009"). Rather, flowback estimates usually relate the amount of produced water
measured at a given time after fracturing as a percentage of the total amount of injected fluid.
Estimates of the fraction of injected hydraulic fracturing fluid that returns as flowback are highly
variable (U.S. EPA, 2016d; Vengosh etal., 2014; Mantell, 2013; Vidic etal., 2013; Minnich, 2011; Xu
etal., 2011"). The maxima are less than 85% in all but one of the examples given in Table 7-1, Table
7-2, and Table 7-3, and most of the median values are less than 30%. In some cases, the amount of
flowback is greater than the amount of injected hydraulic fracturing fluid, and the additional water
comes from the formation (Nicotetal., 2014") or from a conductive pathway from an adjacent
formation (Arkadakskiy and Rostron, 2013"). See Appendix Section E.2.1 for more details.
Table 7-1. Data from one company's operations indicating approximate total water use and
approximate produced water volumes within 10 days after completion of wells.
From Mantell
2013).
Formation
Approx. total average
water use per well
(million gal)
Produced water (flowback)
within the first 10 days
after completion
Produced water as a
percentage of average water
use per well
Low estimate
(million gal)
High or only
estimate
(million gal)
Low estimate
(% of total
water use)
High or only
estimate (% of
total water use)
Gas shale plays (primarily dry gas)
Barnett3
3.4
0.3
1.0
9%
29%
Marcellus3
4.5
0.3
1.0
7%
22%
Haynesville
5.4
-
0.25
-
5%
Liquid plays (gas, oil, condensate)
Mississippi
Lime
2.1
-
1.0
-
48%
7-6
-------
Chapter 7 - Produced Water Handling
Formation
Approx. total average
water use per well
(million gal)
Produced water (flowback)
within the first 10 days
after completion
Produced water as a
percentage of average water
use per well
Low estimate
(million gal)
High or only
estimate
(million gal)
Low estimate
(% of total
water use)
High or only
estimate (% of
total water use)
Cleveland/
Tonkawa
2.7
0.3
1.0
11%
37%
Niobrara
3.7
0.3
1.0
8%
27%
Utica
3.8
0.3
1.0
8%
26%
Granite
Wash
4.8
0.3
1.0
6%
21%
Eagle Ford
4.9
0.3
1.0
6%
20%
a Mantell (2011) reported produced water for the first 10 days at 500,000 to 600,000 gal for the Barnett, Fayetteville and
Marcellus Shales.
Table 7-2. Additional short-, medium-, and long-term produced water estimates.
Location-formation
Produced water as
percentage of injected fluid
Reference
Comment
Estimates without reference to a specific data set
Unspecified Shale
5% - 35%
Haves (2011)
Marcellus Shale
10% - 25%
Minnich (2011)
Initial flowback
ND-Bakken
25%
EERC (2013)
Estimates with reference to specific data evaluation
Short duration
Marcellus Shale
10%
Clark etal. (2013)
0-10 days
TX—Barnett
20%
Clark etal. (2013)
0-10 days
TX—Haynesville
5%
Clark etal. (2013)
0-10 days
AR— Fayetteville
10%
Clark etal. (2013)
0-10 days
Medium duration
WV— Marcellus
8%
Hansen et al. (2013)
30 days
Marcellus Shale
24%
Haves (2011, 2009)
Average from 19 wells, 90
days
7-7
-------
Chapter 7 - Produced Water Handling
Location-formation
Produced water as
percentage of injected fluid
Reference
Comment
Long duration
TX—Barnett
~100%a
Nicot et al. (2014)
72 months
WV—Marcellus
10% - 30%
Ziemkiewicz et al. (2014)
Up to 115 months
TX—Eagle Ford
<20%
Nicot and Scanlon (2012)
Lifetime
Unspecified duration
PA—Marcellus
6%
Hansen et al. (2013)
a Approximate median with large variability: 5th percentile of 20% and 90th percentile of 350%.
Table 7-3. Flowback water characteristics for wells in unconventional reservoirs.
Source: U.S. EPA (2016d). The formation-level data used to develop Tables 7-3 and 7-4 appear in Appendix Table E-l.
Resource
type
Well type
Fracturing fluid
(million gal)
Flowback
(percent of fracturing fluid returned)
Weighted
average
Range
Data
points
Weighted
average
Range
Data
points
Shale
Horizontal
4.2
0.091-24
80,388
7%
0%-580%
7,377
Directional
1.4
0.037-20
340
33%
l%-57%
36
Vertical
1.1
0.015-19
5,197
96%
2%-581%
57
Tight
Horizontal
3.4
0.069-12
7,301
12%
0%-60%
75
Directional
0.05
0.046-4
3,581
10%
0%-60%
342
Vertical
1
0.016-4
10,852
4%
0%-60%
130
7.2.2 Produced Water Volumes
Mantell (2013. 2011) described the amount of produced water over the long term as high,
moderate, or low for several formations. Wells in the Barnett Shale, Cleveland/Tonkawa Sand,
Mississippi Lime, and the Permian Basin can produce more than 1,000 gal (3,800 L) of water per
million cubic feet (MMCF) of gas. The most water-productive of these can be as high as
5,000 gal (19,000 L) per MMCF of gas. As a specific example, a high water producing formation in
the western United States was described as producing 4,200 gal (16,000 L) per MMCF of gas for the
life of the well (McElreath. 20111. The well was fractured and stimulated with about 4 million gal
(15 million L) of water and returned 60,000 gal (230,000 L) per day in the first 10 days, followed by
8,400 gal (32,000 L) per day in the remainder of the firstyear. The Niobrara, Granite Wash, Eagle
Ford, Haynesville, and Fayetteville Shales are relatively dry formations (with small amounts of
naturally occurring formation water) and produce between 500 and 2,000 gal (1,900 to 7,600 L) of
7-8
-------
Chapter 7 - Produced Water Handling
produced water per MMCF of gas (Mantell. 20131. The Utica and Marcellus Shales are viewed as
drier still and produce less than 200 gal (760 L) per MMCF of gas.
Wells producing in various formation show high produced water volume variability, including the
Barnett Shale, which was attributed by Nicotetal. (2014) to a few wells with exceptionally high
water production. Some of these wells produced more than the amount of injected fracturing fluid.
Wells in conventional and unconventional reservoirs produce differing amounts of water.
Individual hydraulically fractured wells producing gas from the Marcellus Shale produced more
water than hydraulically fractured wells in conventional wells in Pennsylvania (Lutz etal.. 20131.
However, on a per-unit of gas produced basis, wells producing from the Marcellus Shale generate
less water (35%), than those in the conventional formations.
The EPA (2016d) reported characteristics of long-term produced water for hydraulically fractured
shale and tight formations (Table 7-4). For shale, horizontal wells produced more water (1,100
gal/day; 4,200 L/day) than vertical wells (500 gal/day; 1,900 L/day). Typically, this would be
attributed to the longer length of the production zone in horizontal laterals than in vertical wells.
Table 7-4. Long-term produced water generation rates (gal/day per well) for wells in
unconventional reservoirs.
Source: U.S. EPA (2016d). The formation-level data used to develop Tables 7-3 and 7-4 appear in Appendix Table E-l.
Resource type
Well type
Long-Term Produced Water Generation Rates
(gal per day per well)
Weighted average
Range
Data points
Shale
Horizontal
1,100
0-29,000
43,893
Directional
820
0.83-12,000
1,493
Vertical
500
4.8-51,000
12,551
Tight
Horizontal
980
10-120,000
4,692
Directional
390
15-8,200
10,784
Vertical
650
0.71-2100
34,624
In an example from the Pennsylvania Marcellus Shale, the EPA determined that, for vertical wells in
unconventional reservoirs, 6% of water came from drilling, 35% from flowback, and 59% from
long-term produced water; for horizontal wells, the corresponding numbers were 9%, 33%, and
58% (U.S. EPA. 2016d). This result agrees with the U.S. Department of Energy (DOE. 2011a) who
concluded that the characteristic small amount of produced water from the Marcellus Shale was
due either to its low water saturation or low relative permeability to water (see Section 6.3.2.1). For
these dry formations, low shale permeability and high capillarity cause water to imbibe into the
formation, where some is retained permanently.
7-9
-------
Chapter 7 - Produced Water Handling
7.2.2.1 Time Trends
High rates of water production (flowback) typically occur in the first few months after hydraulic
fracturing, followed by rates reduced by an order of magnitude fe.g.. Nicot et al.. 20141. In many
cases half of the total produced water from a well is generated in the first year. Similarly, the EPA
(2016d) reported a general rule of thumb that, for unconventional reservoirs, the volume of
flowback (which occurs over a short period of time) is roughly equal to the volume of long-term
produced water. These trends in produced water volumes occur within the timeline of hydraulic
fracturing activities (Section 3.3), and show that the large, initial return volumes of flowback last
for several weeks, whereas the lower-rate produced water phase can last for years (Figure 7-1).
100,000
o»
Q.
>
re
Q
&_
<0
Q.
"re
2,
0)
£
5
i-
V
+*
fO
10,000
1,000
100
0
500 1000 1500
Days on Production
2000
Figure 7-1. Generalized examples of produced water flow from five formations.
Actual produced water flows vary by location, play, basin, and amount of water used for hydraulic fracturing (EWI,
2015). Figure used with permission.
7.2.2.2 Coalbed Methane
Water is pumped from coal seams to reduce pressure so that gas adsorbed to the surface of the coal
can flow to the production well fGuerra et al.. 20111. Consequently, CBM tends to produce large
volumes of water early on: more than conventional gas-bearing formations (U.S. GAP. 20121
(Figure 7-2). Within producing CBM formations, water production can vary for unknown reasons
fU.S. GAP. 20121. As an example, data show that CBM production in the Powder River Basin
produces 16 times more water than that in the San Juan Basin (U.S. GAP. 20121.
ferm/an
Fayetteville
Marce//Us
7-10
-------
Chapter 7 - Produced Water Handling
9,000,000
8,000,000
"S
•22 7,000,000
C
O
tt 6,000,000
3
"O
| 5,000,000
10
7-11
-------
Chapter 7 - Produced Water Handling
fracturing fluid (Section 7.3.4.7 and Appendix E.3.5.). These studies make clear that standard
analytical methods are not adequate for detecting and quantifying the numerous organic chemicals,
both naturally occurring and anthropogenic, that are now known to occur in produced water
fLester etal.. 2015: Maguire-Bovle and Barron. 2014: Thurman etal.. 20141. Similarly, methods
commonly applied for the analysis of radionuclides in drinking water may suffer from analytical
interferences that result in poor data quality (Maxwell etal.. 2016: Ying etal.. 2015: Zhang et al..
2015b: Nelson etal.. 2014: U.S. EPA. 2014i. 2004b], In these instances, alternative methods that
have been developed to support the nuclear materials production and waste industry provide more
reliable approaches to ensure adequate detection limits and avoid sample matrix interferences that
are anticipated for the high salinity and concentrations of organic constituents that may be present
in produced water samples.1 Development of advanced or non-routine methods for both organics
and inorganics (especially radium) suggests that data generated from earlier methods may be less
reliable that those developed by the new methods fNelson et al.. 20141. and that advanced
analytical techniques are needed to detect or quantify some analytes.
The compositional data that follow in this chapter and Appendix E rely on the analytical procedures
used in measurement and were summarized as noted from numerous produced water studies or
compilations, such as the U.S. Geological Survey (USGS) produced water database (Blondes etal..
20141.
7.3.2 Factors Influencing Produced Water Composition
Several interacting factors influence the chemical composition of produced water: (1) the
composition of injected hydraulic fracturing fluids, (2) the targeted geological formation and
associated hydrocarbon products, (3) the stratigraphic environment, and (4) subsurface processes
and residence time fBarbot etal.. 2013: Chapman et al.. 2 012: Dahm etal.. 2011: Blauch etal..
20091.
The mineralogy and structure of a formation are determined initially by deposition, when rock
grains settle out of their transporting medium fMarshak. 20041. Generally, shale forms from clays
that were deposited in deep, oxygen-poor marine environments, and sandstone can form from sand
deposited in shallow marine environments (Ali et al.. 2010: U.S. EPA. 2004a). Coal forms when
carbon-rich plant matter collects in shallow peat swamps. In the United States, coal formed in both
freshwater (northern Rocky Mountains) and marginal-marine environments (Alabama's Black
Warrior formation) fNRC. 2010: Horsey. 19811. Consequently, shale and sandstone produced water
are expected to be saline, and CBM water may be much less so.
7.3.3 Produced Water Composition During the Flowback Period
The chemistry of produced water changes over time, especially during the first days or weeks after
hydraulic fracturing. Generally, produced water concentrations of cations, anions, metals, naturally
occurring radioactive material (NORM), and organics increase as time goes on (Barbot etal.. 2013:
Haluszczak et al.. 2 013: Chapman etal.. 2012: Davis etal.. 2012: Gregory etal.. 2011: Blauch etal..
1 For guidance in planning, implementing, and assessing projects that require laboratory analysis of radionuclides, see
U.S. EPA f2004bl
7-12
-------
Chapter 7 - Produced Water Handling
20091. The causes include precipitation and dissolution of salts, carbonates, sulfates, and silicates;
pyrite oxidation; leaching and biotransformation of organic compounds; and mobilization of NORM
and trace elements. Concurrent precipitation of sulfates (e.g., BaSCU) and carbonates (e.g., CaCOs)
alongside decreases in pH, alkalinity, dissolved carbon, and microbial abundance and diversity
occur over time after hydraulic fracturing (Orem etal.. 2014: Barbotetal.. 2013: Murali Mohan et
al.. 2013: Davis etal.. 2012: Blauch etal.. 2009: Brinck and Frost. 20071. Leaching of organics
appears to be a result of injected and formation fluids associating with shale and coal strata (Orem
etal.. 20141. Concentrations of organics in CBM produced water decrease with time, possibly due to
the depletion of coal-associated water through formation pumping (Orem etal.. 20071.
7.3.3.1 Total Dissolved Solids
Produced water total dissolved solids concentrations (TDS) increase by varying degrees because of
the formation's geological origin. As an example, TDS concentrations increased to upper bound
values in samples from four Marcellus Shale gas wells (Chapman etal.. 2012) (Figure 7-3). The
increased TDS was composed of increased sodium, calcium, and chloride (Chapman etal.. 2012:
Blauch etal.. 2009). Similarly, TDS in flowback from the Westmoreland County wells started low
and exceeded that of typical seawater (35,000 mg/L) within three days f Chapman et al.. 20121. In a
similar study, wells with hydraulic fracturing fluid containing less than 1,000 mg/L saw TDS
concentrations increase above a median value of 200,000 mg/L within 90 days fHaves. 20091.
200,000
180,000
160,000
<140,000
ajo
^120,000
C/l
^ 100,000
in
| 80,000
o
.2 60,000
Q
3 40,000
E-
20,000
0
0 5 10 15 20 25 30
Days Post-Fracturing
Figure 7-3. TDS concentrations measured through time for injected fluid (at 0 days), and
produced water samples from four Marcellus Shale gas wells in three southwest Pennsylvania
counties.
Data from Chapman et al. (2012).
. * *
• Greene Co.
¦ Westmoreland Co. 1
¦ Westmoreland Co. 2
A Washington Co.
¦ 11 1 1 ¦ 11 1 1 11 1 ¦ 1 11 1 ¦ 1 1 ¦ 11 1 11 11 1 ¦
7-13
-------
Chapter 7 - Produced Water Handling
7.3.3.2 Radionuclides
Shales and sandstones naturally contain various radionuclides fSturchio etal.. 200 ll.1 Radium in
pore waters or adsorbed onto clay particles and grain coatings can dissolve and return in produced
water (Langmuir and Riese. 1985). Available data indicate that radium and TDS concentrations in
produced water are positively correlated (Rowan etal.. 2011: Fisher. 19981. likely because radium
remains adsorbed to mineral surfaces when salinity is low, and then desorbs into solution with
increased salinity fSturchio etal.. 20011. As an example, over the course of 20 days, radium
concentration in flowback from a Marcellus Shale gas well increased by almost a factor of four
(Chapman et al.. 2 012: Rowan etal.. 20111 (Figure 7-4).
200,000
180,000
160,000 ^
140,000 I3
120,000 |
100,000 £
-------
Chapter 7 - Produced Water Handling
water contacting the coal may become depleted in DOC to the degree that when outside water of
lower DOC is produced, the resulting DOC concentrations in the produced water are reduced fOrem
et al.. 20141
100,000
90,000
80,000
70,000
§ 60,000
E.
^ 50,000
"C
Z 40,000
o
30,000
20,000
10,000
0
0 50 100 150 200 250 300 350
Days Post-Fracturing
(a)
600
500
400
1
300
8
o
200
100
0
0 50 100 150 200 250 300 350
Days Post-Fracturing
(b)
Figure 7-5. (a) Increasing chloride (CI) and (b) decreasing DOC concentrations measured
through time for samples from three Marcellus Shale gas wells on a single well pad in Greene
County, PA.
Data from Cluff et al. (2014). Reprinted with permission from Cluff, M; Hartsock, A; Macrae, J; Carter, K; Mouser,
PJ. (2014). Temporal changes in microbial ecology and geochemistry in produced water from hydraulically
fractured Marcellus Shale Gas Wells. Environ Sci Technol 48: 6508-6517. Copyright 2014 American Chemical
Society.
As an example, produced water DOC concentrations decreased from their initial levels twofold from
the hydraulic fracturing fluid and initial samples (Figure 7-5b) followed by a decrease of 11-fold
t
t
• Well 1
¦ Well 2
A VVHI 3
7-15
-------
Chapter 7 - Produced Water Handling
over nearly 11 months. The DOC leveled off several months after hydraulic fracturing, presumably
as a result of in situ attenuation processes fCluffetal.. 20141. As DOC was decreasing, chloride
concentrations increased five- to six-fold. These chloride concentrations increased linearly during
the first two weeks fCluff etal.. 20141 and then later approached higher levels (Figure 7-5a). The
pattern in the DOC and chloride levels reflected the changing composition of the produced water—
initially high in DOC from hydraulic fracturing additives and low in salinity, then higher in salinity
and lower in DOC reflecting the chemistry of formation water. The changing composition of
produced water suggests that the potential concern for produced water spills also changes: initially
the produced water may contain more hydraulic fracturing chemicals, and later the concern may
shift to the impact of high salinity water.
7.3.4 Produced Water Composition
The chemical composition of produced water continues to change after the initial flowback period.
Produced water may contain a range of constituents, but in widely varying amounts. Generally,
these can include:
• Salts, including those composed from chloride, bromide, sulfate, sodium, magnesium and
calcium;
• Metals including barium, manganese, iron, and strontium;
• Radioactive materials including radium (radium-226 and radium-228);
• Oil and grease, and dissolved organics (including BTEX);1
• Hydraulic fracturing chemicals, including tracers and their transformation products; and
• Produced water treatment chemicals.2
We discuss these groups of chemicals and then conclude by discussing variability within formation
types and within production zones.
7.3.4.1 Similarity of Produced Water from Conventional and Unconventional Reservoirs
Produced water generated from unconventional reservoirs is reported to be similar to produced
water from conventional reservoirs in terms of TDS, pH, alkalinity, oil and grease, TOC, and other
organics and inorganics (Wilson. 2014: Haluszczaketal.. 2013: Alley etal.. 2011: Hayes. 2009:
Sirivedhin and Dallbauman. 20041. Although produced water salinity varies within and among
shales and tight formations, produced water is typically characterized as saline fLee and Neff. 2011:
Blauch etal.. 20091. Produced water from coalbeds may have low TDS if the coal source bed was
formed in freshwater. Saline produced water is also enriched in major anions (e.g., chloride,
bicarbonate, sulfate); cations (e.g., sodium, calcium, magnesium); metals (e.g., barium, strontium);
1 BTEX is an acronym representing benzene, toluene, ethylbenzene, and xylenes.
2 Some chemicals are added to produced water for the purpose of oil/water separation, improved pipeline flow, or
equipment maintenance, including prevention of corrosion and scaling in equipment fCal/EPA. 2016). Generally the
chemicals serve as clarifiers, emulsifiers, emulsion breakers, floating agents, and oxygen scavengers. Among proprietary
formulations, a few specific chemicals have been disclosed including low concentrations of benzene, toluene, and
inorganics (acetic acid, ammonium chloride, cupric sulfate, sodium hypochlorite].
7-16
-------
Chapter 7 - Produced Water Handling
naturally occurring radionuclides (e.g., radium-226, radium-228) (Chapman etal.. 2012: Rowan et
al.. 20111: and organics (e.g., hydrocarbons) fOrem etal.. 2007: Sirivedhin and Dallbauman. 20041.
7.3.4.2 Variability in Produced Water Composition Among Unconventional Reservoirs
Alley etal. (20111 compared geochemical parameters of shale gas, tight gas, and CBM produced
water. This comparison aggregated data on produced water from original analyses, peer-reviewed
literature, and public and confidential government and industry sources and determined the
statistical significance of the results.
As shown in Table 7-5, Alley etal. f20111 found that of the constituents of interest common to all
three types of produced water from unconventional reservoirs (calcium, chloride, potassium,
magnesium, manganese, sodium, and zinc):
1. Shale gas produced water had significantly different concentrations from those of CBM;
2. Shale gas produced water constituent concentrations were significantly similar to those of
tight gas, except for potassium and magnesium; and
3. Five tight gas produced water constituent concentrations (calcium, chloride, potassium,
magnesium, and sodium) were significantly similar to those of CBM f Alley etal.. 20111.
The degree of variability between produced waters of these three resource types is consistent with
the degree of mineralogical and geochemical similarity between shale and sandstone formations,
and the lack of the same between shale and coalbed formations fMarshak. 20041. Compared to the
others, shale gas produced water tends to be more acidic, as well as enriched in strontium, barium,
and bromide. CBM produced water is alkaline, and it contains relatively low concentrations of TDS
(one to two orders of magnitude lower than in shale and sandstone). It also contains lower levels of
sulfate, calcium, magnesium, DOC, sodium, bicarbonate, and oil and grease than typically observed
in shale and sandstone produced waters f Alley etal.. 2011: Dahm etal.. 2011: Benko and Drewes.
2008: Van Voast. 20031.1
Table 7-5. Compiled minimum and maximum concentrations for various geochemical
constituents in produced water from shale gas, tight gas, and CBM produced water.
Source: Alley et al. (2011).
Parameter
Unit
Shale gas3
Tight GasSandsb
CBMC
Alkalinity
mg/L
160-188
1,424
54.9-9,450
Ammonium-N
mg/L
-
2.74
1.05-59
Bicarbonate
mg/L
ND-4,000
10-4,040
-
Conductivity
US/cm
-
24,400
94.8-145,000
Nitrate
mg/L
ND-2,670
-
0.002-18.7
1 Several regions had low representation in the Alley et al. (2011) data set, including the Appalachian Basin
(western New York and western Pennsylvania), West Virginia, eastern Kentucky, eastern Tennessee, and
northeastern Alabama.
7-17
-------
Chapter 7 - Produced Water Handling
Parameter
Unit
Shale gas3
Tight GasSandsb
CBMC
Oil and grease
mg/L
-
42
-
PH
SUd
1.21-8.36
5-8.6
6.56-9.87
Phosphate
mg/L
ND-5.3
-
0.05-1.5
Sulfate
mg/L
ND-3,663
12-48
0.01-5,590
Radium-226
pCi/g
0.65-1.031
-
-
Aluminum
mg/L
ND-5,290
-
0.5-5,290
Arsenic
mg/L
-
0.17
0.0001-0.06
Boron
mg/L
0.12-24
-
0.002-2.4
Barium
mg/L
ND-4,370
-
0.01-190
Bromide
mg/L
ND-10,600
-
0.002-300
Calcium
mg/L
0.65-83,950
3-74,185
0.8-5,870
Cadmium
mg/L
-
0.37
0.0001-0.01
Chloride
mg/L
48.9-212,700
52-216,000
0.7-70,100
Chromium
mg/L
-
0.265
0.001-0.053
Copper
mg/L
ND-15
0.539
ND-0.06
Fluorine
mg/L
ND-33
-
0.05-15.22
Iron
mg/L
ND-2,838
0.015
0.002-220
Lithium
mg/L
ND-611
-
0.0002-6.88
Magnesium
mg/L
1.08-25,340
2-8,750
0.2-1,830
Manganese
mg/L
ND-96.5
0.525
0.002-5.4
Mercury
mg/L
-
-
0.0001-0.0004
Nickel
mg/L
-
0.123
0.0003-0.20
Potassium
mg/L
0.21-5,490
5-2,500
0.3-186
Sodium
mg/L
10.04-204,302
648-80,000
8.8-34,100
Strontium
mg/L
0.03-1,310
-
0.032-565
Uranium
mg/L
-
-
0.002-0.012
Zinc
mg/L
ND-20
0.076
0.00002-0.59
No value available; ND, non-detect. If no range, but a singular concentration is given, this is the maximum concentration.
a n = 541. Alley et al. (2011) compiled data from USGS (2006); Mcintosh and Walter (2005); Mcintosh et al. (2002) and
confidential industry documents.
bn = 137. Alley et al. (2011) compiled data from USGS (2006) and produced water samples presented in Alley et al. (2011).
c Alley et al. (2011) compiled data from Montana GWIC (2009); Thordsen et al. (2007); ESN Rocky Mountain (2003); Rice et al.
(2000); Rice (1999); Hunter and Moser(1990).
dSU = standard units.
7-18
-------
Chapter 7 - Produced Water Handling
7.3.4.3 General Water Quality Parameters
Data characterizing the content of produced water from unconventional reservoirs in 12 shale and
tight formations and CBM basins were evaluated for this assessment. These reservoirs and basins
include parts of 18 states, but the data do not allow for comparison of trends over time.
For most reservoirs, the amount of available general water quality parameter data is variable (see
Appendix Table E-2 for an example). Average pH levels range from 5.87 to 8.19, with typically
lower values for shales. Larger variations in average specific conductivity are seen among
unconventional reservoirs and range from 213 microsiemens (|j.S)/cm in the Bakken Shale to
184,800 |j.S/cm in Devonian sandstones (Appendix Table E-2). Shale and tight formation produced
waters are enriched in suspended solids, as reported concentrations for total suspended solids and
turbidity exceed those of coalbeds by one to two orders of magnitude.
The average dissolved oxygen (DO) concentrations of CBM produced water range from 0.39-1.07
mg/L (Appendix Table E-3). By comparison, well-oxygenated surface water can contain up to 10
mg/L DO at 59 °F (15 °C) fU.S. EPA. 2012al. Thus, coalbed produced water is either hypoxic (less
than 2 mg/L DO) or anoxic (less than 0.5 mg/L DO) and, if released to surface waters, could
contribute to aquatic organism stress (USGS. 2010: NSTC. 2000).
7.3.4.4 Salinity and Inorganics
The TDS profile of produced water from unconventional reservoirs is dominated by sodium and
chloride, with large contributions to the profile from mono- and divalent cations (Sun etal.. 2013:
Guerra etal.. 2011). Shale and sandstone produced waters tend to be characterized as sodium-
chloride-calcium water types, whereas CBM produced water tends to be characterized as sodium
chloride or sodium bicarbonate water types fDahm etal.. 20111. Elevated levels of bromide, sulfate,
and bicarbonate are also present (Sun etal.. 2013). Elevated strontium and barium levels are
characteristic of Marcellus Shale produced water (Barbot etal.. 2013: Haluszczaketal.. 2013:
Chapman etal.. 20121. Data representing shales and tight formations are presented in Appendix
Table E-4.
Marcellus Shale produced water salinities range from less than 1,500 mg/L to over 300,000 mg/L,
as shown by Rowan etal. (2011). By comparison, the average salinity concentration for seawater is
35,000 mg/L.
Of the CBM data presented in Appendix Table E-5, differences are evident between the Black
Warrior and the three western formations (Powder River, Raton, and San Juan). The Black Warrior
is higher in average chloride, specific conductivity, TDS, TOC, and total suspended solids, and lower
in alkalinity and bicarbonate than the other three. These differences are due to the saline or
brackish conditions during deposition in the Black Warrior, and its older geologic age that contrasts
with the freshwater conditions for the younger western basins. The TDS concentration of CBM
7-19
-------
Chapter 7 - Produced Water Handling
produced water can range from 170 mg/L to nearly 43,000 mg/L (range composited from Dahm et
al. f20111 and Benko and Drewes f20081: see also Van Voast f^OOSll.1
7.3.4.5 Metals
The metals content of produced water from unconventional reservoirs varies by well and site
lithology. Levels of iron, magnesium, and boron were within ranges known for conventional
produced water fHaves. 20091. Produced water from unconventional reservoirs may also contain
low levels of heavy metals (e.g., chromium, copper, nickel, zinc, cadmium, lead, arsenic, and
mercury as found by Hayes). Data illustrating metal concentrations in produced water appear in
Appendix Tables E-6 and E-7.
7.3.4.6 Naturally Occurring Radioactive Material (NORM) and Technologically Enhanced
Naturally Occurring Radioactive Material (TENORM)
Geologic environments contain naturally occurring radioactive material (NORM). Radioactive
materials commonly present in shale and sandstone sedimentary environments include uranium,
thorium, radium, and their decay products. Elevated formation uranium levels have been used to
identify potential areas of natural gas production for decades fFertl and Chilingar. 19881. Shales
that contain significant levels of uranium include the Barnett in Texas, the Woodford in Oklahoma,
the New Albany in the Illinois Basin, the Chattanooga Shale in the southeastern United States, and a
group of black shales in Kansas and Oklahoma (Swanson. 19551.2 When exposed to the
environment in produced water, NORM is called technologically enhanced naturally occurring
radioactive material (TENORM).3 Water soluble forms of TENORM are present in most produced
water from unconventional reservoirs, but particularly so in Marcellus Shale produced water
fRowan etal.. 2011: Fisher. 19981.
Due to insolubility under prevailing reducing conditions encountered within shale formations, only
low levels of uranium and thorium are found in produced water, typically in the concentrated form
of mineral phases or organic matter (Nelson etal.. 2014: Sturchio etal.. 20011. Conversely, radium,
a decay product of uranium and thorium, is known to be relatively soluble within the redox range
encountered in subsurface environments fSturchio etal.. 2001: Langmuir and Riese. 19851. As
noted in Section 7.3.3.2, radium and TDS produced water concentrations are positively correlated
(Rowan etal.. 2011: Fisher. 1998): therefore, in formations containing radium, increasing TDS
concentration indicates likely increasing radium concentration.
1 From a similar dataset, Dahm etal. (2011) report TDS concentrations from a composite CBM produced water database
(n = 3,255] for western basins that often are less than 5,000 mg/L (85% of samples].
2 Marine black shales are estimated to contain an average of 15-60 ppm uranium depending on depositional conditions
fFertl and Chilingar. 19881
3 The U.S. EPA Office of Air and Radiation's website fhttps://www.epa.gov/radiation/technologically-enhanced-naturally-
occurring-radioactive-materials-tenorm] states that TENORM is produced when activities such as uranium mining or
sewage sludge treatment concentrate or expose radioactive materials that occur naturally in ores, soils, water, or other
natural materials. Formation water containing radioactive materials contains NORM, because it is not exposed; produced
water contains TENORM, because it has been exposed to the environment.
7-20
-------
Chapter 7 - Produced Water Handling
Median values of total radium in the Marcellus Shale ranged from about 1,000 pCi/L to less than
6,000 pCi/L, which are values far exceeding the industrial discharge limit of 60 pCi/L fRowan etal..
20111 (Figure 7-6). In the Marcellus Shale, TENORM levels in produced water from unconventional
reservoirs exceeded levels from conventional reservoirs levels by factors of 4 to 26 fPA DEP.
2015b) (Appendix Table E-8). The individual median concentrations in produced water from
unconventional reservoirs of 11,300 pCi/L gross alpha, 3,445 pCi/L gross beta, and total radium of
7,180 pCi/L (Appendix Table E-8). TENORM has been identified in hydraulic fracturing fluid,
presumably due to the reuse of produced water at levels from 2 to 4.5 times lower than produced
water from unconventional reservoirs (PADEP. 2015bl (AppendixTable E-8).
100,000
^ 10,000
QJ
H—'
1
0)
a.
1,000
1 '
ZD
o
o
o
'd.
¦E„ 100
CO
CN1
CM
CD
cc
Is 10
o
CD
cc
1
0
Figure 7-6. Data on radium 226 (open symbols) and total radium (filled symbols) for Marcellus
Shale wells (leftmost three columns) and other formations (rightmost three columns).
Source: Rowan et al. (2011). The dashed line represents the industrial effluent discharge limit of 60 pCi/L set by the
Nuclear Regulatory Commission. The black lines indicate the median concentrations, and the number of points in
each dataset are shown in parentheses. Citations within the figure are provided in Rowan et al. (2011).
7.3.4.7 Organics
The organic content of produced water varies by well and lithology, but consists of naturally
occurring and injected organic compounds (Lee and Neff. 2011). Of the organics detected by either
routine or advanced analytical methods (Section 7.3.1), the majority are naturally occurring
constituents of petroleum (Appendix Tables H-4 and H-5). These organics may be dissolved in
water or, in the case of oil production, in the form of a separate or emulsified phase. Several classes
of organic chemicals have been found in shale gas and CBM produced water, including aromatics,
O
PADEP (2009-2010,
unpub. data)
(25)
O
.
Marcellus Shale data
NYSDEC (2009)
(13)
this study
(14)
A ~
NYSDEC
(Gilday
and others, 1999)
(48)
* ^
PA DEP (1992)
(37)
~
~
~
Dresel and Rose
(2010)
(61
~
7-21
-------
Chapter 7 - Produced Water Handling
polyaromatic hydrocarbons, heterocyclic compounds, aromatic amines, phenols, phthalates,
aliphatic alcohols, fatty acids, and nonaromatic compounds (list from Orem etal. f20141. see also:
Hayes f20091. Benko and Drewes f20081. Orem etal. f20071. and Sirivedhin and Dallbauman
(200411. Compounds found in CBM waters included pyrene, phenanthrenone, alkyl phthalates, C12
through Cis fatty acids, and others. Similarly, compounds found in shale gas produced water
included pyrene and perylene, ethylene glycol, diethylene glycol monodocecyl ether, 2-(2-
butoxyethoxy) ethanol, and others fOrem etal.. 20141. Biomarkers—organic molecules
characteristically produced by life forms, and unique to shale formations—have recently been
suggested to fingerprint produced water (Hoelzer et al.. 2016). More representative examples from
five coal bed and two shale gas formations with reported concentrations are given in Appendix
Tables E-9, E-ll, and E-12, and the complete list of chemicals with CAS registry numbers identified
by the EPA for this assessment appears in Appendix H. (See Appendix Table H-4 for chemicals with
EPA-identified CAS numbers and Appendix Table H-5 for chemicals without) Appendix Table E-13
lists concentrations of organic chemicals that were identified in three specific studies (Khan etal..
2016: Lester etal.. 2015: Orem etal.. 20071.
7.3.4.8 Hydraulic Fracturing Fluid Additives
Several chemicals used in hydraulic fracturing fluids have been identified in produced water.
(Examples are shown in Table 7-6, Appendix Table E-10, and Appendix Tables H-4 and H-5.) Many
of these chemicals were identified through advanced analytical procedures and equipment, and
would not be expected to be found by routine analyses. Of note is that phthalates do not occur
naturally. Their presence in produced water is due to either their use in hydraulic fracturing fluids;
polyvinyl chloride (PVC) in well adhesives, valves, or fittings; or coatings on laboratory sample
bottles (Orem etal.. 2007).1 Phthalates can also be used in drilling fluids, as breaker additives, or as
plasticizers fMaguire-Bovle and Barron. 2014: Hayes and Severin. 2012a1.2 One of the produced
water phthalates has been identified as a component of hydraulic fracturing fluid (di(2-ethylhexyl)
phthalate) (Appendix Table H-2), while others have not, and those may originate from laboratory or
field equipment
Table 7-6. Examples of compounds identified in produced water that can be components of
hydraulic fracturing fluid.
Appendix Tables H-4 and H-5 list chemicals identified in produced water and indicates those also identified as
constituents of hydraulic fracturing fluid. Chemical or class designation in this table is taken directly from the text
of the cited references except where noted, and may or may not reflect the chemical names from the Distributed
Structure-Searchable Toxicity Database (DSSTox) show in Appendix Table H-4 or other chemicals listed in Appendix
Table H-5.
Chemical or class
Use
Reference
2-Butanone
Solvent; microbial degradation
product
Lester et al. (2015)
1 Examples include di(2-ethylhexyl] phthalate, diisodecyl phthalate, and diisononyl phthalate fOrem et al.. 20071
2 Specifically fatty acid phthalate esters fMaguire-Bovle and Barron. 20141
7-22
-------
Chapter 7 - Produced Water Handling
Chemical or class
Use
Reference
2-Butoxyethanol
Acid dispersant, solvent, non-
emulsifier
Thacker et al. (2015)
Acetone
Solvent; microbial degradation
product
Lester et al. (2015)
Cocamidopropyl dimethylamine
(C-7)
Foaming and lubrication enhancer
Lester et al. (2015)
Di(2-ethylhexyl) phthalate3
Derivative of polyvinyl chloride used
in adhesives, valves, fittings or
coatings of sample bottles
Orem et al. (2007)
Diethylene glycol monododecyl
ether
Antifreeze, scale inhibitor, friction
reducer
Orem et al. (2014)
Dioctadecyl ester of phosphate
phosphoric acid
Common lubricant
Maguire-Bovle and Barron (2014)
Ethylene glycol
Antifreeze, scale inhibitor, friction
reducer
Orem et al. (2014)
Fatty acid phthalate esters
(Related to) use in drilling fluids and
breakers
Maguire-Bovle and Barron (2014)
Fluorocarbons
Tracers
Maguire-Bovle and Barron (2014)
Hexahydro-l,3,5-trimethyl-l,3,5-
triazine-2-thione
Biocide
Orem et al. (2014)
Linear alkyl ethoxylates (C-4 to C-8,
C-ll to C-14)
Enhancer of surfactant properties
Lester et al. (2015); Thurman et al.
(2014)
Polyethylene glycol carboxylates
(PEG-C-E02 to PEG-C-EO10)
Friction reducer, clay stabilizer,
surfactants
Thurman et al. (2016)
Polyethylene glycols (PEG-E04 to
PEG-EOIO)
Friction reducer, clay stabilizer,
surfactants
Thurman et al. (2016)
Polypropylene glycols (PPG-P02 to
PPG PO10)
Friction reducer, clay stabilizer,
surfactants
Thurman et al. (2016)
Toluene
Solvent, scale inhibitor
Thacker et al. (2015)
Triethylene glycol monododecyl
ether
Antifreeze, scale inhibitor, friction
reducer
Orem et al. (2014)
Xylenes
Solvent, scale inhibitor
Thacker et al. (2015)
a Di(2-ethylhexyl) phthalate was named di-2-ethyl hexyl phthalate in Maguire-Boyle and Barron (2014).
7-23
-------
Chapter 7 - Produced Water Handling
7.3.4.9 Reactions within Formations
The introduction of hydraulic fracturing fluids into the target formation induces a number of
changes to formation solids and fluids that influence the chemical evolution and composition of
produced water. These changes can result from physical processes (e.g., rock fracturing and fluid
mixing); geochemical processes (e.g., introducing oxygenated fluids of composition unlike that of
the formation); and down hole conditions (elevated temperature, salinity, and pressure) that
mobilize trace or major constituents into solution.
The creation of fractures exposes new formation surfaces to interactions involving hydraulic
fracturing fluids and existing formation fluids. Formations in unconventional reservoirs targeted
for development are composed of detrital, cement, and organic fractions. For example, elements
potentially available for mobilization when exposed via fracturing include calcium, magnesium,
manganese, and strontium in cement fractions, and silver, chromium, copper, molybdenum,
niobium, vanadium, and zinc in organic fractions.
From organic compounds identified in five flowback samples and one produced water sample from
the Fayetteville Shale, three possible types of reactions were identified by Hoelzeretal. (2016):
hydrolysis of delayed acids, oxidant-caused halogenation reactions, and transformation of disclosed
additives. First, delayed acids are used to "break" gel structures and would be intentionally
introduced for their ability to cause in-formation reactions. Second, strong oxidants or other
compounds introduced as breakers, along with elevated temperature and salinity, can trigger
reactions between halogens (chloride, bromide, and iodide) and methane, acetone and pyrane
resulting in halomethane compounds. A similar suggestion was made by Maguire-Bovle and Barron
(2014). Low pH was found to promote oxidation of additives (Tasker etal.. 2016). Third, known
additives may react to form byproducts. Hoelzer etal. f20161 postulate examples from several
types of compounds, two of these are the formation of benzyl alcohol from the hydraulic fracturing
additive benzyl chloride, and abiotic and biotic reactions of phenols. In a study that used synthetic
fracturing fluid, Tasker etal. (2016) reported that surfactants were recalcitrant to degradation
under high pressure and temperature, which may explain the presence of the surfactant glycols in
produced water as reported by Thurman et al. f 20161 (Table 7-6), and the oxidation of other
additives (gelling and some friction reducers (Table 5-1)) may explain their absence.
7.3.5 Spatial Trends in Produced Water Composition
As was reported for the volume of produced water (Section 7.2.2), the composition of produced
water varies spatially on a regional to local scale according to the geographic and stratigraphic
locations of each well within a hydraulically fractured production zone (Bibbv etal.. 2013: Lee and
Neff. 2011). Spatial variability of produced water content occurs: (1) between plays of different
rock sources (e.g., coal vs. sandstone); (2) between plays of the same rock type (e.g., Barnett Shale
vs. Bakken Shale); and (3) within formations of the same source rock (e.g., northeastern vs.
southwestern Marcellus Shale) (Barbotetal.. 2013: Alley etal.. 2011: Breit. 2002).
Geographic variability in produced water content has been established at a regional scale for
conventional produced water. As an example, Benko and Drewes f20081 demonstrate TDS
7-24
-------
Chapter 7 - Produced Water Handling
variability in conventional produced water among fourteen western geologic basins (e.g., Williston,
San Juan, and Permian Basins). Median TDS in these basins range from as low as 4,900 mg/L in the
Big Horn Basin to as high as 132,400 mg/L in the Williston Basin based on over 133,000 produced
water samples from fourteen basins (Benko and Drewes. 2008).1
Average or median TDS of more than 100,000 mg/L has been reported for the Bakken (North
Dakota, Montana) and Marcellus (Pennsylvania) formations; between 50,000 mg/L and 100,000
mg/L for the Barnett (Texas), and less than 50,000 mg/L for the Fayetteville (Arkansas) shale
formations.2 In tight formations, the average TDS was above 100,000 mg/L for the Devonian
Sandstone (Pennsylvania) and Cotton Valley Group (Louisiana, Texas), between 50,000 mg/L and
100.000 mg/L for the Oswego (Oklahoma), and less than 50,000 mg/L for the Mesaverde
Formation (Colorado, New Mexico, Utah, Wyoming). Maximum concentrations above 200,000 mg/L
have been reported for the Marcellus, Bakken, Cotton Valley Group and Devonian Sandstone
(Appendix Table E-2).
CBM produced waters had average TDS of less than 5,000 mg/L in the Powder River (Montana,
Wyoming), Raton (Colorado, New Mexico), and San Juan (Arizona, Colorado, New Mexico, Utah)
basins; while above 10,000 mg/L in the Black Warrior Basin (Alabama, Mississippi), which as noted
above are due to the depositional history of these basins (Appendix Table E-3, Section 7.3.2).
Data further illustrating variability within both shale and tight gas reservoirs, as well as coalbed
methane fields, at both the formation and local scales are presented and discussed in Appendix
Section E.3.
7.4 Spill and Release Impacts on Drinking Water Resources
Surface spills of produced water from oil and gas production have occurred across the country and,
in some cases, have caused impacts to drinking water resources. Released fluids can flow into
nearby surface waters, if not contained on-site, or infiltrate into groundwater via soil. In this
section, we first briefly describe the potential for spills from produced water handling equipment
Next, we address individually reported spill events. These have originated from pipeline leaks, well
blowouts, well communication events, and leaking pits and impoundments. We then summarize
several studies of aggregated spill data, which are based on state agency spill reports.
7.4.1 Produced Water Handling and Spill Potential
Throughout the production phase at oil and certain wet gas production facilities, produced water is
stored in containers and pits that can contain free phase, dissolved phase, and emulsified crude oil.
Since the crude oil is not efficiently separated out by the flow-through process vessels (such as
1 Data were drawn from the USGS National Produced Water Geochemical Database v2.0. Published updates made in
October 2014 to the database (v2.1] are not reflected in this document.
2 Because publications we are comparing may report either average or median values (but not uniformly both], we
combine average and medians in this paragraph.
7-25
-------
Chapter 7 - Produced Water Handling
three-phase separators, heater treaters, or gun barrels), this crude oil can remain present in the
produced water container or pit.
Produced water can be transferred to surface pits for long-term storage and evaporation. Surface
pits are typically uncovered, earthen pits that may or may not be lined.1 Unlined pits can lead to
contamination of groundwater, especially shallow alluvial systems. Recovered fluids can overflow
or leak from surface pits due to improper pit design and weather events.
Produced water that is to be treated or disposed of off-site is typically stored in storage tanks or
pits until it can be loaded into transport trucks for removal (Gilmore etal.. 20131. Tank storage
systems are typically closed loop systems in which produced water is transported from the
wellhead to aboveground storage tanks through interconnecting pipelines fGWPC and IOGCC.
2014). Failure of connections and lines during the transfer process or the failure of a storage tank
can result in a surface release of fluids.
Depending on its characteristics, produced water can be recycled and reused on-site. It can be
directly reused without treatment (after blending with freshwater), or it can be treated on-site
prior to reuse (Boschee. 2014). As with other produced water management options, these systems
also can spill during transfer of fluids.
7.4.2 Spills of Produced Water
7.4.2.1 Pipeline Leaks
Produced water is typically transported from the wellhead through a series of pipes or flowlines to
on-site storage or treatment units fGWPC and IOGCC. 20141. or nearby injection wells. Faulty
connections at either end of the transfer process or leaks or ruptures in the lines carrying the fluid
can result in surface spills. A field report from PA PEP (2009b) described a leak from a 90-degree
bend in an overland pipe carrying a mixture of produced water and freshwater between two pits.
The impact included a "dull sheen" on the water and measured chloride concentration of 11,000
mg/L. The leak impacted a 0.4 mi (0.6 km) length of a stream, and fish and salamanders were killed.
Beyond a confluence at 0.4 mi (0.6 km) with a creek, no additional dead fish were found. The
release was estimated at 11,000 gal (42,000 L). In response to the incident, the pipeline was shut
off, a dam was constructed for recovering the water, water was vacuumed from the stream, and the
stream was flushed with fresh water fPADEP. 2009bl.
Another example of a pipeline release occurred in January 2015, when 70,000 bbls (2,940,000 gal
or 11,130,000 L) of produced water containing petroleum hydrocarbons (North Dakota
Department of Health. 20151 were released from a broken pipeline that crosses Blacktail Creek in
Williams County, ND. The response included placing absorbent booms in the creek, excavating
contaminated soil, removing oil-coated ice, and removing produced water from the creek. The
electrical conductivity and chloride concentration in the water along the creek, the Little Muddy
River, and Missouri River were found to be elevated above background levels, as were samples
1 The use of the terms "impoundments" and "pits" varies and is described in Chapter 8. For the purposes of this section,
the term "pits" will be generally used to cover all below-grade storage (but not above ground closed or open tanks].
7-26
-------
Chapter 7 - Produced Water Handling
taken from groundwater recovery trenches. Remediation work on this site continues as of the date
of this writing (August, 2016).
7.4.2.2 Well Blowouts
Spills of produced water have occurred as a result of well blowouts. Fingerprinting of water from
two monitoring wells in Killdeer, ND, was used to determine that brine contamination in the two
wells resulted from a well blowout during a hydraulic fracturing operation. See the discussion in
Section 6.2.2.1 for more information.
Another example of a well blowout associated with a hydraulic fracturing operation occurred in
Clearfield County, PA. The well blew out, resulting in an uncontrolled flow of approximately
35,000 gal (132,000 L) of brine and fracturing fluid; some of the liquids reportedly reached a
nearby stream (Barnes, 2010], The blowout occurred during drilling of plugs that were used to
isolate fracture stages from each other. An independent investigation found that the primary cause
of the incident was that the sole blowout preventer on the well had not been properly tested. In
addition, the company did not have certified well control experts on hand or a written pressure
control procedure (Vittitow, 2010").
In North Dakota, a blowout preventer failed, causing a release of between 50 and 70 bbls per day
(2,100 gal/day or 7,900 L/day and 2,940 gal/day or 11,100 L/day) of produced water and oil
(Reuters, 2014"). Frozen droplets of oil and water sprayed on a nearby frozen creek. Liquid flowing
from the well was collected and trucked offsite. A 3-ft (0.9-m) berm was placed around the well for
containment. Multiple well communication events have also led to produced water spills ranging
from around 700 to 35,000 gal (2,600 L to 130,000 L) (Vaidyanathan, 2013a"). Well communication
is described in Section 6.3.2.3.
The Chesapeake Energy ATGAS 2H well, located in Leroy Township, Bradford County, PA,
experienced a wellhead flange failure on April 19, 2011, during hydraulic fracturing operations.
Approximately 10,000 gal (38,000 L) of produced water spilled into an unnamed tributaiy of
Towanda Creek, a state-designated trout stock fishery and a tributary of the Susquehanna River
(USGS, 2013b; SAIC and GES, 2011], Chesapeake conducted post-spill surface water and
groundwater monitoring (SAIC and GES, 2011").
Chesapeake concluded that there were short-term impacts to surface waters of a farm pond within
the vicinity of the well pad, the unnamed tributary, and Towanda Creek following the event (SAIC
and GES, 2011"). The lower 500 ft (200 m) of the unnamed tributary exhibited elevated chloride,
TDS, and specific conductance, which returned to background levels in less than a week. Towanda
Creek experienced these same elevations in concentration, but only at its confluence with the
unnamed tributary; elevated chloride, TDS, and specific conductance returned to background levels
the day after the blowout (SAIC and GES, 2011").
7.4.2.3 Leaks from Pits and Impoundments
Leaks of produced water from on-site pits have caused releases as large as 57,000 gal (220,000 L)
and have caused surface water and groundwater impacts (Vaidyanathan, 2013b; Levis,
7-27
-------
Chapter 7 - Produced Water Handling
2011: 2010c: PADEP, 2010]. VOCs have been measured in groundwater near the Duncan Oil Field
in New Mexico downgradient of an unlined pit storing produced water. More example releases
from pits are described in Section 8.4.5.
Two of EPA's retrospective case studies evaluated potential impacts from produced water pits. The
EPA retrospective case studies were designed to determine whether multiple lines of evidence
might be found that could specifically link constituent(s) found in drinking water to hydraulic
fracturing activities using the tiered assessment framework presented in Appendix Section E.6. A
multiple-lines-of-evidence approach was used to evaluate potential cause-and-effect relationships
between hydraulic fracturing activities and contaminant presence in groundwater. Such an
approach is needed, because the presence of a constituent in groundwater that is also found in
hydraulic fracturing fluids or produced water does not necessarily implicate hydraulic fracturing
activities as the cause. This is because some constituents of hydraulic fracturing fluids or produced
water are ubiquitous in society (i.e., BTEX), and some constituents of produced water can be
present in groundwater as background constituents (i.e., methane, iron, and manganese).
Elements of the assessment framework include gathering background information, including pre-
drilling sample results; developing a conceptual model of the site; and assessing multiple analytes
to develop lines of evidence. Development of these requires adherence to sampling and quality
assurance protocols to generate defensible data. Among many other quality assurance
requirements, proper well purging and analyses of field and laboratory blanks are needed
(Appendix Table E-17 and Figure E-15).
In the EPA's Retrospective Case Study in Southwestern Pennsylvania: Study of the Potential Impacts of
Hydraulic Fracturing on Drinking Water Resources (U.S. EPA. 2 015i1. elevated chloride
concentrations and their timing relative to historical data suggested a recent groundwater impact
on a private water well occurred near a pit The water quality trends suggested that the chloride
anomaly was related to the pit, but site-specific data were not available to provide a definitive
assessment of the cause (s) and the longevity of the impact Evaluation of other water quality
parameters did not provide clear evidence of produced water impacts.
In the EPA's Retrospective Case Study in Wise County\ Texas: Study of the Potential Impacts of
Hydraulic Fracturing on Drinking Water Resources fU.S. EPA. 201511. impacts to two water wells
were attributed to brine, but the data collected for the study were not sufficient to distinguish
among multiple possible brine sources, including reserve pits, migration from underlying
formations along wellbores, migration from underlying formation along natural fractures and a
nearby brine injection well.
To aid in assessing impacts, a number of geochemical indicators and isotopic tracers for identifying
oil and gas produced water have been identified. These include (Lauer etal., 2016; Warner etal.,
2014a. hi:
• Common ion ratios, including bromide/chloride and lithium/chloride;
• Isotope ratios, especially Strontium isotope ratios (87Sr/86Sr); and
7-28
-------
Chapter 7 - Produced Water Handling
• Enrichment of certain isotopes: S180, S2H, S7Li, S13C-DIC, S11!}.1
For the case study, twelve geochemical indicators, including the bromine/chlorine (Br/CI) and
strontium isotope ratios, were considered for the well-water samples.2 The results were used to
assess whether the likelihood that the observed values originated with produced water (the
aforementioned sources of brine), sea water, road salt, landfill leachate, sewage/septic tank
leachate, and animal waste. In each sample evaluated, it was found that the water could have
originated with one or more of the six sources. Thus these lines of evidence did not allow
identification of neither a specific source nor a hydraulic fracturing source (Appendix Table E-18).
A third well experienced similar impacts, and a landfill leachate source could not be ruled out in
that case.
The case studies illustrate how multiple lines of evidence were needed to assess suspected impacts
and that no single constituent or parameter could be used alone to assess potential impacts.
7.4.2.4 Other Sources
In the EPA's Retrospective Case Study in Northeastern Pennsylvania: Study of the Potential Impacts of
Hydraulic Fracturing on Drinking Water Resources fU.S. EPA. 2014fl a pond was found to be
impacted due to elevated chloride and TDS, along with strontium ratios (87Sr/86Sr) characteristic of
Marcellus Shale produced water. Here, the suspected source of the impact was a well pad which had
a hydrochloric acid spill, a possible produced water spill and been used for temporary storage of
drill cuttings. The same mulidence fracturing impacts from constituents characteristic of produced
water (TDS, chloride, sodium, barium, strontium and radium) found in three domestic wells located
in an area with naturally occurring saline groundwater. Conversely, at a spring with organic
chemical contamination but no associated chloride or TDS impacts, hydraulic fracturing activities
were also ruled out.
An estimated 6,300 to 57,373 gal (24,000 to 217,280 L) of Marcellus Shale produced water was
discharged through an open valve that drained a tank at XTO Energy Inc.'s Marquardt pad and
flowed into a tributary of the Susquehanna River in November 2010 fU.S. EPA. 2016e: PA DEP.
20Hc). Overland and subsurface flow of released fluids impacted surface water, a subsurface
spring, and soil. Five hundred tons of contaminated soil were excavated, and an estimated 8,000 gal
(30,000 L) of produced water was recovered (Science Applications International Corporation.
20101. Elevated levels of TDS, chloride, bromide, barium and strontium that indicated a release of
produced water were present in the surface stream and a spring for roughly 65 days fU.S. EPA.
2016e). At that time the chloride concentration in the spring dropped below the state surface water
standard of 250 mg/L. The impact extended a distance of approximately 1,400 ft (440 m) to the
spring from the release point Samples were taken in the tributary roughly 500 ft downstream from
the spring, where chloride concentrations remained below the 250 mg/L standard throughout the
sampling period, but were above the upstream concentrations fPA DEP. 2011c: Schmidlev and
Smith. 2011). Similarly, the total barium, total and dissolved iron, manganese and alkalinity
concentrations remained below the Pennsylvania surface water quality standards at the
downstream monitoring location throughout the monitoring period fSchmidlev and Smith. 20111.
1 DIC is dissolved inorganic carbon.
2 The full list was: Br vs. B, CI vs. Mg, CI vs. Br, CI vs. HCCh.Cl vs. Ca, CI vs. K, CI vs. Na, CI vs. SO4, CI/Br, Cl/I, K/Rb, 87Sr/86Sr.
7-29
-------
Chapter 7 - Produced Water Handling
In Pennsylvania, discharges of brine were made into a storm drain that itself discharges to a
tributary of the Mahoning River in Ohio. Analyses of the brine and drill cuttings that were
discharged indicated the presence of contaminants, including benzene and toluene fU.S.
Department of justice. 20141. In California, an oil production company periodically discharged
hydraulic fracturing wastewaters to an unlined sump for 12 days. It was concluded by the
prosecution that the discharge posed a threat to groundwater quality (Bacher. 20131. These
unauthorized discharges represent both documented and potential impacts on drinking water
resources. However, data do not exist to evaluate whether such episodes are uncommon or
whether they happen on a more frequent basis and remain largely undetected. Other cases of
unpermitted discharges have been reported by various sources (Caniglia. 2014: Paterra. 20111.1
7.4.2.5 Data Compilation Studies
Three datasets were examined for produced water spill data. These included two published studies:
a review of spills in Oklahoma that occurred prior to the onset of widespread high-volume
hydraulic fracturing (Fisher and Sublette. 20051. and an EPA study of spills occurring between
February 2006 and April 2012 on the well pads of hydraulically fractured wells fU.S. EPA. 2015ml.
The EPA spills study, Review of state and industry spill data: characterization of hydraulic fracturing-
related spills, is described in Text Box 5-10. Because of data availability, EPA's study was dominated
by data from Pennsylvania (21% of releases) and Colorado (48% of releases). Several difficulties
are encountered in compiling and evaluating data on produced water spills and releases. Because
states have differing minimum reporting levels, more spills are potentially reported in states with
lower reporting limits.2
To include data from another state and to give results current to 2015, data from North Dakota
were reviewed for this assessment.3 Details on the procedures and results for non-produced water
spills are given in Appendix Section E.5. The North Dakota Department of Health (NDDOH) collects
data on environmental incidents and separately compiles oil field incidents; information is made
available to the public at http: //www.ndhealth.gov/EHS/Spills/. Of these incidents, most describe a
release of oil, saltwater, or other liquid. Of the remainder, a few describe releases of gas only.
For the period from November 2012 to November 2013, NDDOH reported 552 releases of produced
water that were retained within the boundaries of the production or exploration facility and 104
that were not (North Dakota Department of Health. 20111. Thus, 16% of the releases were not
contained within facility boundaries and had greater potential for impacting drinking water
resources.
1 Section 8.4 discusses permitted discharges of wastewater.
2 For example, two agencies in the state of California manage different databases that both store information on spills
associated with oil and gas production fCCST. 2015al CCST f2015al reported that the databases contain inconsistencies
as to the number of spills and the details regarding those spills (e.g., quantity, chemical composition ofthe wastewater]
resulting in uncertainty on the impacts spills have on the environment.
3 Wirfs-Brock ("20151 presented an analysis of North Dakota spill data through 2013.
7-30
-------
Chapter 7 - Produced Water Handling
7.4.2.6 Frequency of Spills and Releases
The EPA analyzed these data and found that, in recentyears (2010-2015), there were between five
and seven produced water spills per hundred active production wells (Figure 7-7). Spills declined
between 2014 and 2015 (from 846 to 609), although the number of production wells increased. A
study of 17 states indicated that there was an overall reduction of 8% in spills from 2014 to 2015,
and an increase of 9% in Texas fKing and Soraghan. 20161. More details on the data analysis are
given in Appendix Section E.5, which includes results on North Dakota oil and spills of other types,
including hydraulic fracturing fluids (as noted in Chapter 5).
12
10
5 to 7 salt water
spills per 100
producing wells
-Salt Water (SW)
i-Other Spills
I- Oil Spills
8
6
4
Q-
l/l
2
0
2000
2002
2004
2006
2008
Year
2010
2012
2014
2016
Figure 7-7. Produced water spill rates (spills per active wells) for North Dakota from 2001 to
2015 (Appendix Section E.5).
7.4.2.7 Produced Water Releases—Causes and Sources
The causes and sources identified for releases vary among the three datasets reviewed. North
Dakota releases were dominated by leaks from various pieces of equipment, followed by "others,"
and various overflows (Figure 7-8). While the release rate declined from 2014 to 2015, the causes
remained ranked relatively in the same order; notably fewer releases were attributed to "other"
and more to equipment failure in 2015. The EPA's spills study found on- or near-well pad releases
to be dominated by human error, unknown, and equipment failure (U.S. EPA. 2015m). The earlier
7-31
-------
Chapter 7 - Produced Water Handling
Oklahoma study was dominated by overflows, unpermitted discharges, and storms (Figure 7-9}.1
Storms can cause releases, as was noted after a major flood in northeastern Colorado that caused
damage to produced water storage tanks releasing an estimated 43,000 gal (160,000 L) of
produced water (COGCC. 2013).
The sources of releases are documented for the Oklahoma and EPA studies (Figure 7-10). The EPA
cites storage, unknown, and hoses or lines as the major sources for its 225 well-pad releases. The
earlier Oklahoma study cites unclassified, lines, and tanks as major sources of its 8,874 releases.
250
\A
-------
Chapter 7 - Produced Water Handling
Oklahoma (1993-2003)
981 Releases
US EPA
225 Releases
Other
3%
Container
integrity
v 13%
Equipment
failure
17%
Human error
38%
Unknown
29%
Figure 7-9. Distribution of spill causes in Oklahoma, pre-high volume hydraulic fracturing
years of 1993-2003 (left) and in the EPA study of spills on production pads (right).
Data sources: left, Fisher and Sublette (2005); right, U.S. EPA (2015m).
Oklahoma (1993-2003)
8874 Releases
US EPA
225 Releases
Surface pj.
equipment O0/
7% Z/o
unc assified
Tanks
21%
Lines
27%
Well or wellhead
Equipment 5%
5V^1
Hose\ \
W or line \ \
L_12% \\
Storage
Unknown /
58%
J
Figure 7-10. Distribution of spill sources in Oklahoma, pre-high volume hydraulic fracturing
years of 1993-2003 (left) and in the EPA study of spills on production pads (right).
Data sources: left, Fisher and Sublette (2005); right, U.S. EPA (2015m).
Overflows
49%
Corrosion
4%
Accidents
8%
Storms
13%
Unpermitted
discharges
18%
7-33
-------
Chapter 7 - Produced Water Handling
7.4.2.8 The Volumes of Spilled Produced Water
The 2015 North Dakota spills were ranked from by the median volume, which is the level at which
50% of the spills are below this volume and 50% above (Figure 7-11).1 Of the North Dakota spills in
2015, the highest median spill volume was caused by a blowout (2,400 gal, 91,000 L, left-most red
box). The smallest median volume spill is approximately 10 times lower in volume (84 gal, 320 L).
Spills larger than the median are of interest, because of their potential for impacting drinking water
resources. The largest volume spill occurred from a pipeline break (2,900,000 gal, 11,000,000 L).
The EPA spills study found the highest median volume spill was from equipment failure (1,700 gal,
6400 L), while the highest volume spill was due to container integrity (1,300,000 gal, 4,900,000 L)
(Figure 7-12).
10,000,0001
1,000,0001
100,0001
10,0001
1,0001
1001
10I
1-
(/>
c
o
fc
3S
OJ
E
_3
O
>
QJ
Figure 7-11. Volumes of 2015 North Dakota salt water releases by cause (leftmost 13 boxes in
red), and all causes (last box in blue).
1 These figures are called "box" plots or "box and whisker" plots. The rectangle in the middle represents the range of data
from the 25th to 75fh percentile. The line across the box represents the 50th percentile, also known as the median. Fifty
percent ofthe data are below the median. The lines extending above and belowthe boxes represent the range of data
from minimum to maximum. These concepts are illustrated in Appendix Figure E-6.
7-34
-------
Chapter 7 - Produced Water Handling
in
C
_o
15
go
ai
£
>
a>
w
ro
QJ
CC
10,000,000 E
1,000,000 !
100,000
10,000
1,000
100
10
1 -I
II
X.
/
&
/
Figure 7-12. Volumes of produced water spills reported by the EPA for 2006 to 2012 by cause
(the five left most boxes in red), source (the second five boxes in yellow), and all spills (blue).
Calculated from Appendix B of U.S. EPA (2015m).
From the analyses, half of the spills are less than 1,000 gal (3,800 L) (EPA) and 340 gal (1,300 L)
(North Dakota) (Figure 7-12, Figure 7-13, and medians in Table 7-7). The medians for the
Oklahoma study were higher (overall 1,700 gal or 6,400 L; see Table 7-7 for yearly values) (Fisher
and Sublette. 2005). These occurred in a different state and over an earlier time period, so a direct
connection with the recent North Dakota and EPA results has not been made.
The skewed nature of the distributions are noted by the mean values being considerably higher
than these medians (see Figure 7-13). In each case, this is caused by a small number of large spills.
For 2015 in North Dakota, for example, there were 12 releases of 21,000 gal (79,000 L) or more; 5
of 42,000 gal (160,000 L) or more; and one of greater than 420,000 gal (1,600,000 L) (Appendix
Table E-l 5). The largest spills from these data sets ranged from 1,000,000 gal (3,800,000 L) to
2,900,000 gal (11,000,000 L).
The EPA results give insight into recovery and reuse. Of the volume of spilled produced water, 16%
was recovered for on-site use or disposal, 76% was reported as unrecovered, and the rest was
unknown. The fewest spills occurred from wells and wellheads, but these spills had the greatest
median volumes. Failure of container integrity was responsible for 74% of the volume spilled (U.S.
EPA. 2015ml.
7-35
-------
Chapter 7 - Produced Water Handling
10,000,000.0
I 1,000,000.0
~ro
00
£ 100,000.0
~ro
3
~o
t 10,000.0
SW Median Volume (gallons)
SW Mean Volume (gallons)
SW Maximum Volume (gallons)
o
100.0
2000 2002 2004 2006 2008 2010 2012 2014 2016
Year
Figure 7-13. Median, mean, and maximum produced water spill volumes for North Dakota
from 2001 to 2015.
7-36
-------
Chapter 7 - Produced Water Handling
Table 7-7. Summary of produced water release volumes.
Sources: U.S. EPA (2015m), Fisher and Sublette (2005), and Appendix Section E.5.
Number
Minimum
25th percentile
Median
Mean
75th percentile
Maximum
Study
Year(s)
Total
Quantified
(gal)
(gal)
(gal)
(gal)
(gal)
(gal)
Oklahoma
1993-2002
7,916
2,365
0.0
630
1,700
7,000
4,200
3,400,000
1993
373
161
0.4
420
1,500
3,900
4,200
46,000
1994
844
333
0.4
420
1,600
5,400
4,200
84,000
1995
913
333
0.0
420
1,500
3,700
4,200
63,000
1996
880
333
4.2
630
2,100
6,500
4,200
420,000
1997
806
270
0.4
630
1,900
6,000
4,200
120,000
1998
825
236
2.1
798
4,900
2,100
4,200
105,000
1999
886
218
10.5
840
2,100
6,600
4,200
120,000
2000
853
155
4.2
840
2,100
5,600
5,040
210,000
2001
826
144
21.0
840
2,100
31,000
6,510
3,400,000
2002
710
182
0.8
630
1,700
5,500
3,276
130,000
U.S. EPA
2006-2012
225
2.1
420
1,008
10,920
2,982
1,344,000
North Dakota
2001
97
21.0
168
420
2,646
2,520
42,000
2002
110
4.2
210
756
2,604
2,100
25,200
2003
128
2.1
126
504
3,150
2,562
58,800
2004
159
10.5
126
420
2,478
2,100
88,200
2005
184
5.0
126
420
2,142
1,680
54,600
2006
226
5.0
126
420
3,150
1,680
189,000
2007
248
0.4
210
420
2,814
2,100
210,000
7-37
-------
Chapter 7 - Produced Water Handling
Number
Minimum
25th percentile
Median
Mean
75th percentile
Maximum
Study
Year(s)
Total
Quantified
(gal)
(gal)
(gal)
(gal)
(gal)
(gal)
North Dakota, cont.
2008
248
8.4
84
504
2,520
2,058
54,600
2009
208
2.1
126
630
2,100
2,100
27,300
2010
255
0.1
126
840
2,478
2,310
34,020
2011
381
2.1
126
336
2,436
1,680
58,800
2012
543
7.1
84
336
2,310
1,260
84,000
2013
700
2.1
126
378
3,402
1,428
714,000
2014
846
0.8
84
336
3,528
1,470
1,008,000
2015
609
0.8
84
336
7,560
1,386
2,940,000
7-38
-------
Chapter 7 - Produced Water Handling
7.4.2.9 Environmental Receptors and Transport
Data from the EPA (U.S. EPA. 2015m) were used to show that some spills were known to impact
environmental receptors: soil (141 spills, 340,000 gal, or 1.3 million L); surface water (17 spills,
170,000 gal, or 640,000 L); surface water and soil (13 spills); and groundwater (1 spill, 130 gal, or
490 L).1 Although 1 spill was identified as reaching groundwater, the possible groundwater impact
of 107 of the spills was unknown.
In summary, 18 produced water spills reached surface water or groundwater, accounting for 8% of
the 225 cases and accounting for approximately 170,000 gal (640,000 L) of produced water. Spills
with known volumes that reached a surface water body ranged from less than 170 gal (640 L) to
almost 74,000 gal (280,000 L), with median of 5,900 gal (22,000 L). In 30 cases, it is unknown
whether a spill of produced water reached any environmental receptor.
An assessment conducted by the California Council on Science and Technology (CCST. 2015a) states
that between January 2009 and December 2014, 575 produced water spills were reported to the
California Office of Emergency Services of which nearly 18 percent impacted waterways (CCST.
2015a). These spills occurred in areas where production from both unconventional and
conventional reservoirs occurs. Additional studies of spill impacts are presented in Appendix
Section E.5.3.
Studies of Environmental Transport of Released Produced Water
The processes that affected the fate and transport of spilled produced water (Figure 7-14) are the
same as those processes that impact the fate and transport of spilled chemicals (Section 5.8).
Produced water spills differ from the chemical spills as they are always primarily spills of water
containing multiple chemicals. Additionally, produced water of high salinity is denser than water
and may alter transport and transformation properties of the chemicals and soils.2 If a spill occurs
prior to treatment in an oil and water separator, the produced water can be spilled along with oil. In
the environment, oil is transported as a separate phase liquid as it is immiscible with water. The oil
phase may become trapped (similarly to how oil is trapped in oil reservoirs) and serve as a slowly
dissolving source of hydrocarbons to the environment.
For example, Whittemore (2007) described a site with relatively little infiltration due to moderate
to low permeability of silty clay soil and low permeability of underlying shale units. Thus, most, but
not all, of the historically surface-disposed produced water at the site flowed into surface drainages.
Observed historic levels of chloride in receiving waters resulted from the relative balance of
produced water releases and precipitation runoff, with higher concentrations corresponding to low
stream flows. Persistent surface water chloride contamination was attributed to slow flushing and
discharge of contaminated groundwater.
1 Quoted volumes.
2 Appendix Section E.7 describes the estimation of chemical properties for organic chemical constituents of produced
water for baseline conditions of low TDS. Elevated salinity, as is common for produced water, would alter these values.
7-39
-------
Chapter 7 - Produced Water Handling
Hydraulic fracturing-
related spill or release
Sorption
Unsaturated
Soil j£S£f
Transformation
chemicals penetrating soil
layer above groundwate'r
Uspersion
chemical
groundwater
plume
Groundwater
Surface Water
Sorption
Transformation Dissolution
Dispersion
Schematic of the Fate and Transport Processes
Governing Potential Impacts of Spills
and Releases to Drinking
- Water Resources
Volatilization
Figure 7-14, Schematic view of transport processes occurring during releases of produced
water.
Because it is denser than freshwater, saline produced water can migrate downward through
aquifers. Whittemore f20071 reported finding oilfield brine with a chloride concentration of 32,900
rag/L at the base of the High Plains aquifer. Where aquifers discharge to streams, saline stream
water has been reported, although at reduced concentrations fWhittemore. 20071. likely due to
diffusion within the aquifer and mixing with stream water. The stream flow rate, in part,
determines mixing of substances in surface waters. High flows are related to lower chemical
concentrations, and vice versa, as demonstrated for bromide in the Allegheny River fStates etal..
2013).
7.5 Roadway Transport of Produced Water
Produced water is transported to treatment and disposal sites via pipeline, roadways, or railroad
tankers. Accidents during transportation of hydraulic fracturing produced water are a possible
mechanism leading to potential impacts to drinking water as truck-related releases have been
reported. Nationwide data are not available, however, on the number of such accidents that result
in impacts.
Crash rate estimates for Texas showed that commercial motor vehicle [CMV) crashes were
correlated with oil and gas development activities over a recent period of increased oil and gas
development (Ouiroga and Tsapakis. 2015). As an example of the results, the number of new wells
7-40
-------
Chapter 7 - Produced Water Handling
in the Permian Basin increased (by 61%) and so did rural CMV crashes (by 52%). For the Barnett
Shale region, the number of new wells decreased (by 49%), and so did rural CMV crashes (by 34%).
The correlations were strongest for the rural areas with oil and gas development (Permian and
Eagle Ford).
Based on scenarios presented in Appendix Section E.8, the EPA estimated for this assessment the
number of releases from truck crashes as having a chance of occurrence ranging between 1:110 and
1:13,000 over the lifetime of a producing well. The wide range of these estimates reflects both
variable (distance and volume transported) and uncertain (crash rate) quantities. At 5,300 gal (20
m3) per truckload, the volume from an individual spill would be low relative to the typical volume
of water produced from a well. Several limitations are inherent in this analysis, including differing
rural road and highway accident rates, differing transport distances, and differing amounts of
produced water transported. Further, the estimates present an upper bound on impacts, because
not all releases would reach or impact drinking water resources.
As for other types of impacts to drinking water resources, local effects can be significant despite the
generally small numbers. For example, a brine-truck spill in Ohio resulted in concern for impacts to
a drinking-water-source reservoir (Tucker. 2016).
7.6 Synthesis
Produced water is a by-product of oil and gas production and is that water that comes out of the
well after hydraulic fracturing is completed and injection pressure is reduced. Produced water may
contain hydraulic fracturing fluid, water from the surrounding formation, and naturally present
hydrocarbons. Initially the chemistry of produced water reflects that of the hydraulic fracturing
fluid. With time, the chemistry of the produced water becomes more similar to the water in the
formation. Produced water is directly re-injected or stored at the surface for eventual reuse or
disposal. Impacts to drinking water resources from produced water have been shown where spilled
produced water entered surface water bodies or aquifers.
7.6.1 Summary of Findings
The volume and composition of produced water vary geographically, both within and among
different production zones and with time and other site-specific factors. In most cases, there are
high initial flow rates of produced water that last for a few weeks, followed by lower flow rates
throughout the duration of gas production. The amount of fracturing fluid returned to the surface
varies, and typically is less than 30%. In some formations (e.g., the Barnett Shale), the ultimate
volume of produced water can exceed the volume of hydraulic fracturing fluid because of an inflow
of water.
Knowledge of the composition of produced water comes from analysis of samples. Analysis of an
individual sample is made much easier if the hydraulic fracturing and any equipment maintenance
chemicals have been disclosed. Much of the chemical loading of produced water comes from
naturally occurring material, both organic and inorganic, in the formation along with
transformation products. As such, knowledge of produced water composition is uniquely
7-41
-------
Chapter 7 - Produced Water Handling
dependent on sampling and analysis, which requires appropriate analytical methods. These are
methods that can deal especially with high levels of TDS. Recently developed laboratory methods
have greatly expanded the knowledge of organic chemicals in shale-gas and CBM produced waters,
but these methods rely on advanced equipment and techniques. Routine methods of laboratory
analysis do not detect many of the organic constituents of produced water.
The composition of produced water changes with time as the hydraulic fracturing fluid contacts the
formation and mixes with the formation water. Typically it becomes more saline and more
radioactive, if those constituents are present in the formation, while containing less DOC. The
changing composition of produced water suggests that the potential concern for produced water
spills also changes: initially the produced water may contain more hydraulic fracturing chemicals,
later the concern may shift to the impact of high salinity water. Although varying within and
between formations, shale and tight gas produced water typically contains high levels of TDS
(salinity) and associated ionic constituents (bromide, calcium, chloride, iron, potassium,
manganese, and sodium). Produced water can also contain toxic materials, including barium,
cadmium, chromium, lead, mercury, nitrate, selenium, and BTEX. CBM produced water can have
lower levels of salinity if its coal source was deposited under fresh water conditions, or if
freshwater inflows to coal beds dilutes the formation water (Dahm etal.. 2011). Many organic
compounds have been identified in produced water. Most of these are naturally occurring
constituents of petroleum. With the advent of advanced analytical techniques, more hydraulic
fracturing fluid chemicals have been identified in produced water. These include some known
tracer compounds, but others are known to exist whose identities have not yet been determined.
Work has been done to identify environmentally benign tracers for assessing impacts, but these
tracers have not been fully developed. Despite the presence in produced water of known hydraulic
fracturing chemicals, the majority of organic and inorganic constituents of produced water come
from the formation and cannot be minimized through actions of the operator. Throughout the
formation-contact time, reactions occur between the constituents of the fracturing fluid and the
formation.
Produced water spills have occurred across the country. From evaluation of data from across the
United States and a focused study of North Dakota, the median produced water spill ranges from
336 to 1,000 gal (1,300 to 3,800 L). Although half of the spills are smaller than the median spill size,
small numbers of much higher volume spills occur. In 2015, there were 12 spills in North Dakota
greater than 21,000 gal (80,000 L), and one of 2,900,000 gal (11,000,000 L). From 2010 to 2015,
there were approximately 5 to 7 produced water spills per hundred operating production wells.
The major causes identified for these spills are container and equipment failures, human error, well
communication, blowouts, pipeline leaks, and unpermitted discharges. Section 7.4.2 described
impacts that were both of short and long term duration.
Highway transportation of produced water has resulted in crashes, but the impacts from these are
unknown. Analysis of Texas crashes shows that as the oil and gas development activities increase,
so do crashes, especially in rural areas. The EPA estimated the chance of a crash releasing produced
water to range from 1:110 to 1:13,000.
7-42
-------
Chapter 7 - Produced Water Handling
7.6.2 Factors Affecting the Frequency or Severity of Impacts
The potential of spills of produced water to affect drinking water resources depends upon the
release volume, duration, and composition, as well as watershed and water body characteristics.
Larger spills of greater duration are more likely to reach a nearby drinking water resource than are
smaller spills. Small releases, however, can impact resources where there are direct conduits from a
source to receptor, such as fractures in rock. The composition of the spilled fluid also impacts the
severity of a spill, as certain constituents are more likely to affect the quality of a drinking water
resource.
Potential impacts to water resources from hydraulic fracturing related spills are expected to be
affected by watershed and water body characteristics. For example, overland flow is affected by
surface topography and surface cover. Infiltration of spilled produced water reduces the amount of
water threatening surface water bodies. However, infiltration through soil can lead to groundwater
impacts. Releases from pits can directly impact drinking water resources.
7.6.3 Uncertainties
The volume and some compositional aspects of produced water are known from published sources.
The amount of hydraulic fracturing fluid returned to the surface is not well defined, because of the
imprecise distinction between flowback and produced water. With regard to composition, TENORM
and organics have the most limited data. Most of the available data on TENORM has come from the
Marcellus Shale, where concentrations are typically high in comparison to the limited data available
from other formations. Many organic constituents of produced water have been identified, and
many of them are naturally occurring petroleum hydrocarbons. As methods improve and more data
are collected, an increasing number of hydraulic fracturing fluid chemicals are being identified in
produced water. Little is known concerning subsurface transformations and is reflected in only a
few transformation products have been positively identified. Halogenation of organics has been
noted, though.
Nationwide data on spills of produced water are limited in two primary ways: the completeness of
reported data cannot be determined, and individual states' reporting requirements differ fU.S. EPA.
2015m). Therefore, the total number of spills occurring in the United States, their release volumes,
and associated concentrations can only be estimated because of these underlying data limitations.
Spills vary in volume, duration, and composition, and most spill response focusses on immediate
clean up, so several aspects of spills are not precisely characterized. The volume released is often a
rough estimate, in part, because the spilled liquid spreads across the scene and is inherently
difficult to measure. Simple measurements are often used to characterize the spill, rather than
determining chemical concentrations (e.g., measuring electrical conductivity). As a consequence the
suite of chemicals, and their concentrations, potentially impacting drinking water resources are
usually unknown. Thus, the severity of impacts to drinking water resources is not usually well
quantified.
Spills can originate from blowouts, well communication, aboveground or underground pipeline
breaks, leaking pits, failed containers, human error (including unpermitted discharges, failure to
7-43
-------
Chapter 7 - Produced Water Handling
detect spills, and failure to report spills) or unknown causes. The difference between these causes
affects the location and size of the spill or release. For example, a container that fails may release a
small amount of produced water, and be located on the well pad. A pipeline break may occur at a
distance away from the well pad and release a larger amount of water from a bigger source (i.e., a
pit). In addition, the factors governing transport of spilled fluid to a potential receptor vary by site:
the presence and quality of secondary or emergency containment and spill response; the rate of
overland flow and infiltration; the distance to a surface water body or drinking water well; and
transport and fate processes. Impacts to drinking water resources from spills of produced water
depend on environmental transport parameters, which can, in principle, be determined but are
unlikely to be known or adequately specified in advance of a spill.
Because some constituents of produced water are constituents of natural waters (e.g., bromide in
coastal surface waters) or can be released into the environment by other pollution events (e.g.,
benzene from gasoline releases, bromide from coal mine drainage), baseline sampling prior to
impacts is one way to increase the certainty of an impact determination. Further sampling and
investigation can be used to develop the linkage between a release and a documented drinking
water impact Appropriate sampling and analysis protocols, using quality assurance procedures,
are essential for developing data that can withstand scrutiny. The EPA's northeastern Pennsylvania
case study illustrates that the analytes that can be used to distinguish among types of water vary
depending on the specifics of the situation. No single constituent or parameter could be used alone
to assess impacts, and multiple lines of evidence were needed to assess the suspected impacts.
7.6.4 Conclusions
Produced water has the potential to affect the quality of drinking water resources if it enters into a
surface water or groundwater body used as a drinking water resource. This can occur through
spills at well pads or during transport of produced water. Specific impacts depend upon the spill
itself, the environmental conditions surrounding the spill, water body and watershed
characteristics, and the composition of the spilled fluid. The impacts from the majority of spills and
releases is generally localized in extent as only the largest spills and releases impact large areas.
7-44
-------
Chapter 8- Wastewater Disposal and Reuse
Chapter 8. Wastewater Disposal
and Reuse
Abstract
This chapter addresses the practices and related impacts on drinking water resources that take place
during the final stage of the hydraulic fracturing water cycle. This stage encompasses the management
of wastewater, including disposal, reuse in hydraulic fracturing operations, or other uses. For this
assessment, wastewater is defined as produced water from hydraulically fractured oil and gas wells that
is managed by any of a number of strategies. The constituents of concern in hydraulic fracturing
wastewaters that are most frequently noted include high total dissolved solids (TDS), chloride, bromide,
and radionuclides (radium in particular). Other alkaline earth metals (e.g., barium), organics, and
suspended solids, may be of concern as well.
Most hydraulic fracturing wastewater is managed by injection into Class II disposal wells. There are also
"aboveground" management practices, which include reuse in subsequent hydraulic fracturing
operations; treatment at a centralized waste treatment facility followed by reuse or discharge to surface
water or a publicly owned treatment works; evaporation; irrigation; and direct discharge (under limited
conditions). These practices can affect both surface water and groundwater.
Impacts on surface water arise from discharges of inadequately treated wastewater. In particular,
bromide and iodide found in highly saline wastewaters can contribute to disinfection byproduct
formation in downstream drinking water systems. If not removed during treatment, radium, metals, and
organic compounds can also be discharged. Factors affecting the frequency and severity of impacts on
surface waters include the wastewater's composition, its volume, and the processes used to treat it
(common wastewater treatment processes do not significantly reduce the high TDS content in hydraulic
fracturing wastewaters). In addition, site-specific factors such as local hydrology, size of the receiving
water body, and other activities taking place in a watershed can affect the severity of the impact.
Pits and impoundments used for storage or disposal can impact surface water or groundwater through
spills, leaks, and infiltration through soils. The frequency and severity of such impacts depend on pit
construction and maintenance as well as proximity to drinking water resources. Unlined pits or those
with compromised liners can cause long-lasting impacts on groundwater. Depth to the water table, soil
properties, and the contaminants in the wastewater also affect the likelihood of impacts.
Characterizing the impacts from wastewater management associated with hydraulic fracturing is
challenging given gaps in the data. Specifically, there are limited data on the wastewater volumes
managed, on the influent and effluent concentrations and volumes from facilities that treat wastewater
from hydraulic fracturing operations, and on wastewater residual characteristics and management of
those residuals. Further, there is inadequate monitoring of drinking water resources for specific
contaminants associated with hydraulic fracturing wastewater. However, the data that are available
have shown that management of hydraulic fracturing wastewater through aboveground practices has
affected the quality of water resources.
8-1
-------
Chapter 8 - Wastewater Disposal and Reuse
This page is intentionally left blank.
8-2
-------
Chapter 8- Wastewater Disposal and Reuse
8. Wastewater Disposal and Reuse
8.1 Introduction
The final stage of the hydraulic fracturing water cycle encompasses disposal and reuse of hydraulic
fracturing wastewater. For the purposes of this assessment, "hydraulic fracturing wastewater" is
defined as produced water from hydraulically fractured oil and gas wells that is being managed
using practices that include, but are not limited to, reuse in subsequent hydraulic fracturing
operations, treatment and discharge, and injection into disposal wells.12 3 Although the term
"wastewater" is generally used in this chapter, when more specific information about a wastewater
is known (e.g., a source indicates the wastewater is flowback), that information is also noted.
Wells producing from oil and gas reservoirs generate produced water during the course of their
productive lifespan. This produced water includes the often large volumes of flowback generated
immediately after fracturing in deep wells with long horizontal sections. Flowback estimates vary
by formation and are noted in Section 7.2.1 to range from about 300,000 to 10 million gal (1.14 to
37.8 million L) per well fMantell. 2013: U.S. GAP. 20121. This large volume of initial flowback
necessitates having a wastewater management strategy in place before hydraulic fracturing is
initiated. Also, the longer-term generation of produced water requires ongoing wastewater
management.
The majority of wastewater generated from all oil and gas operations in the United States is
managed via Class II injection wells (Veil. 2015). Injection may be for either disposal or enhanced
recovery. As hydraulic fracturing activity expands or diminishes, choices regarding disposal
practices can change in a given region due to factors such as the quality and volume of the fluids;
regulations; available infrastructure; and the feasibility and cost of treatment, reuse, and disposal
options.
Several articles have noted potential effects of hydraulic fracturing wastewater on water resources
(Vengosh et al.. 2014: Olmstead etal.. 2013: Rahm etal.. 2013: States etal.. 2013: Vidic etal.. 2013:
Rozell and Reaven. 2012: Entrekin etal.. 20111. One study used probability modeling that indicated
water pollution risk associated with gas extraction in the Marcellus Shale is highest for the
wastewater disposal aspects of the operation (Rozell and Reaven. 2012). These concerns arise from
1 The term "wastewater" is being used in this study as a general description of certain waters and is not intended to
constitute a term of art for legal or regulatory purposes. This general description does not, and is not intended to, provide
that the production, recovery, or recycling of oil, including the production, recovery, or recycling of flowback or produced
water, constitutes "wastewater treatment" for the purposes of the Oil Pollution Prevention regulation (with the exception
of dry gas operations], which includes the Spill Prevention, Control, and Countermeasure rule and the Facility Response
Plan rule, 40 CFR 112 et seq.
2 Disposal wells are Underground Injection Control (UIC] Class II wells, including those used for disposal (Class IID],
enhanced oil recovery (Class IIR], and hydrocarbon storage (Class IIH].
3 The term "reuse" is sometimes used to imply no treatment or basic treatment (e.g., media filtration] for the removal of
constituents other than total dissolved solids (TDS], while "recycling" is sometimes used to convey more extensive
treatment (e.g., reverse osmosis (RO]] to remove TDS fSlutz etal.. 20121 In this document, the term "reuse" will be used
to indicate use of wastewater for subsequent hydraulic fracturing, regardless ofthe level oftreatment.
8-3
-------
Chapter 8 - Wastewater Disposal and Reuse
the elevated concentrations of chloride, bromide, radionuclides, and other constituents of concern
found in many hydraulic fracturing wastewaters.
This chapter provides follow-on to Chapter 7, which discusses the per-well volumes of produced
water (Section 7.2) and composition (Section 7.3), as well as the processes involved in its
generation and impacts from a number of types of spills and releases. In this chapter, discussions
are provided on management practices for hydraulic fracturing wastewater, available wastewater
production information, and estimated aggregate volumes of wastewater generated for several
states with active hydraulic fracturing (Section 8.2). As a complement to information on the
composition of wastewaters in Chapter 7, Section 8.3 presents brief information on wastewater
constituents and their relevance to wastewater management. Management methods used in recent
years and their potential impacts on drinking water resources are described (Section 8.4). Based on
background information provided in the earlier sections of the chapter, Section 8.5 discusses
documented and potential impacts on drinking water resources from particular constituents, and a
final synthesis discussion is provided (Section 8.6).1
8.2 Volumes of Hydraulic Fracturing Wastewater
This section provides a general overview of aggregate wastewater quantities generated in the
course of hydraulic fracturing and subsequent oil and gas production, including estimates at
regional and state levels. It also discusses methodologies used to produce these estimates and the
associated challenges. (Chapter 7 provides a more in-depth discussion of the processes affecting
produced water volumes and presents some typical per-well values and temporal patterns.) Wells
also generate drilling fluid waste. Compared to produced water, however, drilling fluid wastewater
can constitute a relatively small portion of the total wastewater produced (e.g., <10% in
Pennsylvania during 2004-2013) fU.S. EPA. 2016dl and is not discussed further in this assessment
Wastewater volume can be relevant to treatment costs, reuse options, and disposal capacities. IHS
Global Insight suggests that as a general rule of thumb, the amount of flowback produced in the
days or weeks after hydraulic fracturing is roughly comparable to the amount of produced water
generated long-term over a span of years, which can vary considerably among wells flHS. 20131.
Thus, on a local level, operators can anticipate a relatively large volume of wastewater in the weeks
following fracturing, with slower subsequent production of wastewater.
Wastewater volumes will most likely vary in the future as the amount and locations of hydraulic
fracturing activities change and as existing wells age and move into the later phases of their
production cycles. Substantial increases in wastewater production have occurred during times of
increasing hydraulic fracturing activity. For instance, the average annual volume of wastewater
1 This chapter makes use of background information collected by the EPA's Office of Water (OW] as part of the
development of its recent pretreatment standards for wastewater from unconventional oil and gas formations HJ.S. EPA.
2016d]- The pretreatment standards apply to wastewater from crude oil and natural gas produced by a well drilled into
shale and tight formations. Coalbed methane is beyond the scope of those standards. In this chapter, we consider
wastewater generated by the hydraulic fracturing of those unconventional oil and gas formations included in the
background research done by OW. But we also consider wastewater generated by hydraulic fracturing in coalbed
methane and conventional formations.
8-4
-------
Chapter 8 - Wastewater Disposal and Reuse
generated by all gas production (both shale gas and conventional] in Pennsylvania quadrupled from
the 2001-2006 period to the 2008-2011 period fWilson and Vanbriesen. 20121.
However, although the total volume of wastewater might be expected to generally increase and
decrease as oil and gas drilling and production changes, it is not necessarily a direct correlation.
Data from the Pennsylvania Department of Environmental Protection (PA DEP) fPA DEP. 2016bl
show trends in volumes of wastewater compared to gas produced from wells in the Marcellus Shale
in Pennsylvania (Figure 8-1). Although the data show some variation, they demonstrate a general
positive correlation between the volume of wastewater and the amount of produced gas until early
2015. At that time, Baker Hughes weekly rig counts also began to drop, declining from 85 in early
January 2014 to 24 in early June 2016 fBaker Hughes. 20161. This suggests that a decline in overall
drilling activity (generally a measure of new wells) can be associated with a decline in wastewater
production, although the exact timing depends on whether there is a time delay between drilling
and completion of a well and the start of production from that well.
3,000
1,200
c _
o tj
+_» Ol
3 £
-a "§
o <->
£ J
l/> —
03 £
(J —
2,500
2,000
1,500
1,000
500
1,000
J2> „(> <2> „<> „(>
9 9 9
^ ^ ^ ^ << ^ <<
vi? 9 9^ 9 9^ 9 9^ 9 9^ 9 9*
4? # # # # # # f / / / / /
/
O)
800
600
400
200
~ fD
3 €
= a>
o sT
-j
O
Q.
C
o
r~+
o"
13
I Gas Production
Wastewater Production
Figure 8-1. Wastewater (i.e., produced water and fracturing fluid waste) and produced gas
volumes from unconventional (as defined by PA DEP) wells in Pennsylvania from January
2010 through June 2016.
Source: PA DEP (2016b).
Estimates of produced water compiled by Veil (20151 indicate that although oil and gas production
in the United States increased by 29% and 22%, respectively, between 2007 and 2012, produced
water volumes increased by less than 1%. There may be a number of factors contributing to this,
including as noted by Veil (20151. a number of uncertainties associated with produced water
8-5
-------
Chapter 8 - Wastewater Disposal and Reuse
estimates. First, wastewater generation varies from well to well, as do oil and gas production (see
Chapter 7, Figure 7-1 for discussion of waste water/produced water decline curves). The rates of
decline in both wastewater volume and hydrocarbon production also vary among reservoirs.
Additionally, some wells are drilled and completed but are not immediately put into production.
Relationships between hydraulic fracturing activity, hydrocarbon production, and produced water
volumes are likely reservoir- (and maybe production zone-) specific, and existing wells and
production need to be considered to anticipate wastewater management needs.
8.2.1 National Level Estimate
Veil (2015) estimated that in 2012, U.S. onshore and offshore oil and gas production generated
889.59 billion gal (21.18 billion bbls) of produced water. This national-level estimate represents
total oil and gas wastewater (from all oil and gas resources, and from wells hydraulically fractured
and wells not hydraulically fractured). The estimate was compiled through a state-by-state analysis
of survey data obtained from oil and gas agencies in the 31 states with active oil and gas production
as well as the Department of Interior and U.S. EPA. However, Veil notes several issues with the data
used for these estimates, including variability among states in data reporting, availability, and
completeness. These issues may result in underestimation of the volumes of water produced fU.S.
GAP. 2012). See Section Error! Reference source not found, for more discussion on methods of
estimating wastewater volumes.
8.2.2 Regional/State Level Estimates
A limited number of studies have described the geographic variability in oil and gas wastewater
volumes. Veil (2015) reported that the top ten states nationwide for wastewater production in
2012 included Texas (35% of total), California (15% of total), Oklahoma (11% of total), and
Wyoming (11% of total). A study by the Bureau of Land Management (BLM) fGuerra etal.. 20111
states that in 2008, more than 80% of all oil and gas wastewater was generated in the western
United States, with Texas producing the highest volume and Wyoming the second highest The BLM
report notes substantial wastewater from CBM wells, in particular those in the Powder River Basin
(Wyoming). Figure 8-2 summarizes the wastewater volumes for these western states,
demonstrating the wide variability from state to state (likely reflecting differences in the number of
oil and gas production wells/production activity and reservoir geology). Although the authors do
not identify all wastewater contributions from production involving hydraulic fracturing, the
regions with established oil and gas production are likely to have methods and infrastructure
available for management of hydraulic fracturing wastewater.
8-6
-------
Chapter 8- Wastewater Disposal and Reuse
.'V
Figure 8-2. Wastewater quantities in the western United States (billions of gal per year).
Data from Guerra et al. (2011).
In the Marcellus region, waste data made public by the PA DEP have enabled analyses of
wastewater volumes and trends since 2009. Estimates of produced water (including flowback or
"fracing fluid waste" as well as "produced fluid") by Wunz (2015) and Shale Alliance for Energy
Research Pennsylvania (SAFER PA. 2015) for 2014 are 1.73 and 1.64 billion gal (41.19 MMbl and
39.05 MMbl, respectively). The estimate compiled for this assessment is 0.65 billion gal (15.48
million bbls) for the first half of 2014 (Table 8-1). Variations among estimates reflect, among other
factors, challenges in working with a dynamic database for which changes and corrections are
ongoing.
Table 8-1 presents estimates of the volumes of hydraulic fracturing wastewater generated and the
associated numbers of wells in North Dakota, Ohio, Pennsylvania, Texas (injected flowback only),
and West Virginia. The data shown in this table were compiled for this assessment (except for West
Virginia) and come primarily from records of produced water made publicly available on state
websites.1 These states are represented in Table 8-1 because the produced water volumes
associated with hydraulic fracturing were readily identifiable. The data show that the increase in
1 Data used for Table 8-1 were downloaded from state agency websites and compiled as needed (in either Microsoft Excel
or Microsoft Access] for each state except West Virginia. Once compiled, data were filtered if needed and summed to
produce estimates of wastewater production by year and a count of the numbers of wells generating the wastewater. Data
were downloaded up through 2014. (Note that 2014 data for Pennsylvania and Texas are for partial years.] Differences in
the years presented for the states are due to differences in data availability from the state agency databases. For West
Virginia, data are from a report bv Hansen et al. ("20131 that compiled available flowback data from West Virginia.
8-7
-------
Chapter 8 - Wastewater Disposal and Reuse
the number of wells producing wastewater and the volumes of wastewater produced are generally
consistent with the timing of the expansion of high-volume hydraulic fracturing and track with the
increase in horizontal wells seen in Figure 3-20.
Several states with mature oil and gas industries (California, Colorado, New Mexico, Utah, and
Wyoming) make produced water volumes publicly available by well as part of their oil and gas
production data, but they do not directly indicate which wells have been hydraulically fractured.
Some states (Colorado, Utah, Wyoming, and New Mexico) specify the producing formation or the
basin along with volumes of hydrocarbons and produced water. Determining volumes of hydraulic
fracturing wastewater for these states is challenging because there is a possibility of either
inadvertently including wastewater from wells not hydraulically fractured or of missing volumes
that should be included. This may be a particular problem where state terminology regarding what
constitutes an unconventional resource or hydraulically fractured well is ambiguous or possibly
different from other states. Appendix Table F-l provides estimates of wastewater volumes in
California, Colorado, New Mexico, Utah, and Wyoming in regions where hydraulic fracturing activity
is taking place, along with notes on data limitations. The data in Table 8-1 and Appendix Table F-l
illustrate the challenges in both compiling a national estimate of hydraulic fracturing wastewater
and comparing wastewater production among states due to dissimilar data types, presentation, and
availability.
8.2.3 Estimation Methodologies and Challenges
Compiling and comparing data on wastewater production at the wide array of oil and gas locations
in the United States presents challenges associated with data reporting and availability. Different
approaches have been used to estimate state-specific and national wastewater volumes. Data from
state agency websites and databases can be a ready source of information, whether publicly
available and downloadable or provided directly by agencies upon request
Veil (2015) notes that the reported volumes of produced water (e.g., reported by well in state
production data) can be inaccurate or imprecise because produced water is not monitored
continuously. Therefore, reported volumes may be estimates. Other issues such as data
transcription errors or extrapolation of data can also affect reported volumes (Veil. 2015).
Using produced water volumes from state records to estimate the volume of wastewater regionally
or nationally presents additional challenges due to a lack of consistency in data collection,
availability, usability, completeness, and accuracy (Malone etal.. 2015: Veil. 2015: U.S. GAP. 2012).
Due to what are sometimes significant differences in the types of data collected and the
mechanisms, formats, and definitions used, data cannot always be directly compared from state to
state. This makes it difficult to aggregate volume data, and estimates may be better and more useful
at a local or state level. Larger-scale estimates across regions or between states should be
interpreted carefully.
8-8
-------
Chapter 8- Wastewater Disposal and Reuse
Table 8-1. Estimated volumes (millions of gal) of wastewater based on state data for selected years and numbers of wells
producing fluid.
State
Basin
Principal
lithologies
Data type
2000
2004
2008
2010
2011
2012
2013
2014
Comments
North Dakota
Williston
Shale
Produced
water
2
3
130
790
1,900
4,500
8,500
9,700
From North Dakota Oil and
Gas Commission website3.
Data provided for six
members of the Bakken
Shale. Produced water
includes flowback (reports
are submitted within 30 days
of well completion.)
Wells
161
152
844
2,083
3,303
5,036
6,913
8,039
Ohio
Appalachian
Shale
Primarily
flowback
3
29
110
Data from Ohio DNR Division
of Oil and Gasb. Utica data for
2011 and 2012. Utica and
Marcellus data for 2013.
Brine is noted by agency to
be mostly flowback.
Wells
-
-
-
-
9
86
400
-
Pennsylvania
Appalachian
Shale
Flowback
plus
produced
water (%
flowback; %
produced
water)
180
(51%;
49%)
740
(46%;
54%)
1,100
(36%;
64%)
1,300
(27%;
73%)
650
(32%;
68%)
Waste data from PA DEPC.
Second half of 2010 and first
half of 2014. Data described
as unconventional as defined
by PADEP. Separate codes
are provided by PA DEP for
flowback and produced
water.
Wells
-
-
-
1,232
2,434
4,039
5,015
5,150
8-9
-------
Chapter 8 - Wastewater Disposal and Reuse
State
Basin
Principal
lithologies
Data type
2000
2004
2008
2010
2011
2012
2013
2014
Comments
Texas
Unspecified
(entire state)
Shale,
Sandstone
Flowback -
injected
volumes
490
2,200
3,100
2,000
Waste injection data from
Texas Railroad Commission^
Monthly totals are provided
for entire state. Oct - Dec for
2011, full years for 2012 and
2013, and Jan - Oct for 2014
West Virginia
Appalachian
Shale
Flowback -
Estimated
total
disposed
120
110
59
Estimates from Hansen et al.
(2013).e
a North Dakota Industrial Commission. Department of Mineral Resources. Bakken Horizontal Wells By Producing Zone: https://www.dmr.nd.gov/oilgas/bakkenwells.asp.
b Ohio Department of Natural Resources, Division of Oil and Gas Resources. Oil and Gas Well Production. http://oilandgas.ohiodnr.goV/production#ARCHl.
c PA DEP Oil and Gas Reporting website, https://www.paoilandeasreportine.state.pa.us/publicreports/Modules/Welcome/Aereement.aspx.
d Railroad Commission of Texas, Injection Volume Query, http://webapps.rrc.state.tx.us/H10/searchVolume.do;isessionid=J3ceVHhK9nkwPrC7ZcWNMevzF9LCYvRlNmvDv3F
lQQ5waXfcGNGN!1841197795?fromMain=ves&sessionld=143075601021612. Texas state data provide an aggregate total amount of flowback injected for the past few years.
(Data on brine volumes injected do not differentiate hydraulically fractured wells and, therefore, well counts are not presented here.) These values are interpreted as estimates
of generated flowback as based on reported quantities of "fracture water flow back" injected into Class IID wells.
0 West Virginia flowback estimates from Hansen et al. (2013) are based on state data. Well counts that are explicitly associated with the flowback and total disposed volumes
were not available.
8-10
-------
Chapter 8- Wastewater Disposal and Reuse
To compile estimates of the production and management of hydraulic fracturing wastewater, there
are additional challenges. Reporting of wastewater volumes may or may not include information
that helps determine whether the producing well was hydraulically fractured (e.g., an indicator of
resource type or formation). It also might not be clear whether volumes listed as 'produced water'
include the flowback component Some states (e.g., Colorado and Pennsylvania) include information
on disposal and management methods along with production data, and others do not.
Given the limitations of comparing state databases, some studies have generated estimates of
wastewater volume using water-to-gas and water-to-oil ratios along with the reports of
hydrocarbon production (Murray. 2013). The reliability of any wastewater estimates made using
this method would need to be evaluated in terms of the quality, timeframe, and spatial coverage of
the available data, as well as the extent of the area to which the estimates will be applied. Water-to-
hydrocarbon ratios are empirical estimates. Because these ratios show a wide variation among
formations, reliable data are needed to formulate a ratio in a particular region.
Another approach to estimating wastewater volumes would entail multiplying per-well estimates
of produced water production rates by the numbers of wells in a given area. Challenges associated
with this approach include obtaining accurate estimates of the number of new and existing wells,
along with accurate estimates of per-well water production both during the flowback period and
during the production phase of the well's lifecycle. In particular, it can be challenging to correctly
match per-well wastewater production estimates, which will vary by formation, with counts of
wells, which may or may not be clearly associated with specific formations. Temporal variability in
wastewater generation would also be difficult to capture and would add to uncertainty. Such an
approach, however, may be attempted for order of magnitude estimates if the necessary data are
available and reliable.
8.3 Wastewater Characteristics
Along with wastewater volumes, wastewater characteristics and the characteristics of residuals
produced during treatment or storage are important for understanding the potential impacts of
management and disposal of hydraulic fracturing wastewater on drinking water resources. This
section provides brief highlights on several important constituents known to exist in hydraulic
fracturing wastewaters and residuals. Chapter 7 provides more in-depth detail on the chemistry of
produced water, and Chapter 9 discusses reference values and health effects associated with
hydraulic fracturing wastewater constituents.
8.3.1 Wastewater
Wastewater composition is the result of naturally-occurring constituents originating in the
formation solids and fluids as well as chemicals associated with the fracturing fluid. Discussion in
this chapter focuses on constituents in hydraulic fracturing wastewater for which adequate
information is available to assess documented and potential impacts on drinking water resources.
There may also be unknown constituents in wastewaters for which analyses have not been
performed. This is due, in part, to a lack of information on fracturing fluid ingredients identified as
confidential business information (CBI). In addition, there are uncertainties about how fracturing
8-11
-------
Chapter 8 - Wastewater Disposal and Reuse
fluid ingredients are degraded or removed in the subsurface. (See Chapter 5, Section 5.8 for a
discussion of processes that can cause chemicals to degrade or transform in the subsurface.)
8.3.1.1 Total Dissolved Solids and Inorganics
Hydraulic fracturing wastewaters are generally high in total dissolved solids (TDS), especially those
from shales and tight formations, with TDS values ranging from less than 1,000 mg/L to hundreds
of thousands of mg/L (Section 7.3.4.4). The TDS in wastewaters from shale formations is typically
dominated by sodium and chloride and may also include elevated concentrations of bromide,
bicarbonate, sulfate, calcium, magnesium, barium, boron, strontium, radium, organics, and heavy
metals (Chapman etal.. 2012: Rowan etal.. 2011: Blauch etal.. 2009: Orem etal.. 2007: Sirivedhin
and Dallbauman. 20041.
Within each formation, the minimum and maximum values presented in Section 7.3.4.4 suggest
spatial variation in TDS content that may need to be accommodated when considering management
strategies such as reuse or treatment. In contrast to shales and sandstones, TDS values for
wastewater from CBM formations are generally lower, with reported concentrations ranging from
approximately 150 mg/L to 62,000 mg/L (DOE. 2014b: Dahm etal.. 2011) (AppendixTable E-3).
This results in fewer treatment challenges and a wider array of management options.
Constituents commonly found in TDS from hydraulic fracturing wastewaters may have potential
health impacts or create treatment burdens on downstream drinking water systems if discharged
at high concentrations to drinking water resources. Bromide, for example, can contribute to the
increased formation of disinfection byproducts (DBPs) during drinking water treatment (Hammer
and VanBriesen. 20121: see Section 8.5.1.
8.3.1.2 Organics
Less information is generally available about organic constituents in hydraulic fracturing
wastewaters than about inorganic constituents, but there are now several studies reporting
analyses of organic constituents (Chapter 7). The organic content in flowback waters can vary
based on the chemical additives (e.g., biocides, antiscalants, gelling agents, breakers) used in
hydraulic fracturing fluids and the chemistry of the formation, but the organics generally include
polymers, oil and grease, volatile organic compounds (VOCs), and semi-volatile organic compounds
(SVOCs) fAkob etal.. 2016: Walsh. 2013: Hayes. 20091. Examples of other constituents detected
include alcohols, naphthalene, acetone, and carbon disulfide, compounds that may be remnants of
hydraulic fracturing fluid chemicals (Hayes and Severin. 2012b: Hayes. 2009) (Appendix E).
Wastewater associated with CBM wells may have high concentrations of aromatic and halogenated
organic contaminants potentially requiring treatment depending on how the wastewater will be
managed fPashin etal.. 2014: Sirivedhin and Dallbauman. 20041. Concentrations of BTEX (benzene,
toluene, ethylbenzene, and xylenes) in CBM produced waters are lower than in shale produced
waters (Appendix Table E-9).
New research is focusing on transformation products generated in the subsurface; experimental
work by Kahrilas etal. f20151 suggests that the biocide glutaraldehyde can be present in
wastewaters along with its transformation products. Low molecular weight organic acids such as
8-12
-------
Chapter 8- Wastewater Disposal and Reuse
acetate, formate, and pyruvate have been detected in Marcellus wastewater, indicating microbial
degradation of organic compounds in the fracturing fluid or formation fAkob etal.. 20151.
8.3.1.3 Radionuclides
Radionuclides are constituents of concern in some hydraulic fracturing wastewaters, with most of
the available data obtained for the Marcellus Shale in Pennsylvania (Appendix Table E-8). Results
from a United States Geological Survey (USGS) report fRowan et al.. 2 0111 indicate that the
predominant radionuclides in Marcellus Shale wastewater are radium-226 and radium-228.
Radionuclides in produced fluids are considered 'technologically enhanced naturally-occurring
radioactive material' (TENORM) because they have been exposed to the accessible environment1
Although data regarding radionuclides in wastewater from formations other than the Marcellus
Shale are limited, there is information on the naturally occurring radioactive material (NORM) in
the formations themselves.2 In particular uranium and thorium can be found in certain organic-rich
black shales. High uranium content has been measured in the Marcellus, Barnett, Woodford, and
other black shales fSwanson. 19551 (Section 7.3.4.6). Radium-226 and -228 are decay products of
uranium and thorium and are soluble (Sturchio etal.. 2001: Langmuir and Riese. 1985). Therefore
wastewater from shales with high concentrations of uranium and thorium can contain radium,
especially where TDS concentrations are also high (Rowan etal.. 2011: Sturchio etal.. 2001: Fisher.
19981. Section 7.3.3.2 provides further information on radionuclides in produced waters and in
formations.
8.3.2 Constituents in Residuals
Depending on the wastewater and the treatment processes used, treatment residuals can consist of
sludges, spent media (used filter materials), or brines. Residuals may require further treatment
(e.g., dewatering sludges) prior to disposal (see Section 8.4.7 for further discussion on management
of residuals). Residuals can contain constituents such as total suspended solids (TSS), TDS, metals,
radionuclides, and organics. These constituents will be concentrated in the residuals, with the
degree of concentration depending on the type of treatment employed. Processes such as
electrodialysis and mechanical vapor recompression have been found to yield residuals with TDS
concentrations in excess of 150,000 mg/L after treating waters with influent TDS concentrations of
approximately 50,000 - 70,000 mg/L fHaves etal.. 2014: Peraki and Ghazanfari. 20141.
Also, TENORM in wastewaters can cause residual wastes to have gamma radiation emissions
(Kappel etal.. 2013). A laboratory study by Zhang etal. (2014b) estimated that the barium sulfate
solids precipitated during treatment to remove barium and strontium from Marcellus Shale
wastewater would also contain between 2,571 and 18,087 pCi/g of radium due to coprecipitation.
Another similar study using mass balances calculated that sludge from a sulfate precipitation
1 Technologically Enhanced Naturally Occurring Radioactive Material (TENORM] is defined by the EPA as naturally
occurring radioactive materials (NORM] that have been concentrated or exposed to the accessible environment as a result
of human activities such as manufacturing, mineral extraction, or water processing.
2 Naturally Occurring Radioactive Materials (NORM] are radioactive materials found in nature that have not been moved
or concentrated by human activities.
8-13
-------
Chapter 8 - Wastewater Disposal and Reuse
process would have an average radium concentration of 213 pCi/g fSilva etal.. 20121. In sludge
from lime softening processes, Silva etal. f 2 0121 estimated a radium-226 concentration of 58
pCi/g a level that would necessitate disposal as a low-level radioactive waste.
8.4 Wastewater Management Practices and Their Potential Impacts on
Drinking Water Resources
Operators have several strategies for management of hydraulic fracturing wastewaters (Figure
8-3], with the most common choice being disposal via Class IID wells (Veil. 2015: Clark etal.. 2013:
Hammer and VanBriesen. 20121. Other practices include reuse in subsequent hydraulic fracturing
operations (with varying levels of treatment), treatment at a centralized waste treatment facility
(CWT) (often followed by reuse), evaporation (in arid regions), irrigation (with no discharge to
waters of the United States), and direct discharge for livestock or agricultural use (allowed west of
the 98th meridian). Up until 2011, treatment of unconventional oil and gas wastewaters (as defined
by PA DEP) at publicly owned treatment works (POTWs) was a common practice for wastewater
management in the Marcellus region (Lutz etal.. 2013): this is discussed further in Text Box 8-1.
The methods shown in Figure 8-3 represent wastewater management strategies, not all of which
would be used at the same facility. Descriptions of incidents of unpermitted disposal and resulting
legal actions have also been publicly reported (Chapter 7). However, such events are not generally
described in the scientific literature, and the prevalence of this type of activity is unclear. Additional
sources of information about potential impacts exist, but some records are sealed (e.g., litigation
records) and are not publicly accessible.
Well
On-site
treatment
Liquid
residuals
Injection
Well
Direct discharge
for agriculture
(low TDS
wastewater),
other uses
Surface
water
discharge
Landfill
(solid
residuals)
Evaporation
pit
Figure 8-3. Schematic of wastewater management strategies.
Gray lines indicate management strategies that involve injection, either for reuse or disposal, and blue lines
indicate management strategies that lead to other end points such as discharge, evaporation, landfills, or other
uses.
8-14
-------
Chapter 8- Wastewater Disposal and Reuse
Each of the wastewater management strategies can potentially lead to impacts on drinking water
resources during some phase of their execution. Such impacts include, but are not limited to:
accidental releases during transport (Chapter 7); discharges of treated wastewaters from CWTs or
POTWs where treatment for certain constituents has been inadequate; migration of constituents to
groundwater due to leakage from pits or land application of wastewaters; leakage from pits that
reach surface waters (Chapter 7, Section 8.4.5); inappropriate management of liquid or solid
residuals (e.g., leaching from landfills); or accumulation of constituents in sediments near outfalls of
CWTs or POTWs that are treating or have treated hydraulic fracturing wastewater.1
A reliable census of oil and gas wastewater management practices nationwide is difficult to
assemble due to a lack of consistent and comparable data among states. In addition, we do not
know how often operators use more than one wastewater management strategy at a site (e.g.,
evaporation and injection), further complicating the tracking of wastewater management As part of
a data survey conducted by Veil ("20151. some state agencies provided estimates of oil and gas
wastewater volumes handled by several management practices (Table 8-2). These estimates
illustrate how widespread injection for both enhanced recovery and for disposal is for managing oil
and gas wastewater. The data also show regional differences in reuse and other practices. For
hydraulic fracturing wastewaters, Table 8-3 illustrates nationwide variability in the primary
wastewater management methods using qualitative and quantitative sources. Similar to Table 8-2,
Table 8-3 shows disposal via underground injection predominates in most regions, and reuse is
predominant in the Marcellus Region. (Table 8-3 does not include wastewater management in areas
of CBM production.)
Management choices are affected by cost and a number of directly and indirectly related factors,
including the chemical properties of the wastewater; the volume, duration, and flow rate of the
wastewater generated; the feasibility of each option; the availability of necessary infrastructure;
local, state, and federal regulations (Text Box 8-2); and operator discretion (U.S. GAP, 2012; NPC,
2011a). The economics (such as transport, storage, and disposal costs) and availability of treatment
and disposal methods are of primary importance (U.S. GAP, 2012). For wastewater composition,
there is limited information on the degradation or removal of fracturing fluid ingredients in the
subsurface. Chemical disclosure requirements vary among states, and some fracturing fluid
ingredients are claimed to be CBI. Therefore, the possible presence of unknown chemical
constituents in wastewater contributes to uncertainty about the effectiveness and potential impacts
of management strategies, particularly with regard to treatment efficacy.
1 The term surface water as used in this assessment refers to surface waters that meet the definition of waters of
the United States under the CWA ("House Bill No. 1950,2011],
8-15
-------
Chapter 8 - Wastewater Disposal and Reuse
Table 8-2. Estimated percentages of wastewater managed by practice and by state.
Source: Veil (2015). Estimates do not identify interstate transport (e.g., wastewater transported from PA to OH or WV for injection into disposal wells). Thus,
there may have been some double counting of volumes in both the generating and receiving states.
Management
practice
Percentage of produced water managed by practice and state (%)
AR
CA
CO
NM
ND
OH
OK
PA
TX
UT
WV
WY
Injection for
enhanced oil
recovery
22
46
32
50 d
18
4.0
47
0
48
40
27
73
Injection for
disposal
76
20
32
50 d
56
91
47
12
37
47
25
27
Surface
discharge
0
2
10
no data
0
0
0
2.3
5.0f
6
0
uncertain
Evaporation
0
21
9.0
no data
0
0
0
0
0
0
uncertain
Offsite
commercial
disposal
0.1a
9
5.7°
no data
26
Included in
injection
for disposal
6.0 e
0
d)
o
1
7s
28 h
uncertain
Beneficial
reuse
1.1b
no data
12 b
no data
0
5.0
0
85
(includes
reuse for
HF)
Est. 15-20
(flowback
fluid)
0.5
uncertain
uncertain
a Land farm.
b Reuse for HF.
c Pits.
d Assumes even split with injection for enhanced oil recovery and injection for disposal.
6 Injection.
f Fresh produced water.
g Evaporation ponds.
h Disposal wells.
8-16
-------
Chapter 8- Wastewater Disposal and Reuse
Table 8-3. Management practices for wastewater from unconventional oil and gas resources.
Source: U.S. EPA (2016d).
Basin
Formation
Resource
type
Reuse
Injection for
disposal
CWT
facilities
Notes
Available
datab
Michigan
Antrim
Shale gas
XXX
Qualitative
Appalachian
Marcellus/Utica (PA)
Shale gas
XXX
XX
XX
Limited disposal wells in east
Quantitative
Marcellus/Utica (WV)
Shale gas/oil
XXX
XX
X
Quantitative
Marcellus/Utica (OH)
Shale gas/oil
XX
XXX
X
Mixed
Anadarko
Granite Wash
Tight gas
XX
XXX
xa
Mixed
Mississippi Lime
Tight oil
X
XXX
Reuse/recycling limited but is being
evaluated
Qualitative
Woodford, Cana, Caney
Shale gas/oil
X
XXX
xa
Qualitative
Arkoma
Fayetteville
Shale gas
XX
XX
xa
Few existing disposal wells; new CWT
facilities are under construction
Mixed
Fort Worth
Barnett
Shale gas
X
XXX
xa
Reuse/recycle not typically used due to
high TDS early in flowback and abundance
of disposal wells
Mixed
Permian
Avalon/Bone Springs,
Wolfcamp, Spraberry
Shale/tight
oil/gas
X
XXX
xa
Mixed
TX-LA-MS Salt
Haynesville
Tight gas
X
XXX
Reuse/recycle not typically used due to
high TDS early in flowback and abundance
of disposal wells
Mixed
West Gulf
Eagle Ford, Pearsall
Shale gas/oil
X
XXX
X
Mixed
Denver Julesburg
Niobrara
Shale gas/oil
X
XXX
X
Mixed
Piceance; Green
River
Mesaverde/Lance
Tight gas
X
XX
X
Also managed through evaporation to
atmosphere in ponds in this region
Qualitative
8-17
-------
Chapter 8 - Wastewater Disposal and Reuse
Basin
Formation
Resource
type
Reuse
Injection for
disposal
CWT
facilities
Notes
Available
datab
Williston
Bakken
Shale oil
X
XXX
Reuse/recycling limited but is being
evaluated
Mixed
a CWT facilities identified in these formations are all operator-owned.
b This column indicates the type of data the EPA based the number of Xs on. In most cases, the EPA used a mixture of qualitative and quantitative data sources along with
engineering judgment to determine the number of Xs.
XXX—The majority (>50%) of wastewater is managed with this management practice; XX—A moderate portion (>10% and <50%) of wastewater is managed with this
management practice; X—This management practice has been documented in this location, but for a small (<10%) or unknown percent of wastewater. Blanks indicate the
management practices have not been documented in the given location.
8-18
-------
Chapter 8- Wastewater Disposal and Reuse
The availability and use of wastewater management strategies in a region can change over time as
oil and gas production increases or decreases, regulations change, costs shift, and technologies
evolve. Text Box 8-1 and Figure 8-4 illustrate shifting wastewater management practices in
Pennsylvania as gas development in the Marcellus Shale increased and concerns over high-TDS
discharges prompted a regulatory response. Reuse has increased substantially at well sites in
Pennsylvania (labeled as "Reuse HF" in Figure 8-4) and wastewater management at CWTs has
moved toward more facilities that provide wastewater for reuse and do not discharge (termed
"zero-discharge facilities"). The estimated total reuse rate in Pennsylvania was 80% in 2012 and
90% in 2013 (PA DEP. 2015a). In contrast, wastewater disposal data in areas of Colorado where
hydraulic fracturing takes place show a steady use of injection wells, an increase in surface water
discharges, and a decrease in the use of on-site pits for evaporation since 2000 (Figure 8-5).
Another factor influencing reuse is the pace of hydraulic fracturing in the area. When hydraulic
fracturing is active, demand for reuse is high. Some formations that are hydraulically fractured such
as the Marcellus Shale and the Utica Shale are still in the early stages of development, with large
potential resources not yet developed. For these plays, the need for wastewater treatment and/or
reuse may remain high for decades to come, and the long-term wastewater management needs
must be considered and addressed (SAFER PA. 2015).1
Researchers have developed optimization models to aid in the minimization of wastewater
management costs as a part of comprehensive water management planning. For example, Yang et
al. (2014) suggest an approach for reusing flowback in scheduled hydraulic fracturing events to
minimize the operational costs of transportation, treatment, storage, and wastewater disposal.
Another modeling study proposes an approach to minimize the total cost of water usage and
wastewater treatment and disposal by optimizing capital costs (such as the costs of treatment units
and storage pits) and operating costs for flowback management, treatment, storage, reuse, and
wastewater disposal (Lira-Barragan et al.. 20161.
Text Box 8-1. Temporal Trends in Wastewater Management - Experience of Pennsylvania.
Gross natural gas withdrawals from shale formations in the United States increased 518% between 2007 and
2012 (EIA. 2014b). This production increase has led to larger volumes of wastewater requiring appropriate
management (Vidic et al.. 2013: Gregory et al.. 2011: Kargbo et al.. 2010). The rapid increase in wastewater
generated from hydraulically fractured oil and gas wells has led to many changes in wastewater disposal
practices in the oil and gas industry. Changes have been most evident in Pennsylvania, which has experienced
a more than 1,400% increase in natural gas production since 2000 fEIA. 2014b).
Lutz etal. (2013) estimated that total wastewater generation in the Marcellus region increased 570%
between 2004 and 2013. The authors concluded that this increase has created stress on the existing
wastewater disposal infrastructure. In 2010, concerns arose over elevated TDS in the Monongahela River
[Text Box 8-1 is continued on the following page.)
1 As noted in Chapter 3, oil and gas prices influence new drilling activity. However, the links between oil and gas prices
and the generation of wastewater (as a byproduct of production] appear to be less direct.
8-19
-------
Chapter 8 - Wastewater Disposal and Reuse
Text Box 8-1 (continued). Temporal Trends in Wastewater Management - Experience of
Pennsylvania
basin, and studies linked high TDS (and, in particular, high bromide levels) to elevated DBP levels in drinking
water systems (PA DEP. 2011a). In response, PA DEP amended Chapter 95 Wastewater Treatment
Requirements under the Clean Streams Law for new discharges of TDS in wastewaters. This regulation is also
informally known as the 2010 TDS regulation. The regulation disallowed any new direct discharges to
streams as well as direct disposal at POTWs of hydraulic fracturing wastewater and set limits on treated
discharges from new CWTs of 500 mg/L TDS, 250 mg/L chloride, 10 mg/L barium, and 10 mg/L strontium.
Existing discharges were exempt.
In April 2011, PA DEP announced a request that by May 19, 2011, gas drilling operators voluntarily stop
transporting wastewater from shale gas extraction (i.e., unconventional resources as defined by PA DEP) to
the eight CWTs and seven POTWs that were exempt from the 2010 TDS regulation.1 Follow-up letters from
PA DEP to the owners of the wells specified that the role of bromides from Marcellus Shale wastewaters in the
formation of total trihalomethanes (TTHM) was of concern due to the their potential public health impacts
(PA DEP. 2011a).
In response to the request, the oil and gas industry in Pennsylvania accelerated the switch of wastewater
deliveries from POTWs to CWTs for better removal of metals and suspended solids (Schmidt. 2013). Effluent
sampling at two POTWs that had accepted Marcellus Shale wastewater showed that concentrations of
bromide, chloride, barium, strontium, and sulfate dropped after the April 2011 request (Ferrar et al„ 2013):
data based on two sampling events, one before and one after May 2011).
Between early and late 2011, although reported wastewater production more than doubled, Marcellus Shale
drilling companies in Pennsylvania reduced their use of CWTs that were exempt from the 2010 TDS
regulation by 98%, and direct disposal of Marcellus Shale wastewater at POTWs was "virtually eliminated"
(Hammer and VanBriesen. 2012).
Along with the decreased discharges from POTWs, there has been increased reuse of wastewater in the
Marcellus Shale region. From 2008-2011, reuse of Marcellus wastewater for hydraulic fracturing increased,
POTW treatment volumes decreased, tracking of wastewater improved, and wastewater transportation
distances decreased fRahm etal.. 20131 Malonev and Yoxtheimer f20121 analyzed data from 2011 and found
that reuse of flowback increased to 90% by volume. Eight percent of flowback was sent to CWTs. Brine water,
which was defined as formation water, was reused (58%), disposed via injection well (27%), or sent to CWTs
(14%). Of all the fluid wastes in the analysis, brine water was most likely to be transported to other states
(28%). Malonev and Yoxtheimer f20121 also concluded that wastewater disposal to municipal sewage
treatment plants declined nearly 100% from 47,221 bbls in the first half of 2011 to 408 bbls in the second
half.
1 An unconventional formation was defined in 2011 by the state of Pennsylvania as "A geological shale formation existing
below the base of the Elk Sandstone or its geologic equivalent stratigraphic interval where natural gas generally cannot be
produced at economic flow rates or in economic volumes except by vertical or horizontal wellbores stimulated by
hydraulic fracture treatments or by using multilateral wellbores or other techniques to expose more of the formation to
the wellbore." The EPA defines unconventional oil and gas as crude oil and natural gas produced by a well drilled into a
shale and/or tight formation (including, but not limited to, shale gas, shale oil, tight gas, and tight oil]. For the purpose of
the rule, the definition of UOG does not include CBM ("U.S. EPA. 2016dl
8-20
-------
Chapter 8- Wastewater Disposal and Reuse
Other
100%
Reuse Non-HF
POTW
Reuse HF
o%
2009 2010 2011 2012 2013 2014
Figure 8-4. Percentages of total unconventional wastewater (as defined by PA DEP) managed
via various practices for the second half of 2009 through the first half of 2014.
The volume sent to POTWs in 2013 was 0%. Note also that a majority of wastewater sent to CWTs is subsequently
reused, so that when combined with "Reuse HF," the total reuse rate was approximately 90% in 2013. "Reuse HF"
indicates on-site reuse. Source: Waste data from PA DEP (2015a).
Text Box 8-2. Regulations Affecting Wastewater Management.
Regulations affect wastewater management options and vary geographically as well as over time. At the
Federal level, the EPA has promulgated national technology-based regulations, known as effluent limitations
guidelines and standards (ELGs], for the oil and gas extraction industry, which can be found in 40 U.S. Code of
Federal Regulations (CFR] Part 435. These ELGs do not apply to CBM discharges which are subject to
technology based limits developed by permit writers on a case-by-case "best professional judgment" basis.
The Onshore subcategory of the oil and gas, ELGs 40 CFR 125.3, Subpart C, prohibits the discharge of
wastewater pollutants to waters of the United States from onshore oil and gas extraction facilities, with one
exception in the arid west as discussed below. This "zero-discharge standard" means that, unless the
exception applies, oil and gas wastewater pollutants cannot be discharged directly to waters of the United
States. Operators have met this requirement through underground injection, reuse, or transfer of wastewater
to POTWs and/or CWTs. The EPA finalized a rule in June 2016 that would prohibit operators from sending
wastewater from unconventional oil and gas extraction to POTWs. Operators can continue to send
wastewater to CWTs, which are subject to regulation under a separate set of ELGs in 40 CFR Part 437.
In addition, Subpart E of the oil and gas ELGs establishes an exception to the zero discharge standard west of
the 98th meridian (the arid western portion of the continental United States), allowing discharges of
produced water from onshore oil and gas extraction facilities to waters of the United States if the produced
water has a use in agriculture or wildlife propagation when discharged into navigable waters. The term "use
in agricultural or wildlife propagation" means that: (1) the produced water is of good enough quality to be
used for wildlife or livestock watering or other agricultural uses; and (2) the produced water is actually put to
[Text Box 8-2 is continued on the following page.)
8-21
-------
Chapter 8 - Wastewater Disposal and Reuse
Text Box 8-2 (continued). Regulations Affecting Wastewater Management.
such use during periods of discharge (40 CFR 135.51(c)). Produced water discharged under this exception
must not exceed an oil and grease concentration of 35 milligrams per liter (mg/L). Subpart E does not allow
for discharge from sources other than produced water (i.e., drilling muds, drill cuttings, produced sands) to
waters of the United States.
In addition to the technology-based limitations discussed above, the Clean Water Act (CWA) and the EPA's
implementing regulations also require that permits include more stringent limits as necessary to meet
applicable water quality standards. CWA Section 301(b)(1)(C); 40 CFR 122.44(d)(1).
100% Commercial
disposal facility
90%
Surface discharge
I I I I I ¦ -
l l I I l
H- Injected on lease
:l I I I
1999-2001 2002-2004 2005-2007 2008-2010 2011-2013
Figure 8-5. Management of wastewater in Colorado in regions where hydraulic fracturing is
being performed.
See footnote for details on disposal codes.1 Production data from Colorado Oil and Gas Conservation Commission
(COGCC, 2015).
The following sections provide an overview of hydraulic fracturing wastewater management
methods, with some discussion of the geographic and temporal variations in practices and their
impacts on drinking water resources. In addition to currently used treatment and disposal
methods, this section also briefly describes past treatment of hydraulic fracturing wastewater at
1 Codes for wastewater disposal from COGCC are described by Veil T20151 as follows:
• Commercial disposal facility: water sent to commercial pits.
• On-site pit: most water evaporates, or excess water is hauled to disposal wells.
• Central disposal pit: Central facilities owned by a single producer to handle water from multiple wells (some
recycled, much is injected].
• Injected on lease: Injected into wells, roughly half for enhanced recovery.
• Surface discharge: water is either fresh or treated to acceptable standards and discharged to surface water.
8-22
-------
Chapter 8- Wastewater Disposal and Reuse
POTWs. More in-depth descriptions of treatment technologies applicable to hydraulic fracturing
wastewater are available in Appendix F.
8.4.1 Underground Injection
Oil- and gas-related wastewater may be disposed of via Class II injection wells (disposal wells are
referred to as Class IID whereas enhanced recovery wells are referred to as Class IIR) regulated by
the UIC Program under the SDWA.1 Nationwide, injection wells receive a large percentage of
wastewater from the oil and gas industry, including wastewater associated with hydraulic
fracturing. Veil (2015) estimates that in 2012, U.S. oil and gas production from onshore wells
generated over 863 billion gal (20.56 billion bbls or 3.27 trillion L) of produced water, and of that
volume, information on management was available for 97%. The study estimated that about 93%
was injected into Class II wells, with about 47% injected into Class IID wells and 46% injected into
Class IIR wells.2
The above national estimates are for the oil and gas industry as a whole. A good national estimate of
the amount of hydraulic fracturing wastewater injected into Class II wells is difficult to develop due
to lack of available information and data on injection of hydraulic fracturing wastewater.
Management of hydraulic fracturing wastewater is not well tracked or made publicly available in
many states (Pennsylvania being a notable exception). The local availability of Class IID wells along
with generally low reuse rates, however, are consistent with Class IID wells being a primary means
of wastewater management in many areas with hydraulic fracturing activity.
According to recently released data from 2012 and 2013, there are about 26,400 active Class IID
wells in the United States, with more than 65% of these located in Texas, Oklahoma, and Kansas
(Table 8-4). In Pennsylvania, on the other hand, there are currently nine operating disposal wells,
and only three of these are commercially operated wells (at one facility) (SAFER PA. 2015). The
location and number of Class IID wells is in part determined by geology (including depth and
permeability of geologic formations appropriate for injection), permitting, and historical demand
for disposal of oil and gas wastewater. The large Class IID well capacity in Texas, for example, is
consistent with the availability of formations with suitable geology and the demand for wastewater
disposal associated with a mature and active oil and gas industry. In contrast, injection plays a
relatively small role in Marcellus Shale wastewater management in Pennsylvania—about 10% in
2013 and the first half of 2014 CPA DEP. 2015a1.
1 States may be given federal approval to run a UIC program under SDWA. UIC Class II wells include those used for
disposal (Class IID], enhanced oil recovery (Class IIR], and hydrocarbon storage (Class IIH).
2 Because some states surveyed by Veil (2015] do not distinguish between volumes injected for disposal versus enhanced
recovery, assumptions and analyses were used in the study to estimate the two types of injection in some states.
8-23
-------
Chapter 8 - Wastewater Disposal and Reuse
Table 8-4. Distribution of active Class IID wells across the United States.
Data are primarily from 2012 and 2013. Source: U.S. EPA (2016d).
Geographic region
(from the EIA)
State
Number of active
disposal wells3
Average disposal
rate per well
(gpd/well)b
State disposal
rate (MGD)
Alaska
Alaska
45
182,000
8.2
East
Illinois
1,054
C
C
Michigan
772
16,200
13
Florida
14
246,000
3.4
Indiana
208
7,950
1.7
Ohio
190
8,570
1.6
West Virginia
64
6,970
0.45
Kentucky
58
4,650
0.27
Virginia
12
17,500
0.21
Pennsylvania
9
6,380
0.057
New York
10d
33.7
0.00034
Gulf Coast/Southwest
Texas
7,876
52,100
410
Louisiana
2,448
40,300
99
New Mexico
736
48,600
36
Mississippi
499
24,200
12
Alabama
85
53,300
4.5
Mid-Continent
Kansas
5,516
25,600
140
Oklahoma
3,837
35,900
140
Arkansas
640e
25,400
16
Nebraska
113
19,100
2.2
Missouri
11
2,270
0.025
Iowa
3
C
C
Northern Great Plains
North Dakota
395
53,300
21
Montana
199
32,700
6.5
South Dakota
15
17,400
0.26
8-24
-------
Chapter 8- Wastewater Disposal and Reuse
Geographic region
(from the EIA)
State
Number of active
disposal wells3
Average disposal
rate per well
(gpd/well)b
State disposal
rate (MGD)
Rocky Mountains
Wyoming
335
107,000
36
Colorado
292
48,800
14
Utah
118
83,400
9.8
West Coast
California
826
86,800
72
Nevada
10
54,600
0.55
Oregon
9
C
C
Washington
1
C
C
Total
26,400
41,300
1,050
Abbreviations: gpd—gal per day; MGD—million gal per day.
a Number of active disposal wells is based primarily on data from 2012 to 2013.
b Typical injection volumes per well are based on historical annual volumes for injection for disposal divided by the number of
active disposal wells during the same year (primarily 2012 to 2013 data).
c Disposal rates and volumes are unknown.
d These wells are not currently permitted to accept extraction wastewater from production in unconventional reservoirs.
6 Only 24 of the 640 active disposal wells in Arkansas are in the northern half of the state, close to the Fayetteville Shale.
The decision to inject hydraulic fracturing wastewater into Class IID wells depends in part on cost,
including transportation costs. Therefore, the distance between the production well and a disposal
well is an important consideration. For oil and gas producers, underground injection is a low cost
management strategy unless significant trucking is needed to transport the wastewater to a
disposal well fU.S. GAP. 20121.
Evaluation of documented or potential impacts on drinking water resources associated with
disposal at Class IID injection wells is outside of the scope of this assessment. However, disposal
wells play a significant role in the overall management of hydraulic fracturing water nationwide,
and their availability and capacity are integral factors in determining which wastewater
management strategies are used by operators in a given region. Should the feasibility of managing
wastewater via underground injection become limited or less economically advantageous,
operators will need to adjust their wastewater management programs. They may evaluate and
implement other local practices such as sending wastewater to a CWT for treatment and discharge
or reuse.
Recent events and studies, for example, have documented a link between wastewater injection and
seismic activity in some locations in several states, including Oklahoma, Colorado, New Mexico,
Arkansas, and Ohio fWeingarten etal.. 2015: Wongetal.. 20151. The Oklahoma Geological Survey
f Andrews and Holland. 20151 "considers it very likely that the majority of recent earthquakes,
particularly those in central and north-central Oklahoma, are triggered by the injection of produced
water in disposal wells." Walsh and Zoback (20151 correlated wastewater injection from
8-25
-------
Chapter 8 - Wastewater Disposal and Reuse
production wells (including hydraulically fractured wells) into Oklahoma's Arbuckle formation to
the steep increase in seismic events observed in that state. Farther west, in the Raton Basin of
southern Colorado and northern New Mexico, Rubinstein et al. f20141 presented several lines of
evidence linking injection well disposal of CBM produced water to seismic events. Horton f20121
attributed a swarm of earthquakes in Northern Arkansas to hydraulic fracturing wastewater
injection, and in a study evaluating multiple states in the mid-continent region, Weingarten et al.
f20151 demonstrated a relationship between Class II wells (including both Class IID and Class IIR
wells) and seismicity.
The local availability of Class IID wells and the capacity to accept large volumes of wastewater
could be affected by these recent findings concerning seismic activity associated with injection fU.S.
EPA. 2014cl. Between 2011 and 2016, some state UIC programs modified their Class II wastewater
injection regulations and permitting requirements. At least eight states (Arkansas, Colorado,
Illinois, Kansas, Ohio, Oklahoma, Texas, and West Virginia) consider an assessment of seismicity in
their Class II programs and have regulatory provisions for banning or shutting injection wells
and/or modifying injection volumes and pressures if evidence indicates that a well is near a fault
and/or is contributing to seismic activity.
As an example, Oklahoma has recently taken steps to reduce the risk of induced seismicity by
implementing a regional strategy intended to reduce wastewater injection in certain regions (OCC
OGCD. 20161. These actions affect over 10,000 square miles and 600 wastewater injection wells in
western and central Oklahoma. The measures are intended to reduce wastewater injection in the
area by 40% below 2014 totals, which will affect wastewater management and disposal practices
across this large area.1
In terms of potential impacts on drinking water resources, Class IID facilities are subject to the
same general considerations regarding wastewater storage and handling as other wastewater
management sites and facilities (e.g., CWTs). Changes in surface water or groundwater quality due
to general wastewater handling at these facilities may be another factor affecting wastewater
management practices in some locations or regions. For example, Kell f20111 identified eight
groundwater contamination incidents in Texas between 1993 and 2008 due to water releases from
storage facilities associated with Class II well sites. A recent study by the United States Geological
Survey documented impacts on surface water from hydraulic fracturing wastewater at a Class II
disposal well site in central West Virginia fAkob etal.. 20161. Water samples collected downstream
from the facility were indicative of wastewater from hydraulic fracturing operations handled at the
site. The authors documented elevated specific conductance and elevated TDS, sodium, chloride,
barium, bromide, strontium, and lithium concentrations, and different strontium isotope ratios
compared to those found in upstream, background waters. The study concluded that activities at
the wastewater facility have affected water quality in a nearby stream. The pathways for the
movement of wastewater into the local stream include several possibilities (e.g., leaks from storage
ponds and tanks, transportation activities, previous site history).
1 For additional information on strategies and initiatives regarding wastewater injection and inducted seismicity, see the
following: KDHE ("20141. States First Initiative ("20141. and U.S. EPA (~2014cl.
8-26
-------
Chapter 8- Wastewater Disposal and Reuse
8.4.2 Publicly Owned Treatment Works
POTWs are designed to treat local municipal wastewater and indirect discharges from industrial
users. POTWs are also used to treat wastewater and other wastes from oil and gas operations in
some eastern states. Although this is not a common method of treatment for oil and gas
wastewaters in the United States, the scarcity of injection wells for waste disposal in Pennsylvania
drove the need for disposal alternatives (Wilson and Vanbriesen. 2012). When development of the
Marcellus Shale began, POTWs were used to treat wastewater originating from these oil and gas
wells fKappel etal.. 2013: Soeder and Kappel. 20091. However, elevated concentrations of
constituents in wastewater from the Marcellus region (halides, heavy metals, organic compounds,
radionuclides, and salts) can pass through the treatment processes commonly used in POTWs and
be discharged to receiving waters fCusick. 2013: Kappel. 2013: Lutz etal.. 2013: Schmidt. 20131. In
addition, sudden, extreme salt fluctuations can disturb POTW biological treatment processes
(Linaric etal.. 2013: Lefebvre and Moletta. 2006).
The annual reported volume of oil and gas wastewater treated at POTWs in the Marcellus Shale
region peaked in 2008 and has since declined significantly (Figure 8-6). As discussed in Text Box
8-1, this was in response to an April 2011 request from PADEP asking operators to cease sending
Marcellus Shale wastewater to 15 POTWs and CWTs that were exempt from the 2010 TDS
regulation fRahm etal.. 20131. Although operators complied with the request in May 2011, non-
Marcellus oil and gas produced water continued to be processed at these facilities fFerrar etal..
2013: Lutz etal.. 2013: Wilson and Vanbriesen. 20121.1 In August 2016, the EPA finalized
pretreatment standards prohibiting discharges of unconventional wastewater pollutants to POTWs
fU.S. EPA. 2016dl.
140
120
"O
QJ
CJjO
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
¦ Marcellus (Mgal) ¦ Conventional (Mgal)
Figure 8-6. Oil and gas wastewater volumes discharged to POTWs from 2001-2011 in the
Marcellus Shale. ("Conventional" is indicated by the authors as non-Marcellus wells and
described as vertically drilled to shallower depths in more porous formations.)
Due to an unrecoverable data loss at the PA DEP, records for 2007 were not available. Source: Lutz et al. (2013).
1 POTWs in Pennsylvania have likely been accepting waste considered conventional by Pennsylvania but unconventional
by others based on the EPA's broader definition (Text Box 8-1].
8-27
-------
Chapter 8 - Wastewater Disposal and Reuse
8.4.3 Centralized Waste Treatment Facilities
A CWT facility is generally defined as one that accepts industrial materials (hazardous or non-
hazardous, solid, or liquid) generated at another facility (off-site) for treatment or recovery (EPA,
2000). (Wastewater may also be treated at on-site mobile or semi-mobile facilities; see Appendix F
for additional information.) The decision to treat hydraulic fracturing wastewater at a CWT and the
level of treatment used depends upon several factors, such as a lack of proximity to Class II disposal
wells; whether the wastewater might be reused for additional hydraulic fracturing jobs; the water
quality needed if it will be reused; whether the treated wastewater can be discharged under the
Subpart E agricultural and wildlife use exception in the arid west; and the water quality needed if it
will be discharged to the waters of the United States. As a group, CWTs that accept oil and gas
wastewater offer a wide variety of treatment capabilities and configurations (Text Box 8-3 and
Appendix F).
Text Box 8-3. Wastewater Treatment Processes.
The constituents prevalent in hydraulic fracturing wastewater include TDS, TSS, radionuclides, organic
compounds, and metals (Section 8.3 and Chapter 7). If the ultimate disposal or use of the wastewater
necessitates treatment, a variety of technologies can be employed to remove or reduce these constituent
concentrations.
The most basic treatment needed for oil and gas wastewaters, including those from hydraulic fracturing
operations, is separation to remove TSS and oil and grease. This is accomplished through separation
technologies including settling, skimming, hydrocyclones, dissolved air or induced gas flotation, media
filtration, or biological aerated filters flgunnu and Chen. 2014: Duraisamv et al.. 2013: Barrett. 2010:
Shammas. 2010).
Other treatment processes that may be used include media filtration after chemical precipitation for hardness
and metals (Boschee. 2014): adsorption technologies for organics, heavy metals, and some anions flgunnu
and Chen. 2014): a variety of membrane processes (microfiltration, ultrafiltration, nanofiltration, reverse
osmosis (RO)); and distillation technologies for metals and organics fDrewes et al.. 2009).
Advanced processes, such as RO, or distillation methods, such as mechanical vapor recompression (MVR), are
needed if the system requires significant reduction in TDS fDrewes et al.. 2009: LEau LLC. 2008: Hamieh and
Beckman. 2006). However, RO is typically only capable of treating TDS concentrations less than 35,000 mg/L
fShaffer et al.. 20131 whereas distillation can effectively treat higher TDS waters fHaves et al.. 2014: Drewes
et al.. 2009). Extremely high TDS waters may require a series of advanced treatment processes, which can be
very costly.
An emerging technology in hydraulic fracturing wastewater treatment is electrocoagulation, which has been
used in mobile treatment systems to remove organics, TSS, and metals (Halliburton. 2014: Igunnu and Chen.
20141
Appendix F provides more in-depth descriptions of technologies used to treat for hydraulic fracturing
wastewaters and the constituents they remove. Also, Appendix Table F-4 provides an overview of influent
and effluent results and removal percentages for constituents of concern at oil and gas treatment facilities
reported in the literature (both conventional and unconventional) and the specific technology(ies) used to
remove them. Section 8.4.7 discusses solid and liquid residuals, including treatment-related wastes.
8-28
-------
Chapter 8- Wastewater Disposal and Reuse
The treated effluent from a CWT can be reused in hydraulic fracturing operations (also called zero-
discharge), discharged directly to a receiving water under a National Pollutant Discharge
Elimination System (NPDES) permit, discharged indirectly to a POTW, or a combination of these.
Some CWTs maybe configured so that they can either (1) partially treat the waste stream to suit
the needs of operators who reuse it or (2) use more advanced treatment (i.e., TDS removal) if the
treated wastewater will be discharged. Generally, the former option is less costly for the CWT, and
some facilities that have permits to discharge do not do so continuously, opting to direct as much of
the wastewater as possible for reuse. There are also CWTs permitted to discharge that do not have
TDS removal capabilities. However, these facilities must still meet TDS discharge limits specified by
their state. Appendix F contains additional information on treatment configurations, including
examples of processes at several facilities treating oil and gas wastewater.
Facilities discharging treated wastewater to waters of the United States or POTWs are regulated
under the Clean Water Act (CWA). For zero-discharge facilities, some states, including Pennsylvania
and Texas, have adopted regulations to control permitting of these facilities or to encourage
treatment and reuse. The PA DEP issues permits that allow zero-discharge CWTs to treat and
release water back to oil and gas industries for reuse (see the Eureka Resources Facility in
Williamsport, PA listed in Appendix Table F-6 as an example of a zero-discharge facility).1
In developing this assessment, we looked at NPDES permit information for several CWTs in the
eastern United States treating wastewater from the Marcellus region and one near the Fayetteville
Shale in Arkansas. The facilities include those with and without TDS removal capabilities, and some
are undergoing upgrades to implement TDS removal. Some of the permits reviewed for this
assessment are current, and others are expired and may be in the process of renewal. The permits
require monitoring (with or without limits) for a range of constituents that may include chloride,
TDS, TSS, total strontium, total barium, oil and grease, heavy metals, 5-day biological oxygen
demand (B0D5), and a range of organic compounds (e.g., phenol, cresol, BTEX, phthalates), with the
specific constituents varying by permit Sample types for the facilities are generally 24-hour
composites. The newer permits set limits for several important constituents such as chloride, TDS,
TSS, total barium, total strontium, oil and grease, and a number of heavy metals. Bromide is
generally either not included or is required to be reported but with no limit specified. However,
limits on TDS will reduce bromide concentrations. Some permits require monitoring for total
radium, uranium, and gross alpha, but no limits are specified. Note that these facilities do not
necessarily discharge consistently because treated wastewater can be sent for reuse.
Although there are CWTs serving hydraulic fracturing operations throughout the country, the
majority serve Marcellus Shale operations in Pennsylvania (Boschee. 2014). Of the 74 CWT facilities
identified by the EPA (U.S. EPA. 2016d) as having accepted or having the ability to accept hydraulic
fracturing wastewater (not counting facilities treating CBM wastewater), 40 are located in
Pennsylvania (Table 8-5). Most are zero-discharge facilities, and many do not have treatment
processes for TDS removal. Although several Pennsylvania facilities are permitted to discharge,
Wunz (2015) found few that currently discharge (two CWTs in Pennsylvania, one in West Virginia,
1 The facility is also permitted for indirect discharge to the Williamsport Sewer Authority.
8-29
-------
Chapter 8 - Wastewater Disposal and Reuse
Table 8-5. Number, by state, of CWT facilities that have accepted or plan to accept wastewater from unconventional oil and gas
activities.
Source: U.S. EPA (2016d).
State
Unconventional
formation(s) served
Zero discharge CWT
facilities3
CWT facilities that discharge
to a surface water or POTWa
CWT facilities with multiple
discharge options3
Total known
facilities
Non-TDS
removal
TDS
removal
Non-TDS
removal
TDS
removal
Non-TDS
removal
TDS
removal
AR
Fayetteville
2
0
0
0
0
1
3
CO
Niobrara, Piceance Basin
3(1)
0
0
0
0
0
3
ND
Bakken
0
1(1)
0
0
0
0
1
OH
Utica, Marcellus
10 (7)
0
1
0
0
0
11
OK
Woodford
2
0
0
0
0
0
2
PA
Utica, Marcellus
22
7(3)
8
0
0
3(1)
40
TX
Eagle Ford, Barnett, Granite Wash
1
3
0
0
0
0
4
WV
Marcellus, Utica
4(2)
0
0
0
1
1
6
WY
Mesaverde and Lance
0
2
0
0
0
2
4
Total
44
13
9
0
1
7
74
a Information is current as of 2014; it is possible that since 2014, some listed CWT facilities have closed and/or new CWT facilities not listed have begun operation. The number
of facilities includes those that have not yet opened but are under construction, pending permit approval, or are in the planning stages. Facilities that are not accepting hydraulic
fracturing wastewater but plan to accept it in the future are noted parenthetically and not included in the sum of total known facilities. Facilities handling CBM wastewater are
not represented here.
8-30
-------
Chapter 8- Wastewater Disposal and Reuse
and one in Ohio). According to EPA research (U.S. EPA. 2016d). the number of CWT facilities serving
operators in the Marcellus and Utica Shales has increased since the mid-2000s, growing from
roughly five in 2004 to over 40 in 2013. A similar trend has been noted for the Fayetteville Shale
region in Arkansas, where there are fewer Class IID injection wells compared to the rest of the state
("U.S. EPA. 2016dl.
In other regions, a small number of newer facilities have emerged in the last several years, most
often with TDS removal capabilities. In Texas, for example, two zero-discharge facilities with TDS
removal capabilities are available to treat wastewater from the Eagle Ford Shale (beginning in 2011
and 2013), and one zero-discharge facility with TDS removal is located in the Barnett Shale region
(operational since 2008). In Wyoming, there are four facilities in the region of the Mesaverde/Lance
formations that started operating between 2006 and 2012. Two are zero-discharge facilities, and
two have multiple discharge options; all are capable of TDS removal (U.S. EPA. 2016d).
Few states maintain a comprehensive list of CWT facilities, and the count provided by the EPA (U.S.
EPA. 2016dl includes facilities that do not currently but plan to accept wastewater from
unconventional formations. Therefore, the data in Table 8-5 do not precisely reflect the number of
facilities currently handling hydraulic fracturing wastewaters. Other sources indicate either use of,
or interest in, development of treatment facilities in other regions such as the Barnett Shale region
(Haves and Severin. 2012b). the Fayetteville (Veil. 2011). and other areas in Texas and Wyoming
fBoschee. 2014. 20121. In addition, news releases and company announcements indicate that other
wastewater treatment facilities are being planned fGreenhunter. 2014: Geiver. 2013: Purestream.
2013: Alanco. 2012: Sionix. 2011).
Use of specific types of CWTs has and will continue to shift due to drivers such as availability and
cost of other disposal options (e.g., disposal wells), operator demand for reuse and the associated
quality needed, developments in treatment, treatment costs, and regulatory changes. Practices in
Pennsylvania over the last several years provide such an example. Between 2010 and 2013, the
percentage of Marcellus wastewater treated at CWTs dropped from 52% to 20% (Figure 8-4), and
the percentage of wastewater reused on-site rose to 65%, reflecting a shift in practice among
operators. Among the percentage of the wastewater sent to CWTs, the portion sent to zero-
discharge facilities for subsequent reuse rose from 10% to 65%. This is consistent with an
increased emphasis on reuse in Pennsylvania. (See Section 8.4.4 for a discussion on reuse as a
waste management practice.)
8.4.3.1 Relationship to Potable Surface Waters
Figure 8-7 shows the relationship between Pennsylvania potable water supplies and the CWTs that
lie in their upstream watersheds. These surface waters, including streams, rivers, and waterbodies
(e.g., lakes and reservoirs) have been evaluated by the PA DEP for attainment of a designated use of
potable water supply as per the CWA Section 305(b) reporting and Section 303(d) listing. Ninety-
four percent of the waterbodies and 98% of the streams and rivers were attaining their designated
use in 2016. These stream segments may or may not currently have intakes for drinking water
treatment plants. The map also shows the locations and types of CWTs that either currently accept
unconventional oil and gas wastewater (as defined by PA DEP) or have accepted such wastewater
8-31
-------
Chapter 8 - Wastewater Disposal and Reuse
WW:
nISBURQ
i - -;a
20 30 40
Centralized Waste Treatment
All Streams
O Non-Discharger
Stream size (order)
0 Discharger: Direct
1
~ Discharger: Indirect
— 2
+ Discharger: Unknown
— 3
Potable Water Supply
# CWTs in upstream watershed
— 6
0
7
^—8
— 2
Reference
— 4
• Major cities
5-6
^—7-8
1 1 Pennsylvania outline
l
-------
Chapter 8- Wastewater Disposal and Reuse
within approximately the last five years.1 CWTs represented include both dischargers (direct and
indirect) as well as zero-discharge facilities. For some facilities, we were not able to determine if the
facility was zero-discharge or if it has a NPDES permit The surface waters have been color-coded to
indicate the number of CWTs that are located upstream. Darker red indicates more CWTs located in
the upstream watershed, while blue indicates no upstream CWTs. Softer grey lines show portions of
the stream network not designated for potable water supply. The thickness of the line indicates the
size of the stream or river, categorized by the "stream order" designation.
The map provides a general illustration of how CWTs are situated within catchments in
Pennsylvania, showing their spatial and general hydrologic relationships to streams that can serve
as potable water supplies. The map shows that a given stream or waterbody may have a number of
CWTs upstream, potentially contributing to combined impacts on surface water if there are spills or
inadequately treated discharges. Note that the upstream catchment areas are large for the major
rivers. Therefore, some rivers, such as the Ohio or Susquehanna, have as many as 15 or 16
upstream CWTs, although most are located far away. The map does not represent the effects of
dilution on either discharges or spills; such an evaluation would necessitate currently unavailable
data required to do a complete analysis of water quality. Note that many of the CWTs are zero-
discharge facilities, and those that are permitted to discharge may do so intermittently. However,
the storage and handling of wastewater at CWTs could impact nearby surface water through leaks
and spills.
To more completely place these facilities in a watershed context, other types of discharges that
could be occurring upstream should be taken into consideration. Impacts from hydraulic fracturing
wastewater may be more problematic if there are additional pollutant sources within the
watershed, increasing the cumulative effects of discharges and spills. For example, an EPA source
apportionment study fU.S. EPA. 2015ol evaluated the relative contributions of bromide, chloride,
nitrate, and sulfate from CWTs primarily treating hydraulic fracturing wastewater to the Allegheny
River Basin and to two downstream public water system intakes. The study considered that the
Allegheny River and its tributaries also receive runoff and discharges from an array of sources that
include acid mine drainage and mining operations, coal-fired electric power stations, industrial
wastewater treatment plants, and POTWs. It was concluded that CWTs treating oil and gas
wastewater and coal-fired power plants with flue gas desulfurization were the primary
contributors of bromide and chloride at the intakes (see Section 8.5.1 for further discussion), while
nitrate and sulfate contributions were from POTWs and Acid Mine Drainage fU.S. EPA. 2015ol
8.4.3.2 Potential Impacts from CWTs
The potential impacts of managing hydraulic fracturing wastewater at CWTs depend on whether
the CWT adequately treats for constituents of concern prior to discharge to surface water or a
POTW, and whether treatment residuals are managed appropriately. Historically, CWTs have not
1 The list of CWTs used to develop this map is based on best available data, including information in the technical
development document supporting the new EPA unconventional oil and gas effluent limitation guidelines HJ.S. EPA.
2016d] as well as data from PA DEP waste records. This information was supplemented with other publicly available
descriptions of the facilities. The information may, however, not be complete, and the symbols may not definitively reflect
the discharge status of a facility.
8-33
-------
Chapter 8 - Wastewater Disposal and Reuse
included processes to treat for constituents that are difficult to remove, such as the high
concentrations of TDS found in wastewater from unconventional reservoirs. As a result, impacts on
drinking water resources have included increased suspended solids and chloride concentrations
downstream of discharging facilities that were accepting hydraulic fracturing wastewater
(Olmstead etal.. 2013) and elevated bromide concentrations and radium concentrations in CWT
effluent (Warner etal.. 2013a): see Sections 8.5.1 and 8.5.2. In addition, spills and leaks can occur in
pits or impoundments associated with the storage of treated wastewater at CWTs (impacts related
to spills and leaks from pits and impoundments are discussed in Section 8.4.5). Wastewater being
transported by truck or pipeline to and from a CWT can also present a vulnerability for spills or
leaks (Easton. 20141 (Chapter 7).
While selection of appropriate treatment processes is critical for CWTs that discharge to surface
waters, there are also two important issues related to completeness of treatment that can have an
impact First, there may be unknown constituents in the wastewater. The effectiveness of treatment
cannot be evaluated for constituents for which the wastewater has not been tested. This makes it
challenging to know the degree to which effluent from a CWT is protective of public health. Second,
even an efficient treatment process may not be able to reduce the concentrations of some
constituents to levels that allow for discharge to a drinking water resource if influent
concentrations are so high that they exceed the capabilities of the treatment technology(ies) to
meet those discharge limits. For example, a facility described by Kennedv/lenks Consultants (2002)
removed a high percentage of boron (88%), but the effluent concentration of 1.9 mg/L (average
influent concentration of 16.5 mg/L) was not low enough to meet California's action level of 1
mg/L. Thus, the influent concentration must be considered together with removal efficiency to
determine whether the effluent quality will meet the requirements dictated by end use or by
regulations.
Relatively few studies describe the ability of individual treatment processes to remove constituents
from hydraulic fracturing wastewater. For this assessment, simple estimated effluent
concentrations were calculated for several combinations of unit treatment processes, wastewater
constituents, and influent concentrations (details are given in Appendix Table F-3). The purpose of
the analysis was to illustrate the relative capabilities of a number of treatment processes and not to
represent a complete treatment system. As an example, the estimates suggest that if wastewater
contains radium with a concentration in the thousands of pCi/L, a 95% removal rate with chemical
precipitation may result in an effluent that exceeds 100 pCi/L. Treatment of the same wastewater
via distillation or reverse osmosis could result in effluent concentrations in the tens of pCi/L. This
analysis suggests that attention should be paid to the capabilities of a planned treatment system for
the full range of anticipated wastewater compositions.
To gain a better understanding of impacts, the USGS has conducted sampling for a wide array of
water quality parameters in surface water and groundwater in the Monongahela River Basin in
West Virginia to establish baseline water-quality conditions (Chambers etal.. 2014). Future water
quality sampling can be compared to this baseline to assess impacts from hydraulic fracturing
activities. To address past impacts, Pennsylvania, having experienced water quality impacts on
receiving streams due to discharges of high-TDS effluent modified their regulations to address
8-34
-------
Chapter 8- Wastewater Disposal and Reuse
these issues by setting water quality standards for CWT dischargers fMauter and Palmer. 2014:
Shaffer etal.. 20131. (See Text Box 8-1.)
8.4.4 Wastewater Reuse for Hydraulic Fracturing
The reuse of hydraulic fracturing wastewater for subsequent hydraulic fracturing operations has
increased in some regions of the country in recent years (Boschee. 2014. 2012: Gregory etal.. 2011:
Rassenfoss. 20111.1 This practice is driven by factors that include cost (including treatment costs),
the lack of availability of other management options (e.g., Class II disposal wells), and changes to
state regulations (Boschee. 2014: Shaffer etal.. 2013). Wastewater quality is a consideration; some
constituents pose challenges for reuse and may necessitate treatment For example, high
concentrations of barium and sulfate can lead to scaling, and the presence of some constituents in
wastewater can hinder crosslinking fAkob etal.. 2016: Boschee. 20141. Hydraulic fracturing fluid
formulations that can use high TDS waters (e.g., as high as 150,000 mg/L to over 300,000 mg/L)
facilitate reuse with minimal treatment (Boschee. 2014: Mauter and Palmer. 20141. See Chapter 5
for more information regarding the chemical composition of hydraulic fracturing fluids and
Appendix F for more discussion of considerations for reuse.
Reuse can be accomplished by blending either untreated or minimally treated hydraulic fracturing
wastewater with fresh water to lower the TDS content (Boschee. 20141. Wastewater may be reused
at a site with multiple wells, eliminating the need for transport to a CWT f Lester etal.. 2015: Easton.
20141. Alternatively, wastewater can be treated at a CWT and then taken by operators for mixing
with other water sources for reuse (Easton. 2014). Flowback may be preferable to later-stage
produced water for reuse because of its lower TDS concentration. Also, it is typically generated in
larger quantities from a single location as opposed to water produced later on, which is generated
in smaller volumes over time from many different locations fBarbotetal.. 2013: Malonev and
Yoxtheimer. 20121. Reuse can reduce the costs associated with water acquisition and produced
water management. Such economic and logistical benefits can be expected to inform ongoing
wastewater management decisions.
Costs can be the most significant driver for reuse. For example, the costs of transporting
wastewater from the generating well to the treatment facility and then to the new well can be
weighed against the costs for transport to alternative locations (e.g., a disposal well). Trucking large
quantities of water can be relatively expensive—from $0.01 to $0.19 per gallon ($0.50 to $8.00 per
bbl)—rendering on-site treatment technologies and reuse economically competitive in some
settings fDahm and Chapman. 2014: Guerra etal.. 20111. Reuse rates may also be driven by
wastewater production rates compared to the demand for reuse, with both production and demand
increasing in a region if more wells go into production or decreasing as plays mature (Lutz etal..
2013: Hayes and Severin. 2012b: Slutz etal.. 20121. Other logistics to consider include proximity of
the water sources for aggregation and sequencing of completion schedules fMauter and Palmer.
1 Reused hydraulic fracturing wastewater is discussed in Chapter 4 of this report (Water Acquisition] as well as in this
chapter, though in a different context. The wastewater reuse rate described in this chapter is the amount or percentage of
generated hydraulic fracturing wastewater that is managed through use in subsequent hydraulic fracturing operations. In
contrast. Chapter 4 discusses reused wastewater as a source water and as one part of the base fluid for new fracturing
fluid.
8-35
-------
Chapter 8 - Wastewater Disposal and Reuse
2014]. A small survey by Mauter and Palmer [2014] indicates that the scheduling of well
completions is complex, requiring optimization of labor, contractual issues, equipment usage, and
water storage capacity among other factors. Boschee ("2014] notes that in the Permian Basin, older
conventional wells are linked by pipelines to a central disposal facility, facilitating movement of
treated water to areas where it is needed for reuse. Companies drilling fewer wells or located in
more remote areas may find reuse difficult because of challenges in consolidating wastewater from
their wells or accessing wastewater from centralized facilities.
Regulations may also encourage reuse. For example, in 2013, the Texas Railroad Commission
adopted rules eliminating the need for a permit when operators reuse on their own lease or
transfer the fluids to another operator for reuse ("Rushton and Castaneda, 2014], Any information
on wastewater management practices in Texas that becomes available for the years after 2013 will
allow evaluation of whether reuse has in fact increased.
A summary of reuse practices throughout the United States is hampered by the limited amount of
data available for many regions of the country. However, current data indicate that extensive reuse
takes place in the Marcellus region. Several studies using data from PA DEP data show that total
reuse rates of oil and gas wastewater in Pennsylvania have risen over the last several years to
between 85 and 90% (Table 8-6], This includes wastewater sentto CWTs to treat for reuse as well
as reuse at the well sites without transfer to a CWT (labeled as "Reuse HF" in Figure 8-4], In
particular, reuse of Marcellus wastewater at well sites in Pennsylvania has risen from about 8% in
the second half of 2010 to nearly 70% in the first half of 2014 (PA DEP, 2015a], Schmid and
Yoxtheimer (2015] report more recent data stating that in 2014, approximately 85% of Marcellus
hydraulic fracturing wastewater was reused. Of that amount, 78% occurred on-site, and 22% was
via CWTs.
Table 8-6. Estimated percentages of reuse of hydraulic fracturing wastewater.
Play or basin
Source and year
2008
2009
2010
2011
2012
2013
2014
East Coast3
Marcellus, PA
Rahm et al. (2013)
9
8
25-48
67-80
Marcellus, PA
Ma et al. (2014)
15-20
90
Marcellus, PA
Shaffer et al.
(2013)
90
Marcellus, PA
Schmid and
Yoxtheimer (2015)
85
Marcellus, PA
Hansen et al.
(2013)
9
6
20
56
Marcellus, PA
Malonev and
Yoxtheimer (2012)
71.6
Marcellus, PA
Tiemann et al.
(2014)
72
87
8-36
-------
Chapter 8- Wastewater Disposal and Reuse
Play or basin
Source and year
2008
2009
2010
2011
2012
2013
2014
Marcellus, PA
Rassenfoss (2011)
~67
(general
estimate)
96 (one
company)
Marcellus, PA
Wendel (2011)
75-85
90
Marcellus, PA
Lutz et al. (2013)
13 (prior to 2011)
56
Marcellus, PA
(SW region)
Rahm et al. (2013)
~10
~15
~25-45
-70-80
Marcellus, PA
(NE region)
Rahm et al. (2013)
0
0
-55-70
-90-100
Marcellus, WV
Hansen et al.
(2013)
88
73
65 (partial
year)
Gulf Coast and Midcontinent
Fayetteville
Veil (2011)
20 (single
company
target)
Barnett
Rahm and Riha
(2014), Nicot et al.
(2012)
5 (general
estimate -
appears
to cover
recent
years)
Eagle Ford
Nicot and Scanlon
(2012)
0
20
(estimate
based on
interviews)
East Texas
Nicot and Scanlon
(2012)
5
Haynesville
Horner et al.
(2014)
0
Haynesville
Rahm and Riha
(2014)
5 (general
estimate -
appears
to cover
recent
years)
8-37
-------
Chapter 8 - Wastewater Disposal and Reuse
Play or basin
Source and year
2008
2009
2010
2011
2012
2013
2014
West Coast and Upper Plains
Denver-
Julesburg
(Weld County),
CO
Sumi (2015)
54 (flow-
back only)
Bakken
Horner et al.
(2014)
0
a Studies focusing on the Marcellus Shale use waste data reports from PA DEP.
Reuse in the Marcellus region is higher in the northeastern part of Pennsylvania than in the
southwestern portion where easier access to Class IID wells in Ohio makes disposal by injection
more feasible (Rahm etal.. 2013). Outside of the Marcellus region, reuse rates are lower. Ma et al.
(20141 note that only a small amount of reuse is occurring in the Barnett Shale. Reuse has not yet
been pursued aggressively in New Mexico or in the Bakken (North Dakota) (Horner etal.. 2014:
LeBas etal.. 20131. Other sources, however, indicate growing interest in reuse, as evidenced in
specialized conferences (e.g., "Produced Water Reuse Initiative 2014" on produced water reuse in
Rocky Mountain oil and shale gas plays), and available state-developed information on reuse (e.g.,
fact sheet by the Colorado Oil and Gas Conservation Commission) (Colorado Division of Water
Resources etal.. 20141.
If hydraulic fracturing activity slows in an area that is currently reusing wastewater, demand for
the wastewater may decrease and wastewater management practices may shift Analysis by Wunz
f20151 and data in Figure 8-1 suggest a decline in wastewater production in Pennsylvania. Wunz
f20151 also notes that in the future, there could be a trend of more wastewater coming from late-
stage produced water and less from flowback as more wells are in the production phase and fewer
wells are being fractured. If the demand drops relative to production due to fewer wells being
drilled and fractured, then the "excess" produced water will need to be managed by other means.
Alternatives to reuse may include increased transport to disposal wells (e.g., those in Ohio),
development of more disposal wells in Pennsylvania, or advanced treatment and discharge to
surface water via CWTs that have TDS removal capabilities f SAFER PA. 2015: Wunz. 2015: Silva et
al.. 2014al.
8.4.4.1 Potential Impacts from Reuse
For companies employing reuse as a wastewater management strategy, surface spills and leaks can
occur during wastewater transport to and from a treatment facility or from storage tanks/pits
located at the treatment facility or at the well site. Releases may be due to failed infrastructure such
as tank or pipe ruptures, from natural disasters such as floods or earthquakes, or incidents such as
overfills, improper operations, or vandalism (CCST, 2015a; NYSDEC, 2011). If the spill or leak is not
contained or otherwise mitigated, these releases could reach groundwater or surface water
(CCST. 2015a; NYSDEC. 20111. See Chapter 7 for more discussion on types of spills associated with
8-38
-------
Chapter 8- Wastewater Disposal and Reuse
hydraulic fracturing activities, including storage and transport See Section 8.4.5 for discussion of
storage pits and associated impacts on drinking water resources.
With reuse there is the potential for accumulation of dissolved solids such as salts and TENORM in
the wastewater over successive reuse cycles (see Section 7.3.4.6 and Section 8.5.2 for more
information about TENORM). Because wastewater is often reused with minimal treatment,
constituents resulting from time spent in the subsurface remain in the wastewater and can increase
during additional hydraulic fracturing. This potentially concentrated wastewater can pose a bigger
issue if a breach occurs in an on-site pit or tank storing this wastewater while awaiting reuse
(Section 8.4.5; Chapter 7).
The issue of concentrating contaminants during reuse has not yet been quantitatively evaluated in
the literature. Also, it is not known how much this problem would be mitigated due to the dilution
of wastewater when reused as new fracturing fluid. Estimates of the percentages of reused
wastewater in new fracturing fluids in Pennsylvania range from about 2% in 2009 to as much as
22% in 2013 fSRBC. 2016: Schmid and Yoxtheimer. 20151 (Chapter 4). However, data from
Pennsylvania's TENORM study fPADEP. 2015bl showed radium in some hydraulic fracturing fluids,
presumably from a reused wastewater component. As reused wastewater continues to accumulate
contaminants, the water will eventually need to be managed, either through treatment or injection.
8.4.5 Storage and Disposal Pits and Impoundments
The use of pits and impoundments as part of a wastewater management strategy is a historic as
well as current practice in the oil and gas industry. These structures are either used for temporary
storage (on-site at oil and gas production wells or off-site at CWTs or disposal wells) or they are
intended for permanent disposal (evaporation or percolation). There are a variety of terms to
describe these structures depending upon their use (Richardson etal.. 2013): "pits,"
"impoundments," and "reserve pits" are some of the more common terms associated with
wastewater management The terms "impoundment" or "pond" are often used to refer to large area
holding structures and are also used by some states for specific applications such as holding
"freshwater" for fracturing fluid formulation fOuarantaetal.. 20121. Definitions and terminology
are not standardized and vary from state to state (Richardson etal.. 2013). For the purposes of this
section, the nomenclature will defer to the term used by the original author/regulating authority.
States govern the use and permitting of pits under their jurisdiction. Regulations vary from state to
state regarding the circumstances in which pits can be used (e.g., chemical composition of the fluid),
how they should be constructed, and whether they must be lined (e.g., proximity to drinking water
resources and/or chemical composition of the fluid) (Richardson etal., 2013). Most states restrict
the use of wastewater pits in environmentally sensitive areas. To avoid contamination events, some
states are moving toward requiring closed loop systems (i.e., tanks) or injection wells rather than
using pits for hydraulic fracturing wastewater storage. For example, Pennsylvania has modified
their regulations (published October 8, 2016) to ban the use of pits for temporary storage of
unconventional (as defined by PA DEP) wastewaters; many operators have already moved to
closed-loop systems (PA DEP. 2016a). This development is particularly notable because of
8-39
-------
Chapter 8 - Wastewater Disposal and Reuse
Pennsylvania's heavy reliance on reuse for wastewater management, necessitating both on-site and
off-site storage.
8.4.5.1 Locations and Numbers of Pits
The locations and number of existing pits (both for storage and for disposal) are not well
documented in all states, and in the data found, pits associated with hydraulic fracturing operations
were not specifically identified. With respect to larger pits for storage or disposal of wastewater,
some states (e.g., Utah and Oklahoma) provide locational data on their websites. In 2016, the state
of California began posting the number of active and inactive oil field produced water "ponds"
(defined as unlined surface impoundments), both permitted and unpermitted, on their website. The
July 2016 posting showed that 64% (682) of the 1,065 unlined ponds identified in the Central
Valley and Central Coast of California were active. Of the active ponds, 21% (144) were not
permitted (CA Water Board. 2016). Active ponds are primarily found in the southern San Joaquin
Valley (CCST. 2015a). The EPA Region 8 conducted a survey of pits associated with oil and gas
operations in Colorado, Montana, North Dakota, South Dakota, Utah, and Wyoming from 1996
through 2002. Results indicated there were approximately 28,000 pits atthattime (U.S. EPA.
2003b).
In the absence of an inventory of pits in Pennsylvania, the organization SkyTruth led an effort using
volunteers to produce a map of pits believed to be associated with drilling and hydraulic fracturing
the Marcellus Shale fManthos. 20141. The identification of pits was based on USDA aerial imagery
taken in 2005, 2008, 2010, and 2013. SkyTruth acknowledges the uncertainties associated with
identifying pits based on aerial images and volunteer labor. They have described their methodology
as including multiple reviewers and QA/QC procedures. The study cannot differentiate ponds for
drilling fluids and fracturing fluids from those for wastewater. Their preliminary findings indicate
that the estimated number of ponds rose from 11 in 2005 to 529 in 2013, with the structures
themselves increasing in size from a median size of 3,713 ft2 (345 m2) in 2005 to 66,844 ft2 (6,210
m2 in 2013. SkyTruth also notes that impoundments are not permanent and that of 581 ponds
delineated in 2010, only 116 of them were found in the images from 2013.
Evaporation ponds, referred to as Commercial Oil Field Waste Disposal Facilities (COWDFs), are a
waste management strategy most commonly used in the western states such as Utah, Wyoming,
and Colorado (USFWS, 2014). According to a 2016 list of approved COWDFs posted by the Utah
Division of Oil, Gas, and Mining (Utah Division of Oil. 2016), 20 facilities in Utah are approved to
accept produced water. All are in the eastern part of the state where the Uinta and Paradox basins
are found (unconventional shale formations). The Wyoming Department of Environmental Quality
website, accessed in 2016, lists 35 active COWDFs fWDEO. 2016b"). The increase in hydraulic
fracturing activity in Wyoming has resulted in significant increase in wastewater disposed of in
COWDFs (USFWS. 2014). Data from the Colorado Oil & Gas Conservation Commission includes
eight active evaporation pits, five of which are unlined (COGCC, 2016). Ninety-five other active pits
are listed in Colorado, with descriptors such as "production," "multi-well pit," "skim," or "produced
water." Seventy-one of these are unlined, and 22 have synthetic liners. Eleven pits are located in
8-40
-------
Chapter 8- Wastewater Disposal and Reuse
Garfield County, where there is hydraulic fracturing activity. The Colorado data do not distinguish
pits at centralized commercial facilities from on-site pits.
8.4.5.2 Unlined Storage Pits and Percolation Pits
Whether an unlined pit is designed and intended to percolate wastewater into the ground for
disposal or if it is built for storage, it provides a pathway for wastewater to infiltrate into the
subsurface and potentially reach groundwater. Such pits have been used historically for
conventional oil and gas wastewater. More recently, they have received wastewater in areas where
hydraulic fracturing takes place. States such as Montana and Wyoming allow unlined pits to be used
for storage if the quality of the waste fluid meets specified limits and the pit is not in close
proximity to environmentally sensitive areas such as drinking water resources, wetlands, and
floodplains ("Kuwayama etal., 2015b; Richardson et al., 2013").
In the past, several states have allowed unlined pits designed to dispose of wastewater via
percolation into the subsurface. For example, until July 2015, percolation pits were permitted for
wastewaters from hydraulically fractured wells in the Central Valley Region in California ( Grinberg,
2016"). The California Department of Conservation's Division of Oil, Gas, and Geothermal Resources
(DOGGR) listed "evaporation-percolation" as the management method for almost 60% (190 million
gal) of the wastewater generated via well stimulation in Kern County between 2011 and 2014
("CCST, 2015a"). However, according to DOGGR's 2015 report addressing well stimulation activities
in Kern County from January 1, 2014 through September 30, 2015, evaporation/percolation was
not employed as a disposal option during that period (98% of the produced water was disposed of
via operator-owned Class II injection wells, 1.75% was disposed of via commercial Class II injection
wells, and 0.16% was reused).
While the practice of disposal via percolation pits has been discontinued in most states, as of July
2016, Wyoming's regulations still allow the use of percolation for disposing produced water
specific to CBM operations in the Powder River Basin. To be permitted, the operator must
demonstrate that the disposed fluid will comply with water quality standards of the Department of
Environmental Quality (WYOGCC, 2015).
8.4.5.3 Evaporation Ponds
Evaporation is a simple water management strategy involving transporting wastewater to a pond
or pit with a large surface area and allowing passive evaporation of the water from the surface
(NETL, 2014; Clark and Veil, 2009). As discussed above, this disposal option, often referred to as a
COWDF, is practical for drier climates of the western United States. Evaporation ponds have been
used for oil and gas wastewater disposal in Montana, Colorado, Utah, New Mexico, and Wyoming
(Veil etal, 2004). However, New Mexico no longer allows the use of pits for disposal (NM EMNRD
OCD, 2013), and in Montana, evaporation ponds are no longer allowed because they do not put
extracted water to a beneficial use (NRC, 2010). Figure 8-8 shows an example of a lined evaporation
pit in Montana (DOE, 2006).
8-41
-------
Chapter 8 - Wastewater Disposal and Reuse
Figure 8-8. Lined evaporation pit in the Battle Creek Field (Montana).
Source: DOE (2006). Reproduced with permission from ALL Consulting.
As the water component of the wastewater is subject to evaporation, the fluid remaining in the
pond becomes concentrated, and a sludge layer is formed. Remaining residual brines in the pond
can be collected and disposed of via an underground injection well, and the solids can be taken to a
landfill (see Section 8.4.7 for more details]. In cold, dry climates, a freeze-thaw evaporation method
has also been used to purify water from oil and gas wastewater (Bovsen et al.. 19991.
Nowak and Bradish f20101 describe the design, construction, and operation of two large
commercial evaporation facilities in Southern Cross, Wyoming and Danish Flats, Utah. Each facility
includes 14,000 gal (53,000 L) three-stage concrete receiving tanks, a sludge pond, and a series of
five-acre (20,234 m2) evaporation ponds connected by gravity or force-main underground piping.
The Wyoming facility, which opened in 2008, consists of two ponds with a total capacity of
approximately 84 million gal (2 million bbls or 318 million L). The Utah facility, open since 2009,
consists of 13 ponds with a total capacity of approximately 218,4 million gal (5.2 million bbls or
826.6 million L). Each facility receives 0.42 to 1.47 million gal (10,000 to 35,000 bbls; 1.59 million
to 5.56 million L) of wastewater per day from oil and gas production companies in the area.
Evaporation ponds or pits are subject to state regulatory agency approval and must meet state
standards for water quality and quantity (Bovsen et al.. 20021.
8.4.5.4 Impacts and Potential Impacts from Pits and impoundments
Pits containing hydraulic fracturing wastewater have the potential to impact drinking water
resources if spills and overflows cause runoff to surface water or if wastewater percolates through
the soil and reaches groundwater. In addition to contaminants in the wastewater itself, wastewater
8-42
-------
Chapter 8- Wastewater Disposal and Reuse
that reaches groundwater may mobilize constituents in pit bottoms or soils, and it may also reach
hydrologically connected surface water. These impacts are amplified with increasing
pit/impoundment size fOuaranta etal.. 20121. Percolation may be accidental (through tears or
improper installation of liner) or by design in unlined pits (Sumi. 2004).
Compromised pit liners can result in leaks, and extreme weather events, such as floods, can cause
pits to overflow. An analysis of three state databases (New Mexico, Oklahoma, and Colorado) where
pits and tanks have been used for storage of hydraulic fracturing wastewater found that for pits, the
most common causes of spills were from overflows and liner malfunctions (Kuwavama et al..
2015b). For instance, of the 106 pit-related spills reported in New Mexico between 2000 and 2014,
33% were due to overflows and 26% were caused by liner malfunctions. Of the 62 tank spills
reported, 44% were due to leaks, and 27% were related to overfilling fKuwavama etal.. 2015bl.
The types of constituents in pits that may be of concern from such events include VOCs, metals,
TDS, oil, and TENORM (Kuwavama et al.. 2015b).
Operational factors also influence potential impacts from pits and impoundments. These can
include water level management (influent, seepage, spillage), the length of time water is stored in
the pit/impoundment, the composition of the water, the local climate (rainfall and/or evaporation),
and the transmission method (piped or delivered in an open channel) fNRC. 20101.
Construction and Capacity Issues
Construction requirements typically include specifications for features that can reduce the potential
for impacts on groundwater or surface water. These can include liner specifications, depth to
groundwater, secondary containment, setback requirements, freeboard, leak detection, and water
quality monitoring (Kuwavama etal., 2015b).12 For example, in a 2012 review of 19 states with
shale gas development or potential for shale gas development, many states had setback
requirements for pits in sensitive areas including surface water, wetlands, and floodplains. As of
December 2015, however, 12 of the 19 states surveyed did not include setback requirements in
their regulations. Many states did address the vertical separation of pits from the water table (e.g.,
20 in (0.5 m) to seasonal high water table in PA; 10 ft (3 m) in WY; 50 ft (15 m) in NM) fKuwavama
etal., 2015b).
Despite construction standards, impacts on groundwater or surface water due to overflows, liner
breaches, and other construction issues have been documented. In 2007 in Knox County, Kentucky,
retention pits holding hydraulic fracturing flowback fluids overflowed into Acorn Fork Creek during
the development of four natural gas wells (CCST. 2015a: Papoulias and Velasco. 2013). The incident
caused the pH of the creek to drop from 7.5 to 5.6 and the conductivity to increase from 200 to
35,000 |j.S/cm. In addition, organics and metals including iron and aluminum formed precipitates in
the stream. Fish and aquatic invertebrates were killed or distressed in the area of the stream
affected by the release fPapoulias and Velasco. 20131.
1 Setback is the distance between the pit and a stream, lake, building, or other feature or structure that needs protection.
2 Freeboard is the vertical distance between the level of the water in an impoundment and the overflow elevation (an
outfall or the lowest part of the berm].
8-43
-------
Chapter 8 - Wastewater Disposal and Reuse
Similarly, in 2009, Marcellus wastewater stored in an impoundment from a hydraulic fracturing
operation in Washington County, Pennsylvania overflowed the bank of the impoundment and
reached surface water (a tributary of Dunkle Run) fCCST. 2015a). NRC f20101 reported continuous
overfilling of an impoundment in the Powder River Basin (Wyoming) with CBM produced water,
resulting in significant erosion of a seasonal water channel. The CBM operator was required
through litigation to manage flows to the impoundment to prevent overflows. The literature did not
report specific impacts on groundwater or surface water from the Pennsylvania or Wyoming
incidents.
In Pennsylvania in 2010, pit liner failure was reported to have impacted groundwater through
leakage of Marcellus wastewater from six impoundments fColaneri. 20141. Ziemkiewicz et al.
f20141 note that a study of 15 pits and impoundments in West Virginia found that slope stability
and liner deficiencies were common problems. Construction quality control and quality assurance
were often inadequate; the authors found a lack of field compaction testing, use of improper soil
types, excessive slope lengths, buried debris, and insufficient erosion control, although no breaches
were reported. A statistical analysis of oil and gas violations in Pennsylvania found that structurally
unsound impoundments or inadequate freeboard were the second most frequent type of violation,
with 439 instances in the period from 2008 to 2010 (Olawovin etal.. 20131.
Unlined Pits
Impacts on groundwater from historic and current uses of unlined pits in the oil and gas industry
have been documented. In a review of records spanning 25 years (1983 - 2007), 63 incidents of
private water supply contamination from the infiltration of saline fluids from unlined or
inadequately constructed reserve pits were identified in Ohio fKell. 20111. The same study fKell.
20111 identified 57 legacy (pre-1984) incidents in Texas involving groundwater contamination
from unlined produced water disposal pits. Such pits were phased out in Texas by 1984, prompting
a move towards disposal of oil and gas wastewater in disposal wells.
Kern County, California has experienced impacts on groundwater associated with unlined
percolation pits. A 2014 study notes that there are hundreds of pits across Kern County and
elsewhere in the state, stretching state resources for regulatory oversight (Grinberg. 20141. Past
sampling of water in percolation pits has shown exceedances of California's Tulare Lake Basin Plan
(Basin Plan), which specifies maximum levels permitted for discharges of oil field well wastewater
to unlined ponds overlying groundwater fGrinberg. 20141.1 For example, the McKittrick 1 and 1-3
pits are large percolation pits in Kern County near oil fields where most of the hydraulic fracturing
in California takes place (Grinberg. 20141. The pits are situated close to a number of important
resources. They are located within a few miles of the Kern River Flood Channel, the California State
Water Project, farmland, and are in an area of high quality groundwater (Grinberg. 20141. Sampling
of fluids in the pits dating back to 1997 showed consistent exceedances of Tulare Basin Plan
standards for TDS, chlorides, and boron. Sampling also revealed the presence of BTEX, gasoline
range organics (GRO), and diesel range organics (DRO) (MTA. 20141. Sampling of three monitoring
1 The Basin Plan sets limits for salinity (1,000 nmhos/cm measured as electrical conductivity], chloride (175 mg/L], and
boron (1 mg/L] fCalifornia Regional Water Quality Control Board Central Valley Region. 20151.
8-44
-------
Chapter 8- Wastewater Disposal and Reuse
wells indicated that in 2004, a plume had migrated at least 4,000 ft (1,000 m) from the pits and was
still detected in test wells in 2013. As of July 1, 2015, California's Code of Regulations includes a
provision that no longer allows the use of pits, including percolation pits, for fluids produced from
stimulated wells (Grinberg. 2016).
Unlined pits that were used from the 1960s until the mid-1990s for disposal of drilling muds and
flowback and produced waters associated with hydraulic fracturing operations have been linked to
groundwater contamination in Pavillion, Wyoming ("Digiulio and Tackson, 2016; AME, 2015). A
report by the Wyoming Oil and Gas Conservation Commission (WYOGCC) fWYOGCC, 2014a)
summarizes site investigations and reclamation activities conducted by WOGCC, the Wyoming
Department of Environmental Quality (WDEQ), and Encana Oil and Gas for pits in the Pavillion Well
Field. The report includes information on samples collected between 2006 and 2013 from shallow
groundwater in the vicinity of the pits. Some sites had detections for one or more of the following
contaminants: GRO, DRO, BTEX, and/or naphthalene. Of the shallow groundwater sites with
detections, some were associated with pits located within one-quarter mile of a domestic well. One
of these sites exceeded clean-up levels established by the WDEQ Voluntary Remediation Program
for DRO (13,000 |ig/L) and benzene (110 [ig/L).1The report noted that there was insufficient
evidence to determine whether or not drinking water supply wells in the vicinity of the pits were
contaminated by disposal of hydraulic fracturing wastewater in those pits fWYOGCC. 2014a").
Other examples in the literature include the detection of VOCs in groundwater downgradient of an
unlined pit containing oil and gas wastewater near the Duncan Oil Field in New Mexico fSumi.
2004) (Section 8.5). Groundwater impacts downgradient of an unlined pit in Oklahoma included
high salinity (3500-25,600 mg/L) and the presence of VOCs (Kharaka etal.. 2002). Neither New
Mexico nor Oklahoma currently allows unlined pits for disposal or storage (OCC OGCD. 2015: NM
EMNRD OCD. 20131.
Mobilization and Transport of Constituents
Groundwater impacts may result not just from constituents in the wastewater but also from
mobilization of existing constituents in the soil or sediment A CBM produced water impoundment
in the Powder River Basin of Wyoming was studied for its impact on groundwater (Healv etal..
2011: Healv etal.. 20081. Infiltration of water from the impoundment was found to create a perched
water mound in the unsaturated zone above bedrock in a location with historically little recharge.
Elevated concentrations of TDS, chloride, nitrate, and selenium were found at the site, with one
lysimeter sample exceeding 100,000 mg/L of TDS (Healv etal.. 2008). Most of the solutes found in
the groundwater mound did not originate with the CBM produced water, but rather were the
consequence of dissolution of previously existing salts and minerals fHealv etal.. 20111.
Generally, the deeper that wastewater can move into an aquifer, as impacted by the volume and
timing of the release, the longer the duration of contamination (Whittemore. 2007). Kharaka et al.
(2007) reported on studies at a site in Oklahoma with one abandoned and two active unlined pits.
1 WDEQ cleanup levels are derived from a combination of promulgated levels (MCL, state-assigned water quality
standards] and risk-based cleanup level concentrations ("WDEO. 2016a").
8-45
-------
Chapter 8 - Wastewater Disposal and Reuse
Produced water from these pits penetrated 10 to 23 ft (3 to 7 m) thick shale and siltstone units,
creating three plumes of high-salinity water (5,000 to 30,000 mg/L TDS). The impact of these
plumes on the receiving water body (Skiatook Lake) was judged to be minimal, although the
estimate was based on a number of notably uncertain transport quantities (Otton etal.. 2007).
Vadose (unsaturated) zone transport was illustrated at a site in Oklahoma where two abandoned
pits were major sources for releases of produced water and oil. Saline water from the pits flowed
through thin soils and readily percolated into underlying permeable bedrock. Deeper, less-
permeable bedrock was contaminated by saltwater later in the history of the site, presumably due
to fractures. The mechanisms proposed were vertical movement through permeable sand bodies,
lateral movement along shale fractures, and possibly increased clay permeability due to the
presence of highly saline water (Otton etal.. 2007).
Summary
Collectively, the above examples show that regardless of the purpose of pits (storage or disposal),
they present a potential pathway for wastewater constituents to impact groundwater or surface
water. Good construction standards and practices, including liners, adequate freeboard, and
setbacks, are important for minimizing potential impacts on both surface water and groundwater.
Proper monitoring and maintenance (e.g., avoiding overfilling, maintaining the integrity of liners
and berms) are also important for protecting surface water and groundwater. Unlined pits, in
particular, can lead to groundwater contamination. This can be long-lasting, as evidenced by legacy
impacts from older pits. Most states have phased out unlined disposal pits and unlined storage pits,
but if such pits are still in use, they can provide ongoing potential sources of groundwater
contamination fCCST. 2015a: Grinberg. 20141.
8.4.6 Other Management Practices and Issues
Additional strategies for wastewater management in some states include directly discharging to
surface waters and land application. In particular, wastewater from CBM fracturing and production
generally has lower TDS concentrations than wastewater from other types of unconventional
formations and more readily lends itself to other uses.
8.4.6.1 Land Application and Road Spreading
Road spreading has been used as a disposal option for high-TDS wastewaters (brines) from
conventional oil and gas production. Road spreading can be done for dust control and de-icing.
Although recent data are not available, an American Petroleum Institute (API) survey estimated
that approximately 75.6 million gal (1.8 million bbls or 286.2 million L) of wastewater was used for
road spreading in 1995 fAPI. 20001. The API estimate does not specifically identify hydraulic
fracturing wastewater. There is no current nationwide estimate of the extent of road spreading
using hydraulic fracturing wastewater.
Road spreading with hydraulic fracturing wastewater is regulated primarily at the state level
(Hammer and VanBriesen. 2012) and is prohibited in some states. For example, with annual
approval of a plan to minimize the potential for pollution, PA DEP allows spreading of brines from
8-46
-------
Chapter 8- Wastewater Disposal and Reuse
conventional (as defined by PA DEP) wells for dust control and road stabilization. Hydraulic
fracturing flowback, however, cannot be used for dust control and road stabilization fPADEP.
2011b). In West Virginia, use of gas well brines for roadway de-icing is allowed per a 2011
memorandum of agreement between the West Virginia Division of Highways and the West Virginia
Department of Environmental Protection, but the use of "hydraulic fracturing return fluids" is not
permitted (Tiemann et al.. 2014: West Virginia DEP. 20111.
Concerns about road application center on contaminants such as barium, strontium, and radium. A
report from PA DEP analyzed several commercial rock salt samples and compared results with
contaminants found in Marcellus Shale flowback samples. The results noted elevated barium,
strontium, and radionuclide levels in Marcellus Shale brines compared with commercial rock salt
fTitler and Curry. 20111. Another study found increases in metals (radium, strontium, calcium, and
sodium) in soils ranging from 1.2 to 6.2 times the original concentrations (for radium and sodium,
respectively), attributed to road spreading of wastewater from conventional oil and gas wells for
de-icing fSkalak etal.. 20141.
Potential impacts on drinking water resources from road spreading have been noted by Tiemann et
al. (2014) and Hammer and VanBriesen (2012). These include potential effects of runoff on surface
water and migration of brines to groundwater. Snowmelt can carry salts and other chemicals from
the application site, and transport can increase if application rates are high or rain occurs soon after
application fHammer and VanBriesen. 20121. Research on the impacts of conventional road salt
application has documented long-term salinization of both surface water and groundwater in the
northern United States (Kelly. 2008: Kaushal etal.. 2005). When conventional oil field brine was
used in a controlled road spreading experiment, elevated chloride concentrations were detected in
shallow groundwater (Bair and Digel. 1990). The amount of salt attributable to road application of
hydraulic fracturing wastewaters has not been quantified.
To evaluate land application of solid wastes from oil and gas production, a laboratory study
mimicking land spreading of conventional oilfield scales and sludges indicated that 20% of the
radium in barite sulfate scales was released by microbial processes during incubation with soil
fMatthews etal.. 2006: Swann etal.. 20041. Although the radium was then complexed with the soil,
it would be more mobile and more bioavailable than when it was associated with the barite.
Overall, potential effects on drinking water resources from land spreading are not well understood,
including the amounts of hydraulic fracturing wastes that are managed by land spreading.
8.4.6.2 Management of Coalbed Methane Wastewater
Many, but not all, CBM wells are hydraulically fractured to enhance recovery, using fluids that range
from water alone to more complex gel formulations with proppant (e.g., Engle etal.. 2011:
McCartney. 2011: NRC. 2010: Halliburton. 2008: U.S. EPA. 2004al. The literature indicates that
hydraulic fracturing of CBM formations is being conducted in the San Juan, Raton, Piceance, and
Uinta Basins, among others. Literature such as NRC (2010) notes that hydraulic fracturing may not
be common in the Powder River Basin. Additionally, when CBM well stimulation does take place, it
can be accomplished using very simple hydraulic fracturing fluid formulations (Chapter 3).
8-47
-------
Chapter 8 - Wastewater Disposal and Reuse
Wastewater from CBM wells can be managed like other hydraulic fracturing wastewater discussed
above. However, the wastewater from CBM wells can also be of higher average quality (typically
lower TDS content) than wastewater from other hydraulically fractured wells. The lower TDS
content makes it more suitable for certain management practices and uses. A number of
management strategies have been proposed or implemented, with varying degrees of treatment
required depending on the quality of the wastewater and the intended use (Hulme. 2005: DOE.
2003. 20021. Although specific volumes managed through the practices discussed below are not
well documented, qualitative information and considerations for feasibility are available and
presented. The discussion below covers both dilute and higher-TDS wastewater from CBM
formations.
The quality of CBM wastewater plays a large role in how the wastewater is managed. The TDS
content can range from an average of nearly 1,000 mg/L in the Powder River Basin to an average of
about 14,000 mg/L (and as high as approximately 62,000 mg/L) in the Black Warrior Basin
(Appendix Table E-3). Data sources from about 2002 through 2008 indicate that operators in some
basins such as the San Juan, Uinta, and Piceance, and Raton (in New Mexico), where TDS is typically
higher compared to other basins (e.g., Powder River), manage most wastewater by injection into
disposal wells fNRC. 2010: U.S. EPA. 201 Oal.
Discharge to rivers and streams, a management option governed by the CWA, may be permitted in
cases where wastewater is of high quality.1 To be discharged, the wastewater must meet
technology-based effluent limitations established by the permitting authority on a case-by-case
"best professional judgment" basis as well as any more stringent limitations necessary to meet
applicable water quality standards. For example, as a means of protecting high-quality waters of the
state, the Montana Supreme Court ruled in 2010 that treatment is required for all CBM produced
water prior to discharge to surface water (NRC. 2010).
A 2008 EPA survey of CBM operators found that of the projects represented in the results, direct
discharge to surface water was by far most prevalent in the Powder River Basin but was also
reported as a management practice in the Green River, Raton, Black Warrior, Cahaba, Illinois, and
Appalachian basins fU.S. EPA. 2013e. 2010a).2 Discharges to surface water can provide habitat
maintenance, restoration of wildlife-waterfowl fishery habitat, and flow augmentation to benefit
downstream water users. However, hydrologic changes from such discharges could also have
unanticipated effects on ecosystems previously adapted to intermittent streamflow.
Some CBM wastewater can be put to agricultural use, including livestock and wildlife watering, and
crop irrigation. Livestock watering with CBM wastewater can be done using on-channel or off-
channel impoundments, and irrigation is an area of active research fe.g.. Engle etal.. 2011: NRC.
2010). However, wastewater from some higher-salinity CBM basins (e.g., San Juan, Uinta, and
Piceance) would need blending or treatment before such uses. Irrigation with treated CBM
1 Although discharge to rivers and streams is generally prohibited under the EPA's oil and gas ELGs, the ELGs do not apply
to CBM.
2 These reports did not describe certain non-discharging wastewaters management strategies in basins with few
operators in order to preserve CBI. The reports also do not provide information on hydraulic fracturing activities in the
basins. Not also that results are presented by numbers of projects, which may vary in the number of wells they contain.
8-48
-------
Chapter 8- Wastewater Disposal and Reuse
wastewater would be most suitable on coarse-textured soils for cultivation of salt-tolerant crops
fDOE. 20031. NRC f20101 remarks that "use of CBM produced water for irrigation appears practical
and sustainable," provided that appropriate measures are taken such as selective application,
dilution or blending, appropriate timing, and rehabilitation of soils.
Although CBM wastewater is generally lower in TDS than wastewater associated with shale gas
development, it can still have higher TDS concentrations than stream water. This poses concerns
regarding the sodium adsorption ratio (SAR) for agricultural soils. A USGS study performed trend
analysis of water quality at sampling sites in the Tongue and Powder River watersheds (Powder
River Basin) (Sando etal.. 20141. One of the study objectives was to determine possible effects of
CBM produced water particularly in areas where the water was discharged to impoundments or
upper reaches of in-stream channels for infiltration. Trend analysis showed potential effects of CBM
production on downstream water quality (increases in sodium, alkalinity, and SAR) in the main-
stem Powder River but found mixed results at the Tongue River sites (some appeared to be
impacted by CBM activities while others did not) f Sando etal.. 20141.
Sando etal. f20141 found that CBM pumping rates (i.e., discharge of produced water) were high
relative to streamflow in the Powder River Basin. For the three main-stem Powder River sites, the
CBM pumping rates were 26-34% of the 2001-2010 median streamflows. For one site in the Little
Powder River watershed, the CBM pumping rate was 360% of 2001-2010 median streamflow. This
underscores that in arid climates in the western United States, permitted discharges from CBM
activities (whether hydraulically fractured or not) at a particular site may be large relative to the
size of the receiving water and may sometimes dominate flows.
As noted above, a degree of treatment is needed (or required) for some uses. Plumlee etal. f20141
examined the feasibility, treatment requirements, and potential costs of several hypothetical uses
for CBM wastewater. In several cases, costs for these uses were projected to be comparable to or
less than estimated disposal costs. In one case study, use of CBM wastewater for streamflow
augmentation or crop irrigation could potentially cost between $0.26 and $0.27 per bbl. For
comparison, reported disposal costs in 2000-2001 ranged from $0.01 per bbl for a pipeline
collection system with impoundment to $2.00 per bbl for hauling to disposal or treatment. The
2010 NRC report (NRC. 2010) noted that 15 to 18% of CBM produced water in the Powder River
Basin was being treated to reduce SAR in order to satisfy NPDES permits for discharge.1 If
wastewater is treated to address SAR, reported costs are approximately $0.12 to $0.60/bbl fNRC.
20101.
The applicability of particular uses may be limited by ecological and regulatory considerations as
well as the irregular nature of CBM wastewater production (voluminous at first, and then declining
and halting after a period of years). Legal issues, including overlapping jurisdictions at the state
level and senior water rights claims in over-appropriated basins (in western states) can also
determine the use of CBM wastewater (Wolfe and Graham. 20021.
1 SAR is the relative proportion of sodium to other cations in water. It is also an indication of risk to soil from alkalinity.
The higher the SAR, the less suitable the water is for irrigation, and long-term use can damage soil structure.
8-49
-------
Chapter 8 - Wastewater Disposal and Reuse
8.4.6.3 Other Documented Uses of Hydraulic Fracturing Wastewater
Uses of wastewater from shales or other hydraulically fractured formations face many of the same
possibilities and limitations as those associated with wastewater from CBM operations. The biggest
difference is in the quality of the water. Wastewaters vary widely in water quality, with TDS values
from shale and tight sand formations ranging from less than 1,000 mg/L TDS to hundreds of
thousands of mg/L TDS fDOE. 20061 (Chapter 7). Wastewaters on the lower end of the TDS
spectrum could be reused in many of the same ways as CBM wastewater, depending on the
concentrations of potentially harmful constituents and applicable federal, state, and local
regulations. High TDS wastewaters have more limited uses, and pre-treatment may be necessary
(Shaffer etal.. 2013: Guerra etal.. 2011: DOE. 20061. Agricultural and wildlife uses are subject to the
produced water daily effluent discharge limit of 35 mg/1 for oil and grease.1
Potential uses for wastewater in the western United States include livestock watering, irrigation,
streamflow supplementation, fire protection, road spreading, and industrial uses, with each having
their own water quality requirements and applicability (Guerra et al.. 2 011). Guerra etal. (2011)
summarized the least conservative TDS standards for five possible uses in the western United
States that include 500 mg/L for drinking water (the drinking water secondary maximum
contaminant level (SMCL)), 625 mg/L for groundwater recharge, 1,000 mg/L for surface water
discharge, 1,920 mg/L for irrigation, and 10,000 mg/L for livestock watering. The authors
estimated that wastewater from 88% of unconventional wells in the western United States could be
used for livestock watering without TDS removal based on a maximum TDS concentration of 10,000
mg/L. However, wastewater from only 10% of unconventional wells could be used for surface
discharge without treatment for TDS based on the least conservative standard among the western
states of 1,000 mg/L TDS (Guerra etal.. 2011). Guerra etal. (2011) indicate that in several basins in
the western United States (e.g., Wind River, Green River, and Powder River), wastewater from 50%
or more of oil and gas wells is suitable for agricultural use. In other basins (e.g., San Juan, Piceance,
and Permian) over 50% of oil and gas wastewater is unsuitable for use without treatment A 2006
Department of Energy (DOE) study pointed out that the quality necessary for use in agriculture
depends on the plant or animal species involved and that in the Bighorn Basin in Wyoming, low-
salinity wastewater is used for agriculture and livestock watering after minimal treatment to
remove oil and grease (DOE. 2006).
Although TDS is a common criterion for water quality, there are also recommended limits or
considerations for some metals, alkalinity, and nitrate in water for use in livestock watering, and for
metals, SAR, electrical conductivity (ECw), and pH for water for irrigation f Guerra etal.. 20111. Also,
using TDS/salinity as the primary criterion may not be appropriate if wells contributing to the
produced water have undergone hydraulic fracturing or if maintenance chemicals are being used
on the well.
The water quality standards and monitoring requirements for direct discharge for use in irrigation
or livestock watering include few specifications. In California, the California Council on Science and
Technology (CCST. 2015a) notes that the testing and treatment required by the regional water
140 CFR 435.52(b).
8-50
-------
Chapter 8- Wastewater Disposal and Reuse
quality control boards prior to use of produced water for irrigation do not include assessment for
chemicals associated with hydraulic fracturing and that there are no policies prohibiting the use of
hydraulic fracturing wastewaters for irrigation.
In the Wind River Basin in Wyoming, three NPDES permits were appealed by environmental groups
due to concerns that the permits failed to address maintenance and hydraulic fracturing chemicals
(Natural Resources Defense Council. 2015: PEER. 20151. The environmental groups argued that the
EPA's regulations do not allow for the discharge of produced water containing chemicals from well
treatment, and that, moreover, the EPA lacked sufficient information regarding the well treatment
chemicals to determine whether the discharge would be "good enough quality" for wildlife and
agricultural use, as required under the ELG regulations. As an example, the environmental groups
pointed to MSDS information provided upon request for six maintenance products, which included
toxic chemicals such as ethylene glycol, benzyl chloride, isopropanol, naphthalene, benzene, and
xylene, among others. This raised concerns that produced water permitted for direct discharge may
contain toxic chemicals or their degradation products. Ultimately, pursuant to a settlement
agreement with the environmental groups and permittees, the EPA issued modified permits that
included additional conditions for handling of and reporting about well stimulation and well
maintenance chemicals.
8.4.7 Management of Solid and Liquid Residuals
Solid and liquid residuals associated with hydraulic fracturing wastewater are formed from
treatment processes at CWTs, buildup of sludges in tanks and pits, and scale formation on pipes and
equipment These residuals must be managed and disposed of properly to avoid impacts on ground
and surface water resources. (Note that drill cuttings and drilling muds are outside the scope of this
chapter.)
8.4.7.1 Solid Residuals
The solid residuals produced at a CWT depend on the constituents in the untreated water and the
treatment processes used and are likely to contain TSS, TDS, metals, radionuclides, and organics.
Solid residuals can consist of sludges (from precipitation, filtration, settling units, and biological
processes), spent media (filter media, adsorption media, or ion exchange media), and other
material such as spent filter socks used to remove gross particulates. In addition, solids that
accumulate in storage tanks and pits and scale that deposits on equipment are part of the residual
load from a site. These residuals can constitute a considerable fraction of solid waste in an oil or gas
production area.
Handling and disposal of residual sludges from treatment processes can present some of the biggest
challenges associated with these technologies (Igunnu and Chen. 2014). Additional treatment may
be applied to solid residuals including thickening, stabilization (e.g., anaerobic digestion), and
dewatering processes prior to disposal. The solid residuals are then typically sent to a landfill, land
spread on-site, or incinerated (Morillon et al.. 2002). Land spreading is a waste management
method in which wastes are spread over the soil surface and tilled into the soil to allow the
hydrocarbons in the wastes to biodegrade (Smith etal.. 1998): note that inorganic constituents
8-51
-------
Chapter 8 - Wastewater Disposal and Reuse
(e.g., salts, metals, metalloids, and radionuclides) will not degrade. In addition, pits or
impoundments that have reached the end of their useful life have accumulated residuals. Practices
used to decommission these pits include draining and leveling the pit in place or land farming the
residual materials into the ground fRich and Crosby. 20131. although more information is needed
on the potential for these practices to affect water resources.
A particular concern for the management of residual wastes is TENORM that originates from the
geologic formation and was present in the produced water fSAFERPA. 20151. Studies have found
TENORM in solid residuals at oil and gas operations including the filter cake (PA DEP. 2015b). filter
socks (Harto etal.. 2014). and pit sludges (Rich and Crosby. 2013). Researchers have assessed
Marcellus produced water samples, finding that many with low barium and high radium-226 levels
would generate sludges that exceed the maximum acceptable radium-226 activity for
nonhazardous landfill disposal in Pennsylvania fSilva etal.. 2014b: Silva etal.. 2014al. In scales that
build up on hydraulic fracturing and treatment equipment and sludges that accumulate in tanks and
pits, radium can coprecipitate with barium, strontium, or calcium sulfates (Smith etal.. 1999). (See
Section 8.5.2 for additional discussion of TENORM associated with residuals.)
The accumulation of TENORM in the solids generated can limit or preclude landfills as a disposal
option. Walter etal. (2012) point out that wastes containing TENORM can be problematic due to
the possibility of radon emissions from the landfill. Regulatory limits on permissible radionuclide
levels accepted at non-hazardous landfills vary by state fSilva etal.. 2014a!.1 Some states have
volumetric limitations on TENORM in their landfill permits (e.g., Colorado). Also, some states write
criteria, such as gamma exposure rates (radiation) and radioactivity concentration limits, into
permits for many landfills that are permitted to accept TENORM. Silva etal. f2014al note that there
are 50 nonhazardous (RCRA-D) disposal facilities in Pennsylvania, but no TENORM disposal
facilities. Texas and other states have disposal facilities for TENORM.
8.4.7.2 Liquid Residuals
Liquid residuals include concentrated brines (from membrane or evaporation processes) and
regeneration or cleaning chemicals (from ion exchange, adsorption, and membrane processes)
(Fakhru'l-Razi etal.. 2009). Practices for managing liquid residual streams from treatment
processes are generally the same as for untreated hydraulic fracturing wastewaters, although the
treated volumes are smaller, resulting in lower costs (Hammer and VanBriesen. 2012).
Concentrations of contaminants in liquid residuals, however, will be higher. The most common
disposal method is injection into disposal wells.
If the liquid is not injected into a disposal well, treatment to remove salts would be required for
surface water discharge to meet NPDES permit requirements and protect the water quality for
downstream users such as drinking water utilities (Section 8.5). Because some constituents of
concentrated liquid residual waste streams can pass through or impact municipal wastewater
treatment processes (Linaric etal.. 2013: Hammer and VanBriesen. 2012). these residuals would
1 Examples of permissible radionuclide levels at non-hazardous landfills: Pennsylvania requires alarms to be set at all
municipal landfills, with a trigger set at 10 |iR/hr above background radiation. Texas sets a radioactivity limit, requiring
that any waste disposed by burial contains less than 30 pCi/g radium or 150 pCi/g of other radionuclides.
8-52
-------
Chapter 8- Wastewater Disposal and Reuse
not be appropriate for discharge to a POTW. Elevated salt concentrations, in particular, can have
detrimental effects on microbiological treatment at municipal wastewater systems, such as
activated sludge treatment (Linaric etal.. 2013).
Liquid residuals can also be mixed with a solidifying agent such as Portland cement and then
disposed of in landfills, or they can undergo advanced treatment processes to generate products
such as road salt or industrial chemicals fSAFER PA. 20151.
8.4.7.3 Potential Impacts from Solid and Liquid Residuals
Residual wastes have the potential to impact the quality of drinking water resources if
contaminants leach to groundwater or reach surface water. In a recent study by PA DEP, radium
was detected in leachate from 34 of 51 landfills that accept waste from the oil and gas industry
(Marcellus in particular). Radium-226 concentrations ranged from 54 to 416 pCi/L, and radium-
228 ranged from 2.5 to 1,100 pCi/L (PADEP. 2015b). (See also Section 8.5.2 and see Chapter 9 for
health effects associated with radium). Countess etal. f20141 studied the potential for a wide array
of elements to leach from sludges generated at a CWT handling hydraulic fracturing wastewater in
Pennsylvania. Tests used strong acid solutions (to simulate the worst case scenario) and weak acid
digestions (to simulate environmental conditions). The data illustrate the possibility of leaching of
these constituents from landfills. The extent of leaching varied by constituent and by fluid type, but
the authors concluded that boron, bromide, calcium, magnesium, manganese, silicon, sodium, and
strontium had high potential to migrate from the residual solids, with bromide and sodium having
the highest leaching potential (Countess etal.. 2014). (See also Section 5.8 in Chapter 5 for
discussion of the processes governing the movement of constituents in the subsurface.)
In another study assessing the leaching behavior of residuals from hydraulic fracturing operations,
Sharma etal. (2015) found that alkali metals, alkaline earth metals, and bromide had the highest
leaching potential of the constituents tested. The authors also found that disposing of hydraulic
fracturing residuals along with other solids (e.g., at a municipal landfill) produces a greater leaching
potential than if the residuals are disposed of by burying or land disposal designed for solely the
hydraulic fracturing residuals. This is due to the more acidic leachate formed at the co-disposal
locations (Sharmaetal.. 2015).
Sang etal. (2014) studied the potential for hydraulic fracturing fluid to mobilize colloidal particles
in the soil. The study used microspheres and sand particles as surrogates for contaminant particles.
The authors note that the chemistry of hydraulic fracturing fluid favors transport of colloids and
mineral particles through rock cracks, and they found that infiltration of flowback fluid can
transport existing pollutants such as heavy metals, radionuclides, and pathogens, in unsaturated
soils (Sang etal.. 2014). Heavy metals can also move through soil. Although not specific to hydraulic
fracturing wastes, Camobreco etal. f!9961 report high levels of heavy metal transport in soil
columns, with 12% recovery for lead, 15% for copper, 23% for zinc, and 30% for cadmium
fCamobreco etal.. 19961.
Residuals, whether liquid or solid, are the most concentrated wastes and waste streams associated
with hydraulic fracturing operations. Contaminants in the produced water will accumulate in the
8-53
-------
Chapter 8 - Wastewater Disposal and Reuse
sludges in storage tanks/pits, in scale on the equipment, and in treatment facilities. Proper
management and disposal of these highly concentrated wastes is critical to minimize the potential
for impacts on water resources.
8.5 Potential Impacts of Hydraulic Fracturing Wastewater Constituents on
Drinking Water Resources
The previous section discussed the potential impacts of specific wastewater management strategies
on drinking water resources. The severity of impacts, however, depends largely on the constituents
in the wastewater, the concentrations of those constituents, and their health and ecological effects.
This section will discuss the potential impacts of several specific types of hydraulic fracturing
wastewater constituents on drinking water resources.
The impacts or potential impacts discussed in the literature are heavily focused on discharges from
CWTs, including treated wastewater that is discharged indirectly through POTWs. Available
evidence suggests that the effects of hydraulic fracturing on surface water quality are related to
discharges of partially treated wastewater (Kuwavama etal.. 2015a). Other avenues of
contamination for both surface water and groundwater include leaks from pits and impoundments,
landfill leachate, and leaching from contaminated sediments and other improperly managed solid
wastes.
As noted, an important consideration regarding the potential impacts of hydraulic fracturing
wastewater on receiving water is whether there are constituents of concern known to have health
effects or that can give rise to compounds with health effects. See Chapter 9 for discussion of the
health effects of wastewater constituents. For some classes of constituents, such as DBP precursors,
considerable research exists regarding concentrations in the waste stream and impacts on
downstream drinking water treatment plants or the finished drinking water after treatment. For
other constituents, information is limited, especially within the context of hydraulic fracturing
activities. There may also be unknown constituents because some ingredients in the original
hydraulic fracturing fluids are claimed to be CBI. The following subsections identify several classes
of constituents known to occur in hydraulic fracturing wastewater, discuss whether potential
impacts are likely, and detail information gaps.
8.5.1 Bromide, Iodide, and Chloride
Halides, including bromide, chloride, and iodide, are commonly found in high-TDS hydraulic
fracturing wastewater. As noted in Section 8.3.1.1, chloride is a regulated contaminant with a
secondary MCL of 250 mg/L. Bromide and iodide are not regulated, but are of concern due to their
role in the formation of DBPs fParker etal.. 2014: Krasner. 20091. (See Appendix F for information
on DBP formation.) High-TDS wastewaters from the Marcellus Shale have been the focus of concern
due to the state's history of treating these wastewaters at POTWs (without pretreatment) and at
CWTs without TDS removal capabilities (Text Box 8-1). Discontinuing the practice of sending shale
gas wastewater to POTWs without pretreatment fStates etal.. 20131. and compliance with the new
EPA pretreatment standards for discharges of unconventional oil and gas wastewaters helps
mitigate this problem. This section describes the role of some constituents in high-TDS fluids in the
8-54
-------
Chapter 8- Wastewater Disposal and Reuse
formation of DBPs and provides more details on the effects on surface waters as observed in
Pennsylvania. The lessons learned and steps taken in the Marcellus region can provide valuable
knowledge for operators and state agencies in other parts of the country that treat and discharge
high-bromide and high-iodide wastewaters.
8.5.1.1 Influence of Bromide and Iodide on Formation of Disinfection Byproducts
Disinfection byproducts (DBPs) are formed when organic material comes in contact with
disinfectants (e.g., chlorine, chloramine, chlorine dioxide, or ozone). Of particular concern are DBPs
formed in the presence of halides (e.g., bromide or iodide). The type of DBP formed depends on the
organic precursors in the source water and the disinfectant used. Regulated DBPs include total
trihalomethanes (TTHM), five haloacetic acids (HAA5), bromate, and chlorite. There are, however,
many additional DBPs that are not regulated and may in fact be of greater concern than the
regulated species. Brominated forms of DBPs, for example, are considered to be more toxic and
carcinogenic than chlorinated species (McGuire etal.. 2014: Parker etal.. 2014: States etal.. 2013:
Krasner. 2009: Richardson et al.. 2007). Another halide, iodide is also found in some hydraulic
fracturing wastewater (Chapter 7), and although its effects have not been as well documented as
those associated with bromide, iodide raises some of the same concerns regarding formation of
toxic DBPs as bromide (Xu etal.. 2008).
Studies have found that elevated bromide levels in water correlate with increased DBP formation in
the drinking water that is delivered to customers (also called "finished drinking water") fObolenskv
and Singer. 2008: Matamoros etal.. 2007: Hua etal.. 2006: Yang and Shang. 2004). Harkness et al.
(2015) studied the chemical composition of flowback, produced waters, treated wastewaters,
instream flows downstream from discharges, and accidental spill sites. The study found high
concentrations of bromide and iodide in the flowback and produced waters, concluding that the
elevated levels of these constituents could promote the formation of toxic brominated and
iodinated DBPs in downstream drinking water systems (Harkness etal.. 2015).
In terms of the resulting DBP formation, laboratory experiments using hydraulic fracturing
wastewater from the Marcellus and Fayetteville shales and river water from the Allegheny and Ohio
rivers suggest that a relatively small portion of hydraulic fracturing wastewater can notably affect
DBP formation (Parker etal.. 2014). In particular, trihalomethanes (THM; a category of DBPs) were
shown to shift towards greater brominated and iodinated species with a little as 0.01% hydraulic
fracturing wastewater in disinfected samples. Modeling work by Landis etal. f 20161 evaluated the
impact of CWT discharges on DBP formation at a drinking water system and suggested that
although only a 3% increase in overall TTHM formation was predicted, the model predicted a
decrease in chlorinated THM and a substantial shift toward a higher percentage of the more-toxic
brominatedTHMs fLandis etal.. 20161.
States etal. (2013) found a strong correlation between bromide concentrations in source water
from the Allegheny River in Pennsylvania and the percentage of brominated THMs in finished water
at a drinking water facility using Allegheny source water. Bromide concentrations in the river water
measured during the study ranged from less than 25 |ig/L to 299 |ig/L. The authors noted that
source water containing 50 |ig/L of bromide resulted in treated drinking water with approximately
8-55
-------
Chapter 8 - Wastewater Disposal and Reuse
62% of total THMs consisting of brominated species. When the source water contained 150 |ig/L
bromide, the brominated THM percentage was 83% fStates etal.. 20131.
Pope etal. (2007) reported that increased bromide levels are the second best indicator of DBP
formation, with pH being the first. Furthermore, bromine (which may be formed from bromide in
the water during disinfection) reacts as much as ten times faster and more efficiently with DBP
precursors than chlorine (Westerhoff et al.. 2004). These studies show that increased bromide
concentration in a drinking water resource shifts the DBP formation towards more-toxic
brominated forms.
If disinfection is accomplished using ozonation instead of or in addition to chloramination or
chlorination, bromide and iodide in the source water can form two additional constituents: bromate
and iodate. Iodate, although formed during disinfection by ozonation, is not considered a DBP and is
non-toxic (Allard etal.. 2013). Bromate, however, is a DBP of concern and has an MCL of 0.010
mg/L CIJ.S. EPA. 19981.
Another category of DBP that is not regulated is the nitrogenous DBPs, including nitrosamines. Data
are lacking on the formation of nitrogenous DBPs specifically linked to hydraulic fracturing
wastewater, but their formation is possible. During chloramination, bromide can enhance the
formation of the nitrosamine N-nitrosodimethylamine (NDMA) in waters containing the precursor
dimethylamine (DMA) fLe Roux etal.. 2012: Luh and Marinas. 20121. As with some other non-
regulated DBPs, nitrogenous DBPs may be more toxic than the regulated ones (Harkness etal..
2015: McGuire etal.. 2014: Parker etal.. 2014).
As discussed in Section 8.4 and Text Box 8-3, removal of dissolved solids, including chloride and
bromide, requires advanced treatment processes such as reverse osmosis (RO), distillation,
evaporation, or crystallization. Unless the treatment plant receiving the high-TDS wastewater
employs processes specifically designed to remove these constituents, effluent discharge may
contain high levels of bromide and chloride. Drinking water systems with intakes downstream of
these discharges may receive water with correspondingly higher levels of bromide and chloride
and may have difficulty complying with SD WA regulations related to DBPs.
8.5.1.2 Effects on Receiving Waters
Studies show that discharges from oil and gas wastewater treatment facilities can elevate TDS,
bromide, and chloride levels in receiving waters, and potential impacts may be detectable far
downstream (> 1km) of an outfall (States etal.. 2013: Warner et al.. 2013a: Wilson and Van Briesen.
20131. The work by Landis etal. f 20161 in the Allegheny River mentioned above is consistent with
these findings. The authors studied the impacts of a CWT accepting oil and gas wastewater on water
quality at a downstream drinking water intake. They found that compared to data from upstream
(background) locations, bromide concentrations at the intake were increased by 53% at low
streamflow and 22% during high streamflow.1
1 Background samples are those taken from locations upstream of, and therefore unaffected by, permitted facilities.
8-56
-------
Chapter 8- Wastewater Disposal and Reuse
Elevated TDS, chloride, and bromide can serve as indicators of potential influence from hydraulic
fracturing wastewater in surface water and can also raise concerns about DBP formation in
downstream drinking water systems. Elevation of bromide has been shown to place a burden on
downstream drinking water systems. The Pittsburgh Water and Sewer authority (PWSA) drinking
water system concluded that elevated bromide in their source water led to elevated TTHMs in their
finished drinking water, with a substantial increase in the percentage of brominated TTHMs (States
etal.. 20131. The utility modified their treatment process and proposed improvements to their
storage facilities to address the elevated TTHM levels in the distribution system fChester Engineers.
20121.
Conversely, changes in regional wastewater handling that reduce bromide discharges can be
reflected in receiving waters. A three-year study at water intakes downstream of wastewater
discharges on the Monongahela River in western Pennsylvania evaluated water chemistry in the
context of flow measurements. The authors concluded that an overall decrease in bromide
concentrations at drinking water intakes from 2010 to 2012 was likely associated with shale gas
operators voluntarily ceasing the practice of sending high-bromide wastewaters to treatment
facilities that discharge to surface waters without adequate TDS removal fWilson and Van Briesen.
20131.
Elevated TDS and halides need to be interpreted in the context of other inputs into a watershed. An
EPA source apportionment study of the Allegheny River in Pennsylvania fU.S. EPA. 2015ol found
that CWTs accounted for almost 90% of the bromide at one drinking water treatment plant intake
and 37% of the bromide at another intake. Other sources include coal-fired power plants and acid
mine drainage. Furthermore, although effluent is diluted when discharged to a water body, this may
not always be sufficient to avoid water quality problems if there are existing pollutant loads in the
waterbody from other contributors (e.g., such as acid mine drainage or power plant effluent)
(Ferrar etal.. 2013). Warner etal. (2013a) evaluated effluent from the Josephine Brine Treatment
Facility, which treated both conventional and unconventional (as defined by PA DEP) oil and gas
wastewater at the time of the study. The authors concluded that even a 500 to 3,000-fold dilution of
the wastewater would not reduce bromide levels to background. Modeling by Weaver etal. (2016)
suggests that bromide levels in receiving streams can be improved by reducing concentrations in
the effluent, discharging during periods of high streamflow, and discharging intermittently
(pulsing). (See Appendix F for additional description of modeling studies.)
In addition to concerns about formation of DBPs within downstream drinking water systems,
treatment at the upstream CWTs and POTWs themselves can also produce DBPs if the facilities
disinfect prior to discharge. The DBPs may then be released into receiving waters and increase
concerns about the total loads of brominated and iodinated DBPs at downstream drinking water
systems fHladik et al.. 20141. A study by Hladik etal. f 20141 documented brominated and iodinated
DBPs at the outfalls of CWTs and POTWs treating both conventional and unconventional
wastewater and noted that this DBP signature was different than for those plants that did not
accept oil and gas wastewater.
8-57
-------
Chapter 8 - Wastewater Disposal and Reuse
8.5.1.3 Other Constituents That Can Affect Downstream DBP Formation
In addition to halogens, organic matter and ammonium can also be present in hydraulic fracturing
wastewater (Chapter 7; Appendix E) and can have an influence on the formation of DBPs at
downstream drinking water systems (Harkness etal.. 2015). Experimental work by Parker et al.
(20141 found that a mixture of river water with 1-2% flowback by volume could contribute to DBP
formation due to the higher dissolved organic carbon content of the flowback. Harkness et al.
f20151 studied the chemical composition of water associated with oil and gas production and found
high concentrations of ammonium in the flowback and produced waters. Elevated levels of
ammonium can convert chlorine to chloramines at downstream drinking water treatment plants.
This could have an impact on the plant's disinfection practices because chloramines are a weaker
disinfectant than chlorine f Harkness etal.. 2015: Parker etal.. 20141.
8.5.1.4 Mitigating Impacts from TDS and Halides on Drinking Water Utilities
High bromide concentrations and low flow conditions in waterways have been shown to increase
DBP formation in downstream drinking water systems (States etal.. 20131. Most drinking water
treatment plants are not designed to address high concentrations of TDS (including bromide and
iodide), limiting their options for restricting the formation of brominated and iodinated DBPs when
these halides are present.
To mitigate these impacts, one strategy that was implemented in Pennsylvania was to disallow
influent of high-TDS wastewaters to POTWs and CWTs that discharged to streams and were not
designed to treat for TDS. Wilson and Van Briesen (2013) showed that this strategy was effective
for reducing bromide concentrations at drinking water utilities downstream from POTWs and
CWTs that had formerly accepted hydraulic fracturing wastewaters f States etal.. 2013: Warner et
al.. 2013a: Wilson and Van Briesen. 20131. Alternatively, advanced treatment processes such as
reverse osmosis, distillation, evaporation, and crystallization, can be employed to reduce
constituents that contribute to high TDS (e.g., such as chloride, bromide, and iodide), reducing
impacts on surface waters and, subsequently, downstream drinking water utilities. Strategies such
as discharging during higher streamflow periods and using a pulsing or intermittent discharge
could also reduce the frequency and severity of potential impacts on drinking water systems from
elevated TDS.
8.5.2 Radionuclides
Potential impacts on drinking water resources from TENORM associated with hydraulic fracturing
wastewater can arise through a number of pathways, including: treated wastewater in which
radionuclides were not adequately removed; accumulation of radionuclides in surface water
sediments downstream of wastewater treatment plant discharge points; migration or mobilization
from soils that have accumulated radionuclides from previous activities such as pits or land
application; and inadequate management of treatment plant solids (such as filter cake), landfill
leachate, or sediments in pits or tanks that have accumulated radionuclides.
An additional concern is the potential for underestimation of radium concentrations in hydraulic
fracturing wastewater due to the high TDS content When using wet chemical techniques, high TDS
8-58
-------
Chapter 8- Wastewater Disposal and Reuse
concentrations can result in poor recovery of some chemical constituents. For radium, recovery
may be as low as <1% in a high-salt matrix fNelson et al.. 20141. This may lead to the inability to
identify an impact on drinking water resources or an underestimation of the severity of an impact.
Research suggests that spectroscopic methods are more appropriate for analysis of radium in high-
TDS wastewaters (Nelson et al.. 20141.
A recent study by the PA DEP (PA DEP. 2015bl provides information that helps fill a general data
gap regarding TENORM content in oil and gas wastes that are treated and discharged to surface
waters. The study, although not exclusive to Marcellus wastes, was motivated by concerns over an
increase in radionuclides in oil and gas wastes observed during the expansion of Marcellus Shale
production. The study began in 2013 and quantified radionuclide (radium-226, radium-228, K-40,
gross alpha, and gross beta) levels at CWTs, POTWs, well sites, and landfills and discussed human
health and environmental implications. Other relevant studies addressing radionuclides focus on
CWTs that have handled Marcellus wastewater, evaluation of solids in storage pits, and analysis of
scale on pipes and tanks.
8.5.2.1 Effluent from POTWs
In Pennsylvania between 2007 and 2010, TENORM-bearing wastewaters were sent to POTWs,
which are generally not required to monitor for radioactivity (Resnikoffetal.. 20101. Although
management of Marcellus wastewaters via POTWs has declined, there is still potential for input of
radionuclides to surface waters via discharge of CWT effluent either directly to surface water or
indirectly through discharge to POTWs. The potential for TENORM to pass through treatment at
POTWs is one of the concerns addressed in the EPA's recently promulgated pretreatment standards
for unconventional oil and gas wastewaters that discharge to POTWs.
Six of the POTWs in the PA DEP TENORM study received effluent from a CWT along with municipal
wastewater. Note that the CWTs in the study are not described as receiving exclusively Marcellus
wastewater. The POTWs that receive both CWT effluent and municipal waste had radium in their
effluent (overall average effluent radium-226 concentration of 103 pCi/L, with a range of <35 to
340 pCi/L). Those POTWs receiving only municipal wastewater also contained radium, with an
average effluent radium-226 concentration of 145 pCi/L.1 These concentrations are many times
higher than the MCL for radium (5 pCi/L) and are also orders of magnitude higher than typical
background values; radium-226 in river water generally ranges from 0.014 pCi/L to 0.54 pCi/L
(IAEA. 20141.2
1 These values are for unfiltered samples. In filtered samples, the POTWs that receive both CWT effluent and municipal
waste had higher average radium-226 values than those for POTWs only treating municipal waste (497 pCi/L vs. 85
piCi/L]. Filtered samples are passed through a filter to remove fine particles; concentrations of constituents in filtered
samples are often lower than in unfiltered samples. However, liquid samples in this study were filtered after preservation
with acid. Therefore, the difference between unfiltered and filtered samples may not be reliable.
2 A confounding issue for this study is that it was not clear why the radium-226 concentrations were comparable or
higher for those POTWs not receiving oil and gas CWT effluent. However, sample sizes were small and possible
alternative sources for the radium were not discussed. The report also did not describe how it was verified that the
POTWs did not receive contributions from oil and gas wastewater.
8-59
-------
Chapter 8 - Wastewater Disposal and Reuse
8.5.2.2 Effluent from CWTs
Four of the ten CWTs sampled during the PA DEP TENORM study fPA DEP. 2015bl discharged to
surface water under a National Pollution Discharge Elimination System (NDPES) permit, and the
other six discharged to POTWs. The average radium-226 concentration in the effluent from the
CWTs (1,840 pCi/L for unfiltered samples) was an order of magnitude higher than in effluent from
the POTWs. Samples of treated wastewater from zero-discharge facilities contained higher
concentrations, averaging 2,610 pCi/L radium-226 and 295 pCi/L radium-228 fPADEP. 2015bl
The treated wastewater from these zero-discharge facilities will likely be reused for subsequent
hydraulic fracturing jobs, postponing the need for disposal, but reuse could result in overall
increases in some constituents of concern due to repeated passage through the subsurface. In
addition, there is also a potential for impacts on drinking water resources from spills and leaks
associated with wastewater storage and handling at these facilities.
Sampling done at the Josephine Brine Treatment Plant in western PA from 2010-2012 ("Warner et
al.. 2013a) detected radium in the effluent (mean values of 4 pCi/L of radium-226 and 2 pCi/L of
radium-228). Treatment at the facility removes radium by coprecipitation with barium sulfate. The
authors note that if the activities of radium-226 and radium-228 in Marcellus brine influent at the
CWT are similar to those reported by other researchers ("Rowan etal.. 20111. then the CWT
achieved a 1,000-fold reduction in radium content (This facility also accepted conventional oil and
gas wastewater.) The detection of radium in the effluent from this CWT suggests that if the influent
concentration is extremely high, radium will still be found in the effluent of a treatment plant even
if the treatment process removes a high percentage (see Section 8.4 and Appendix F for additional
discussion on constituent removal efficiencies at CWTs).
8.5.2.3 Accumulation in Sediments
In addition to concerns about TENORM in discharges to surface waters, studies have shown the
potential for a legacy of radionuclide accumulation in surface water sediments. The PA DEP
TENORM study (PA DEP. 2015b) found radium in sediments near the outfalls for CWTs (averages of
84.2 pCi/g and 19.8 pCi/g for radium-226 and -228, respectively) and three POTWs receiving
treated oil and gas wastewater from CWTs (radium-226 and radium-228 concentrations ranging
from 1.8 to 18.2 pCi/g). Typical background soil levels of radium are approximately 1 to 2 pCi/g (PA
DEP. 2015bl.
Warner etal. f2013al measured radium-226 levels in stream sediment samples at the point of
discharge of a CWT that had treated both conventional oil and gas wastewater and unconventional
Marcellus wastewater. They found concentrations approximately 200 times greater than upstream
and background sediments. This indicates the potential for accumulation of contaminants in
localized areas near wastewater discharge facilities. Although the CWT studied by Warner et al.
f2013al also accepted conventional oil and gas wastewater, the authors observed that the radium-
228/radium-226 ratio in the river sediments near the discharge (0.22 - 0.27) is consistent with
ratios in Marcellus wastewater. The authors indicate that the radium likely accumulated in the
sediments, originating from the discharge of treated unconventional Marcellus oil and gas
wastewater. Accumulation of TENORM can also occur in sediments receiving discharged effluent
8-60
-------
Chapter 8- Wastewater Disposal and Reuse
from landfills that accept oil and gas wastes. In the PA DEP TENORM study (PA DEP. 2015b).
samples of impacted soils were collected atthree landfill outfalls. Radium-226 and -228 were
detected in all samples (2.82 to 4.46 pCi/g and 0.979 to 2.53 pCi/g, respectively).
A study by Skalak etal. (20141. on the other hand, did not find elevated levels of alkali earth metals
(including radium) in sediments just downstream of the discharge points of five POTWs that had
previously treated Marcellus wastewater. These inconsistencies among studies suggest that
accumulation of contaminants in sediment may depend on treatment processes and their removal
rates for each constituent as well as stream chemistry and hydrologic characteristics.
Contamination with radium-226 would potentially be long lived because of the long half-life of
radium.1
The association of radium with sediments near discharge points is attributed to adsorption of
radium to the sediments, a process governed by factors such as the salinity of the water and
sediment characteristics. Increased salinity promotes desorption of radium from sediments, while
lower salinity promotes adsorption, with radium adsorbing particularly strongly to sediments high
in iron and manganese (hydr)oxides fPorcelli etal.. 2014: Gonneea etal.. 20081. Warner et al.
(2013a) speculate that when saline CWT effluent is discharged into stream water, the lower salinity
of the stream environment facilitates sorption of radium onto streambed sediments. The long-term
fate of radium sorbed to sediments depends upon changes in water salinity and the sediment
properties, including any reduction/oxidation chemical reactions that affect iron and manganese
minerals in the sediments. Additionally, the sediment may be physically transported downstream
due to high flows or if sediment is disturbed and resuspended.
8.5.2.4 Pits and Tanks
Where pits or impoundments are used, radionuclides may accumulate in the bottom sludges and
can also be found in soils once the pit is closed and leveled. A study of three centralized wastewater
storage impoundments in southwestern Pennsylvania (Zhang et al.. 2 015a) showed that radium-
226 accumulated in various components of the bottom solids, including through coprecipitation
with barium sulfate. Sludge from one pit showed a substantial increase in radium-226 between
sampling events 2.5 years apart (from 8.8 pCi/g to 872 pCi/g). The authors attributed the steep
increase to enrichment in radium during cycles of wastewater reuse. In Texas, accumulation of
radionuclides (potassium, thorium, bismuth, radium, and lead) was documented for two pits that
stored fluids associated with hydraulic fracturing fRich and Crosby. 20131. One pit was
decommissioned and used as farmland, and the other was active at the time of sampling. Analyses
of soil and sludge samples detected a number of radionuclides, including radium-226, radium-228,
thorium-228, strontium-90, and potassium-40 (radium-226 was only found atthe former pit). Rich
and Crosby f20131 note a total beta radiation value of 1,329 pCi/L in one sample from the active pit
They note that this value exceeded regulatory guidelines even though the values for individual
1 The half-life of radium-226 is approximately 1,600 years, while the half-life of radium-228 is 5.76 years. The half-life is
the time it takes for half of the nuclei in a sample of a radioactive element to decay. After two half-lives, one fourth of the
original sample will be left, and after three half-lives there will be one eighth of the original sample remaining, and so
forth.
8-61
-------
Chapter 8 - Wastewater Disposal and Reuse
radionuclides did not exceed regulatory guidelines, suggesting that using a single radionuclide (i.e.,
radium) as an indication of exposure can underestimate total radioactivity.
Although the sample sizes were small for both the Zhang etal. (2015a) and the Rich and Crosby
(2013) studies, the results suggest that radionuclides associated with sediments from some pits
could have potential impacts on surface water or groundwater. These studies illustrate the need for
appropriate management where wastes have high TENORM content. Rich and Crosby (2013) note
that pits are often found in agricultural regions. If pit solids that are incorporated into soils (e.g.,
during draining and leveling or during land application) contain radionuclides, they may reach
surface water in runoff or leach from the solids and migrate to groundwater. In active pits, Rich and
Crosby f20131 note that TENORM in the contents may be deposited onto crops and soil through
aerosolization or breaching. The Pennsylvania study f Zhang etal.. 2015al suggests that landfill
leachate may be affected by receiving sludges from impoundments that store produced water and
will need to be managed appropriately.
With radium-226 values of 121 pCi/g and 872 pCi/g, sludges from the pits studied by Zhang et al.
f2015al exceeded the limit for disposal as a nonhazardous solid in a municipal or industrial solid
waste landfill but would meet the radium-226 limits for disposal in a hazardous waste landfill.
There are currently no federal requirements to test solid residuals for radionuclides before
disposal. At landfills studied in the PA DEP TENORM report (PA DEP. 2015b). seven samples of
treated effluent from nine facilities that accept oil and gas waste had radium-226 values ranging
from 105 pCi/L to 378 pCi/L and radium-228 values ranging from <6 pCi/L to 1,100 pCi/L.
Untreated effluent from the nine landfills had radium-226 contents ranging from 70 to <139 pCi/L.
The study authors conclude that there is "limited potential" for environmental impacts from spills
or discharges of leachate from these facilities.
Where wastewater is stored in tanks, TENORM concentrations can increase through radioactive
ingrowth.1 Radium-226 and radium-228 are generally considered the radionuclides of greatest
concern in wastewaters and are the most frequently measured. But recent research indicates that
in closed environments such as tanks, where the radium decay product radon cannot escape, total
radioactivity may increase due to ingrowth of other decay products of radium such as Pb-210, Po-
210, and Th-228 (Nelson etal.. 2015). Experimental work by Nelson et al. found that concentrations
of these decay products in Marcellus produced water that was stored in a sealed drum started
growing immediately. Concentrations started at zero and reached 10.49 pCi/L for Po-210 and 155
pCi/L for Th-228 over the first 50 and 66 days of storage, respectively. The authors note that these
decay products are not soluble, would be associated primarily with particulates, and could be
bioavailable. This study demonstrates that analyzing for radium will not provide a complete
indication of sample radioactivity if the water is stored in a closed environment and that
subsequent management decisions would need to take into account possible increases in
radioactivity due to ingrowth.
1 The ingrowth, or growth within a sample, of radioactive daughter products from radionuclides initially present in the
sample can cause greater radioactivity than that resulting from the parent radionuclides alone.
8-62
-------
Chapter 8- Wastewater Disposal and Reuse
8.5.2.5 Other Solids
Other solid wastes associated with unconventional oil and gas production that may contain
radionuclides include solid residuals from POTWs and CWTs and scale in oil and gas equipment.
Filter cake samples from POTWs were found by PA PEP (2015b) to have highly variable radium-
226 concentrations, with an average of 16 pCi/g, while typical soil concentrations in Pennsylvania
have been found to be less than 2.5 pCi/g fGreeman et al.. 19991. Filter cake from CWTs had an
average radium-226 concentration of 111 pCi/g. The authors conclude thatthere could be impacts
on surface waters through spills or effects on groundwater from landfill leachate containing
contaminants originating in residuals sent to landfills.
Accumulation of TENORM-bearing scale in CWTs or POTWs may continue to affect the treatment
plant even after discontinuing treatment of wastewaters containing high radionuclide
concentrations. Radium can adsorb onto scales in pipes and tanks and can also be removed from
water by coprecipitation if sulfate or carbonate is added to hydraulic fracturing wastewater to
precipitate calcium, barium, or strontium (Kappel etal.. 2013: USGS. 2013al. Pipe scale in oil and
gas production facilities has been found to have radium concentrations as high as 154,000 pCi/g,
although concentrations of less than about 13,500 pCi/g are more common fSchubert etal.. 20141.
A similar issue, the potential for accumulation and possible release of radionuclides and other trace
inorganic constituents in water distribution systems, has gained attention, with the potential for
drinking water concentrations to exceed drinking water standards fWater Research Foundation.
20101. Scale eventually removed from pipes or other oil and gas equipment can end up in landfills
and then leach into groundwater or run off to surface water fUSGS. 2013al. Also, laboratory
research suggests that radium in land-applied barium sulfate scales from conventional oil and gas
operations may become mobilized by microbial processes, rendering the radium more mobile and
bioavailable f Matthews etal.. 2006: Swann etal.. 20041: see discussion in Section 8.4.6.1.
Monitoring would be needed in order to ascertain the potential for accumulation and release of
radionuclides from systems that have treated or continue to treat hydraulic fracturing wastewaters
with elevated TENORM concentrations.
8.5.2.6 Road Spreading
Salt and radionuclide accumulation can occur near road spreading sites; one study in Pennsylvania
describes a roughly 20% increase in average radium-226 concentrations in soils near five roads
where wastewaters from conventional operations had been spread for de-icing (Skalak et al.. 20141.
However, the standard deviation for the samples was large (24 pCi/g), and background
concentrations were approximately 1 pCi/g. Should significant accumulation of radionuclides in
soils near roads occur, it would present a vehicle for potential impacts on drinking water resources.
The frequency with which hydraulic fracturing wastewater contributes to this type of impact
depends on state-level regulations dictating whether the wastewater can be used for road
spreading.
8-63
-------
Chapter 8 - Wastewater Disposal and Reuse
8.5.2.7 Potential for Monitoring
Effluent from treatment plants (e.g., CWTs, POTWs) and receiving waters can be monitored for
radionuclides. Research suggests thatradium-226 and radium-228 are the predominant
radionuclides in Marcellus Shale wastewater, and they account for most of the gross alpha and
gross beta activity in the waters studied (Rowan etal.. 20111. Gross alpha and gross beta
measurements may, therefore, serve as an effective screening mechanism for the presence of
radionuclides in hydraulic fracturing wastewater. This in turn can help in evaluating management
strategies. Portable gamma spectrometers allow rapid screening of wastewater effluent. Sediments
can also be measured for radionuclide concentrations at discharge points. If an accurate assessment
of total radioactivity is needed rather than a screening, measuring radium content may not be
adequate depending upon how wastewater has been stored. Analyses of other radionuclides such
as Pb-210, Po-210, and Th-228 may be warranted, especially if the wastewater has been stored in
closed loop systems.
8.5.3 Metals
Given the presence in hydraulic fracturing wastewaters of some heavy metals, as well as barium
and strontium concentrations that can reach hundreds or even thousands of milligrams per liter
(Table 7-5), surface waters may be impacted if discharges from CTWs or POTWs indirectly
receiving oil and gas wastewater via CWTs are not managed appropriately or if spills occur.
Common treatment processes, such as chemical precipitation, are effective at removing many
metals (Section 8.4). Effluent sampling results collected between October 2011 and February 2013
from seven facilities in Pennsylvania treating oil and gas wastewaters were requested by the EPA.
The results revealed low to modest concentrations of copper (0-50 ju.g/L), zinc (14-256 ju.g/L),
and nickel (8 - 22 |Jg/L) (U.S. EPA. 2015f. g). However, metals such as barium and strontium were
found to range from low to elevated concentrations in the effluent for some of the facilities. The
data showed effluent barium concentrations ranging from 0.35 to 25 mg/L (median of 3.5 mg/L
and average of 6.7 mg/L). For results that were greater than 2 mg/L, the drinking water MCL for
barium was exceeded. Strontium concentrations ranged from 0.36 to 546 mg/L (median of 297
mg/L and mean of 236 mg/L) (U.S. EPA. 2015g). (See Chapter 9 for information on health effects for
barium and strontium.)
Volz etal. f20111 discussed a December 2010 effluent sampling effort at a Pennsylvania CWT that
had been treating both conventional and Marcellus wastewater; they measured average barium and
strontium concentrations of 27 mg/L and nearly 3,000 mg/L, respectively (eight samples from the
one plant) (Volz etal.. 20111. NPDES compliance data submitted for 2011 shows that effluent from
the same CWT had average barium effluent levels ranging from 26 to 98 mg/L in the months prior
to PA DEP's April 2011 request to cease sending hydraulic fracturing wastewater to this and other
facilities exempt from the 2010 TDS regulation (U.S. EPA. 2015f. g). After May, 2011, barium
effluent concentrations dropped to average values of 9 to 22 mg/L. The facility is scheduled to
upgrade its TDS removal capabilities, which should help decrease concentrations of metals in the
effluent.
8-64
-------
Chapter 8- Wastewater Disposal and Reuse
Limited data are available on metal concentrations in wastewater and treated effluent that are
directly discharged; additional information would be needed to assess whether there could be
downstream effects on drinking water utilities. NPDES discharge permits, which restrict TDS
discharge concentrations, would likely reduce metal effluent concentrations due to the additional
treatment necessary to minimize TDS.
8.5.4 Volatile Organic Compounds
Benzene is a common constituent in hydraulic fracturing wastewater, and it is of concern due to
recognized human health effects. A wide range of concentrations of BTEX compounds occurs in
wastewater from the Barnett and Marcellus shales. Natural gas formations generally produce more
BTEX than oil formations (Veil etal.. 20041. and lower concentrations of BTEX naturally occur in
wastewater from CBM production (Appendix Table E-9). The organic chemistry of Marcellus
wastewater has been found by Akob etal. (2016) to be more variable than that of inorganic
constituents, indicating the need to consider the concentrations of organic compounds when
planning wastewater management
Processes such as air stripping or dissolved air flotation can remove VOCs during treatment, but if
treatment is not adequate prior to discharge, the VOCs may reach water resources. For example, the
average benzene concentration measured in the discharge from a Pennsylvania CWT in December
2010 was 12 |ig/L fVolz etal.. 20111 exceeding the MCL for benzene of 5 [ig/L.1 The facility was
receiving wastewater from both conventional and unconventional operations at that time. Ferrar et
al. (2013) measured for BTEX in effluent from the same facility, and mean concentrations among
the four compounds ranged from approximately 2 to 46 |ig/L. Concentrations were lower for
samples taken after May 19, 2011 (when Marcellus operators voluntarily stopped sending
wastewater to POTWs and CWTs exempt from the 2010 TDS regulation), and the difference
between pre and post May 2011 sampling was considered statistically significant
Spills and leakage from pits creates another potential route of entry to drinking water resources, as
described in Section 8.4.5. Akob etal. T20161 documented the microbial degradation of organic
compounds in Marcellus produced water and note that more research is needed to evaluate how
this could mitigate the migration of organic constituents in the event of spills or leaks.
8.5.5 Semi-Volatile Organic Compounds
Little is known about the fate of the SVOC, 2-butoxyethanol (2-BE) (an antifoaming and anti-
corrosion agent used in slick-water) (Volz etal.. 2011) or its potential impact on surface waters,
drinking water resources, or drinking water systems. This compound is very soluble in water and is
subject to biodegradation, with an estimated half-life of approximately 1-4 weeks in the
environment fWess etal.. 19981. It is classified by the EPA's Integrated Risk Information System
(IRIS) as not likely to be carcinogenic to humans, and the International Agency for Research on
Cancer (IARC) classifies it as having insufficient evidence to determine carcinogenicity (see Chapter
9 for more information). 2-BE was detected in the discharge of a Pennsylvania CWT at
1 Among the BTEX compounds, the MCL for benzene is the lowest at 5 ng/L; the MCL for ethylbenzene is 700 ng/L, the
MCL for toluene is 1,000 ng/L, and the MCL for xylenes is 10,000 ng/L.
8-65
-------
Chapter 8 - Wastewater Disposal and Reuse
concentrations of 59 mg/L (Volz etal.. 20111. Ferrar etal. (20131 detected 2-BE in the effluent from
a CWT in western Pennsylvania at average concentrations of 34 - 45 mg/L; the latter value was
measured when the CWT was receiving only conventional oil and gas wastewater. Data are lacking
on 2-BE concentrations in surface waters that receive treated effluent from hydraulic fracturing
wastewater treatment systems.
Polyaromatic hydrocarbons (PAHs; a group of SVOCs) have been found in hydraulic fracturing
wastewater (Section 7.3.4.7, Table 7-6). PAHs detected in an unlined pit containing oil and gas
wastewater near the Duncan Oil Field in New Mexico were also detected in soils 82 ft (25 m)
downgradient at concentrations ranging from 2,000 to 4,900 |J.g/kg and 164 ft (50 m)
downgradient, with concentrations ranging from 22 to 370 |J.g/kg fSumi. 2004: Eiceman. 19861.
8.5.6 Oil and Grease
Oil and grease in oil and gas wastewater can come from the formation or from oil-based drilling
fluids. Typically, oil and grease are separated from the wastewater before discharge either by heat
treatment or by gravity separation followed by skimming. If these processes are inefficient, oil and
grease can be integrated with the discharge to surface waters. For example, in some cases, oil and
grease are allowed to separate in pits, and water is then withdrawn from the lower part of the pit If
the oil layer is allowed to drop to the level of the standpipe or if the water is agitated, oil and grease
may be discharged along with the water. Oil and grease are also often dispersed in wastewater in
the form of small droplets that are 4 to 6 microns in diameter. These droplets can be difficult to
remove using typical oil/water separators (Veil etal.. 20041.
A study was conducted in Wyoming by the U.S. Fish and Wildlife Service from 1996 to 1999 of sixty
five oil and gas sites that discharge to ephemeral streams and subsequently to wetlands. Fifteen
percent of the wetlands receiving wastewater contained oil-stained vegetation and had a visible oil
sheen on the sediments. In addition, ten of twelve sites that were randomly selected for water
sample collection (from oil field separator or skim pit effluent) exceeded the discharge limit of 10
mg/L for oil and grease with one site as high as 54 mg/L fRamirez. 20021.
8.6 Synthesis
A variety of strategies may be considered for the management of hydraulic fracturing wastewater.
Important factors for planning management include cost, logistics, wastewater composition,
wastewater volumes, and regulations. Available information suggests that Class IID wells regulated
under the UIC Program are the most frequently used wastewater management practice, but reuse,
sending to a CWT, and various other methods are also employed.
8.6.1 Summary of Findings
8.6.1.1 Wastewater Volumes
The most current national estimate of the total wastewater volume generated in the oil and gas
industry (both onshore and offshore) was 889.59 billion gal (21.18 billion bbls or 3.37 trillion L) in
2012, although this estimate is subject to a number of uncertainties (Veil. 20151. The total amount
8-66
-------
Chapter 8- Wastewater Disposal and Reuse
of wastewater generated may increase if hydrocarbon production increases in a region, although
Veil f20151 suggests that this trend may not hold true at the national level. Geographically, a large
portion of onshore oil and gas wastewater in the United States is reported to be generated in the
western part of the country, consistent with the areas where most oil and gas wells are located and
most production takes place.
Obtaining reliable national estimates of the amount of wastewater attributable to hydraulic
fracturing is a challenge. State data collection efforts vary, and in many states, production data do
not identify which wells have been hydraulically fractured. However, annual estimates compiled
from those states where hydraulic fracturing wastewater is identified range from hundreds of
millions to billions of gallons of wastewater generated each year. Data from individual states
indicate that along with an increase in the numbers of hydraulically fractured wells, associated
wastewater volumes have generally increased over the last several years into 2014. However, while
there is a general correlation between unconventional oil and gas production and wastewater
volume, the relationship is complicated by several factors such as timing of drilling and production.
More complete and comparable estimates of local, state, and regional wastewater volumes would
facilitate wastewater management on the part of operators as well as planning on the part of
agencies that oversee wastewater management.
8.6.1.2 Wastewater Management Practices
Hydraulic fracturing wastewater is managed in a variety of ways, including disposal via Class IID
wells; minimal treatment and reuse (in subsequent fracturing operations); more complete
treatment followed by reuse; sending to CWTs for treatment followed by direct discharge or
transfer to POTWs; evaporation; and other uses such as agriculture and wildlife use (allowed only
in the arid west when the wastewater is of good enough quality for such uses). All of these
strategies have the potential to affect drinking water resources. Wastewater management practices
continue to shift in response to evolving understanding of environmental concerns, emplacement of
new regulatory controls, changes in costs, and changes in technology and operator practices.
Unauthorized discharges of hydraulic fracturing wastewater have also been documented, and such
discharges can potentially impact drinking water resources.
As of 2015, available information suggests that Class IID disposal wells are a primary wastewater
management practice for operators in most of the major unconventional reservoirs in the United
States, with the notable exception of the Marcellus Shale region in Pennsylvania. Class IID wells
tend to be economically favorable, especially if they are located within a reasonable transportation
distance from well sites (U.S. GAP. 2012). In particular, large numbers of active injection wells are
found in Texas (7,876), Kansas (5,516), Oklahoma (3,837), Louisiana (2,448), and Illinois (1,054)
fU.S. EPA. 2016dl.
Pennsylvania is somewhat unique in having only nine Class IID wells (as of February 2015), along
with having experienced significant growth of shale gas production in the Marcellus and
corresponding production of large volumes of wastewater. Operators producing from
unconventional formations (as defined by PA DEP) have managed their wastewater through the use
of POTWs (a practice that is subject to recently promulgated regulations), CWTs, extensive reuse
8-67
-------
Chapter 8 - Wastewater Disposal and Reuse
for hydraulic fracturing operations, and hauling to disposal wells (to a lesser degree). The
wastewater management history in Pennsylvania provides an example of evolving strategies to
manage the treatment, discharge, storage, and reuse of hydraulic fracturing wastewaters that are
high in constituents of concern (e.g., bromide, TDS, and TENORM).
Reuse of hydraulic fracturing wastewater to formulate fluid for subsequent hydraulic fracturing
jobs is most prevalent in Pennsylvania (as high as 90%), with much of the reuse happening on-site
fPA DEP. 2015bl Reuse is practiced in other regions as well (e.g., Haynesville Shale, the Fayetteville
Shale, the Barnett Shale, and the Eagle Ford Shale), but at much lower rates (about 5 - 20%).
Reliable estimates are not available for all areas of the United States because waste management
practices are not consistently reported across all states. If hydraulic fracturing activity slows,
demand for wastewater for reuse will also likely decrease, and other forms of wastewater
management will be needed. Potential impacts associated with reuse center on concerns over the
storage of untreated or minimally treated wastewater on-site or transport to CWTs.
Treatment of hydraulic fracturing wastewater may be done at CWTs or using mobile or semi-
mobile systems designed for on-site use. Treatment at a CWT may be followed by direct discharge
by the CWT to surface water, indirect discharge to a POTW in accordance with recently
promulgated regulations, or reuse. Most CWTs treating hydraulic fracturing wastewater are located
in Pennsylvania (about 40 facilities), with a limited number in other states. CWTs vary widely in
treatment capabilities, ranging from producing high-quality effluent to minimal treatment for reuse.
Other wastewater management practices, such as evaporation and agricultural uses, represent a
smaller fraction of wastewater management nationally. These practices can, however, be locally
significant. Although specific instances of contamination were not identified for this assessment,
these practices could lead to impacts on drinking water resources if facilities are not properly
constructed and maintained or if water quality is not adequately characterized to ensure that
management is appropriate.
8.6.1.3 Treatment and Discharge
Wastewater that is treated and subsequently discharged by CWTs can result in impacts due to
inadequate treatment A frequently cited concern is the high TDS content in wastewaters from
unconventional formations, which poses challenges for treatment, discharge, and reuse. Treatment
processes such as sedimentation, filtration, flotation, and chemical precipitation are capable of
removing constituents such as oil and grease, major cations, metals, and TSS. They do not, however,
adequately reduce TDS in high-salinity wastewaters. More advanced processes such as reverse
osmosis (RO) or distillation are needed if TDS removal is required (Shaffer etal.. 2013: Younos and
Tulou. 2005). Most available information on treatment of hydraulic fracturing wastewater is based
on practices used in Pennsylvania because that is where most data have been collected.
Hydraulic fracturing wastewater discharged from treatment facilities without advanced TDS
removal processes has resulted in elevated TDS concentrations (including bromide, iodide, and
chloride levels) in receiving waters. Impacts from these discharges is due largely to the role of
bromide and iodide in DBP formation at downstream drinking water systems, potentially causing
8-68
-------
Chapter 8- Wastewater Disposal and Reuse
higher levels of harmful DBPs in finished drinking water.1 Modeling suggests that very small
percentages of hydraulic fracturing wastewater added to a river used as a source for drinking water
systems could cause a notable increase in DBP formation.
Radionuclides (i.e., TENORM), which are present in some hydraulic fracturing wastewaters, can
cause impacts if the wastewater is discharged without adequate treatment TENORMs have been
measured in effluent from wastewater treatment facilities receiving Marcellus wastewater (which
includes effluent sent for reuse and not discharged to surface water). Radium-226, radium-228,
gross alpha, and gross beta are most cited as the radioactive constituents of concern, likely due to
the availability of test methods for these constituents in wastewater. Radium concentrations can
range up to thousands or tens of thousands of pCi/L. Fewer data are available on concentrations of
uranium and other radionuclides in hydraulic fracturing wastewaters. Also, fewer data exist on
radionuclide concentrations in wastewaters from unconventional formations other than the
Marcellus, limiting our ability to assess potential impacts from TENORM on a nationwide basis.
Other constituents posing health or environmental concerns that can be discharged in inadequately
treated hydraulic fracturing wastewater include organic compounds, barium, strontium, and other
metals. Chemicals used in the fracturing fluid or their degradation products could also be present. A
variety of treatment processes can be used for removal of these contaminants, from commonly used
methods such as chemical precipitation and filtration to more advanced and more costly
techniques, such as reverse osmosis, distillation, and mechanical vapor recompression.
8.6.1.4 Storage and Disposal Pits and Impoundments
Regardless of the wastewater management practices used, some type of temporary storage of fluids
is generally required. Storage can be in the form of tanks as well as pits and/or impoundments. Pits
encompass a variety of structures, from on-site pits for storage at the well site to larger, centralized
facilities (typically referred to as "impoundments" or "ponds"). Some states allow evaporation pit
facilities or percolation pits as a means of wastewater disposal. The locations and number of pits
are not well documented in most states, nor are pits associated with hydraulic fracturing
operations necessarily identified, despite efforts by the U.S. EPA fU.S. EPA. 2003bl and
environmental groups such as SkyTruth to identify pits in use. Information that is typically
available on state websites includes permitted centralized commercial evaporation facilities
(COWDFs) most commonly used in the western United States.
Impacts on both groundwater and surface water resources due to inadequate pit capacities,
overfilling, and leaks have been documented. In extreme precipitation events, pits can be
overtopped. Leaks can occur if liners are compromised or were not used. With an increased
emphasis on reuse in some regions, the need for temporary storage of high-TDS wastewater
increases the potential for leaks and spills from pits and during fluid handling.
1 Some types of DBPs are regulated under SDWA's Stage 1 and Stage 2 DBP Rules, but a subset of DBPs, including a
number of chlorinated, brominated, nitrogenous, and iodinated DBPs, are not regulated. Brominated and iodinated DBPs
are more toxic than other species of DBPs.
8-69
-------
Chapter 8 - Wastewater Disposal and Reuse
Unlined pits, in particular, provide a pathway for contaminants to reach groundwater, and impacts
on groundwater from historic and current uses of unlined pits in the oil and gas industry have been
documented. The resulting contamination can be long-lasting. States have taken measures to phase
out the use of unlined disposal and storage pits, but unlined pits that are still in use can provide an
ongoing potential source of contamination for groundwater (Grinberg. 20141.
8.6.1.5 Residuals
Solid and liquid residuals associated with hydraulic fracturing wastewater (treatment residuals
from CWTs, sludges from tanks and pits, and pipe scale) could have impacts on drinking water
resources if not managed and disposed of properly. Liquid residuals are inappropriate for surface
water discharge or discharge to a POTW due to high concentrations of salts and other
contaminants; they are commonly disposed of in an injection well. Solid residuals may leach a
number of constituents, such as alkali metals, alkaline earth metals, and bromide. They can also
contain TENORM if radionuclides are present in the wastewater being treated. Given that residuals
are commonly disposed of in landfills, TENORM can be problematic due to the possibility of radon
emissions and radioactivity in the landfill leachate. Solids from pits or tanks can also contain
TENORM if the wastewater contains radionuclides, and one study has shown the potential for
radioactivity to increase in the closed environment of tanks.
8.6.2 Factors Affecting the Frequency or Severity of Impacts
The frequency and severity of impacts on drinking water resources from hydraulic fracturing
wastewater will depend upon the wastewater composition and volumes, and the mix of wastewater
management strategies used.1 The types of potential impacts (along with frequency and severity)
may shift in time as management practices change in response to evolving environmental,
regulatory, economic, or logistical drivers. The frequency and severity of impacts can also depend
on the size and initial quality of the drinking water resource and its proximity to wastewater
management operations.
8.6.2.1 Role of Changing Wastewater Management Practices
The most common disposal option for hydraulic fracturing wastewater is injection into Class II
disposal wells. If this option becomes restricted in a given location, the wastewater management
options could shift, at least locally, towards other options such as sending wastewater to CWTs for
treatment and either discharge or reuse. Although reuse avoids the immediate need to discharge
wastewater by directing it to ongoing hydraulic fracturing activities, the practice could concentrate
radionuclides or other constituents as fluid moves through cycles of reuse. Whether such
concentrations would be significant depends on the ratio of recycled to "fresh" water when the
wastewater gets reused. Alternatively, wastewater might need to be transported to more distant
1 Both national and state regulations affect the wastewater management practices used. At a national level, although the
EPA's oil and gas ELG regulations generally prohibit the direct discharge of oil and gas wastewater to waters of the U.S.,
treatment and discharge of hydraulic fracturing wastewater can occur under certain limited circumstances, such as under
an exemption authorizing discharge for agricultural and wildlife use in the arid west, or by Centralized Waste Treatment
facilities. For additional information on national regulations relevant to hydraulic fracturing wastewater management, see
Text Box 8-2.
8-70
-------
Chapter 8- Wastewater Disposal and Reuse
Class IID wells. This option, while attractive from the perspective of limited disposal impacts, could
increase the frequency of impacts from spills and leaks during transportation (see Chapter 7 for
discussion of roadway transport of produced water).
8.6.2.2 Treatment and Discharge
Both the frequency and severity of potential impacts on drinking water resources from treated
hydraulic fracturing wastewater depend on the influent concentrations of the constituents in the
wastewater and the type and adequacy of the treatment processes employed. If treatment and/or
blending is inadequate, the resulting quality in a receiving water could, for example, influence
formation of DBPs during subsequent drinking water treatment, impair biological treatment
processes, and release TENORMs into receiving waters.
The volume of treated effluent discharged relative to the size of the receiving water body is an
important local factor affecting the frequency and severity of potential impacts. Because of dilution
effects, drinking water systems drawing from smaller rivers will likely face greater challenges in
dealing with contaminants in their source water than systems drawing from larger rivers receiving
the same volume of effluent Seasonal changes in streamflow will also affect frequency and severity
by affecting the degree of dilution. Existing loadings of pollutants from other sources in a watershed
can increase the frequency and severity of potential impacts if the additional contributions from
hydraulic fracturing wastewater cause concentrations to exceed thresholds.
Direct discharges of wastewaters with lower TDS concentrations to ephemeral streams are allowed
in parts of the country where the wastewater is considered to be "of good enough quality" for
livestock watering and wildlife use, and the discharges may constitute a large portion of
streamflow. Permits authorizing such discharges may only require monitoring for a limited set of
constituents. In particular, they may not necessarily require monitoring for specific constituents
associated with hydraulic fracturing. The potential for water quality impacts from such discharges
depends upon whether chemicals used for fracturing fluid or maintenance (or their degradation
products) are present and at what concentrations. Long-term discharges to these ephemeral
streams could result in ongoing impacts if there are unrecognized or unaddressed water quality
issues.
Concerns about radionuclides in hydraulic fracturing wastewater have received considerable public
attention, especially in the Marcellus region. The severity and frequency of impacts on receiving
waters and sediments from TENORM depends upon the TENORM content in the wastewater
(highest in regions with NORM-rich formations), temporal variability in the wastewater
composition, and the treatment processes used. There are insufficient data to indicate whether
radionuclides from these wastewaters have reached drinking water intakes. However, data do
suggest that radionuclides can accumulate in sediments at or near discharge points from facilities
that treat and discharge oil and gas wastewater. A recent PA DEP study (PA DEP. 2015b) reported
8-71
-------
Chapter 8 - Wastewater Disposal and Reuse
radium in the effluent of both CWTs handling oil and gas wastewater and POTWs receiving effluent
from such facilities.
Analysis of TENORM concentrations in hydraulic fracturing wastewaters prior to treatment,
selection of appropriate treatment processes that adequately address the TENORM levels, and
monitoring of TENORM in the treated effluent and receiving waters could help address the
frequency and severity of potential impacts on drinking water resources in these areas. However, a
confounding issue is underestimation of radium concentrations when using traditional wet
chemical methods with high-TDS waters. This could consequently cause underestimation of
frequency or severity of impacts. Newer studies have begun to use gamma spectroscopy for better
recovery, which could help with more accurate assessment of frequency and severity of impacts
(Nelson et al.. 2014).
Accumulation of other contaminants such as organic compounds or metals in sediments at or near
discharge points is also possible. If the sediments are disturbed or entrained due to dredging or
flood events, contaminated sediments could be transported downstream closer to drinking water
systems. The fate of such sediments and likelihood of mobilization of contaminants originating from
hydraulic fracturing wastewaters have not been explored. The frequency and severity of impacts
related to contaminated sediments would depend on a number of site-specific factors such as
concentrations in the sediments, effluent quality, volume from the discharging facility, stream
water quality, and stream hydrodynamics.
8.6.2.3 Storage and Disposal Pits and Impoundments
Tanks, pits, and impoundments, ever-present at oil and gas operations and CWTs, provide an
opportunity for impacts on drinking water resources. Tanks are generally regarded as being safer
than pits in terms of containment, although recent research has shown the potential for an increase
in radioactivity in tank sediment if the wastewater contains TENORM. For pits and impoundments,
the likelihood and severity of impacts due to spills and leaks depends in part on state construction
and maintenance requirements for pits and how well these are observed. Frequency and severity of
impacts will be lessened by attention to design standards, competent construction, and operational
practices.
Liners, in particular, are an important measure to protect groundwater resources and are a
common aspect of pit construction requirements. Liner specifications address materials, thickness,
and leak detection. If a liner is compromised or nonexistent, the severity of impacts on groundwater
will be affected by the volume leaked, the composition of the water in the pit, the depth to the water
table, soil permeability, and the capacity of the soil to retain certain pollutants as the water
percolates through. If substantial sediment has built up in the bottom of the pit, then in the event of
a liner breach, contaminants may leach if the sediments permit water to pass through and into the
soil. The fate and transport of wastewater contaminants in the subsurface is governed by a complex
set of physical, chemical, and biological processes that dictate interactions between wastewater
constituents and soil minerals, degradation or transformation of wastewater constituents, and
possible mobilization of constituents in the soil under a pit (see Section 5.8 in Chapter 5 for a
thorough discussion of processes affecting movement of constituents in the subsurface). Duration
8-72
-------
Chapter 8- Wastewater Disposal and Reuse
of use is also a consideration; the longer a pit with a faulty or nonexistent liner receives
wastewater, the more severe the ultimate impact could be on underlying sediment and
groundwater.
In the event of overtopping of a pit due to overfilling or extreme weather, the severity of impacts on
surface water or groundwater will depend on the volume that overflows, wastewater composition,
distance to surface water (if wastewater flows over land), depth to the water table, and soil
properties (if the overflow infiltrates into the soil). If the overflow reaches a stream or river, the
size of the spill relative to stream size and flow rate could also affect the severity of the impact The
combined factors that can contribute to overflows include capacity of the pit, the volume of fluid
stored in the pit (i.e., freeboard) at the start of the precipitation event, and failure to monitor/
reduce pit fluid levels if needed.
As with concerns over discharges, the potential for impacts will be tied to other, existing stresses
within a watershed. If the surface water is already receiving pollutant loadings from other sources,
then an additional contribution from a pit-related leak or spill may not be as readily accommodated
without causing water quality impairment
8.6.2.4 Other Management Practices and Management of Residuals
Other management strategies such as irrigation, road spreading, and evaporation are less
frequently employed for hydraulic fracturing wastewaters. The severity of impacts on surface
waters from irrigation and road spreading will depend on the constituents in the wastewater (e.g.,
salts, radionuclides, and chemicals used in hydraulic fracturing), the distance to a receiving water,
and whether stormwater management measures exist to mitigate runoff. The factors influencing
whether constituents will migrate to shallow groundwater include depth to the water table,
precipitation, soil permeability, and the soil's ability to retain pollutants that can adsorb to
particles. If irrigation and road spreading are long-term management practices, the frequency of
impacts will likely be proportional to the frequency with which the practices are employed.
Liquid and solid residuals generated from the treatment, storage, and handling of hydraulic
fracturing wastewater have highly concentrated waste constituents. This could increase the
potential severity of impacts due to spills that reach surface water resources or leach to
groundwater. Potential impacts from management of residuals can be lessened in frequency and
severity through careful handling, adequate characterization (including TENORM content), and
selecting an appropriate disposal method, including use of a landfill that can accept TENORM waste
if needed.
8.6.3 Uncertainties
A full understanding of hydraulic fracturing wastewater management is limited by a lack of
available data in several areas. First, it is difficult to assemble a complete national- or regional-level
picture of wastewater volumes and the management practices used because the tracking and
availability of data vary from state to state. Although some states provide well-organized and
relatively thorough data, not all states collect or make such information available. It can be difficult
to identify wastewater volumes specifically associated with hydraulic fracturing (as compared to all
8-73
-------
Chapter 8 - Wastewater Disposal and Reuse
oil and gas production activities). Such data would be needed to place hydraulic fracturing
wastewater in the broader context of all oil and gas wastewaters. It is also generally difficult to
determine whether hydraulic fracturing wastewater is being injected under a given disposal well
permit because the permit rarely identifies which production wells are contributing to the
wastewater stream. Data are also generally difficult to locate for wastewater production volumes,
the chemical composition and concentrations in wastewater, and the management and disposal
strategies for residuals.
Up-to-date information on the volume of hydraulic fracturing wastewater disposed of via
underground injection by state is not uniformly available. Without this information, it is difficult to
assess whether disposal well capacity will become an issue in areas where hydraulic fracturing
activity is expected to increase or where use of disposal wells may become restricted locally or
regionally.
For CWTs permitted to discharge to surface water, the ability to assess the potential effects of these
discharges on drinking water resources is limited by the lack of effluent water quality data. Some
monitoring data are required by the permit, but the list of monitored constituents may be limited.
Selection of the appropriate water quality parameters to be monitored is critical to ensure that
potentially problematic constituents are identified (e.g., chemicals associated with hydraulic
fracturing fluids, maintenance chemicals, and degradation products of those chemicals). Some
chemicals used in fracturing fluids are not disclosed, and analytical methods are lacking for some
chemicals of concern and their degradation products.
Pollutant removal capabilities of the treatment facilities would also be valuable information to have,
but this would require well-coordinated collection of both influent and effluent samples; this type
of data is even less commonly available. In addition, the use of inappropriate analytical methods for
the high TDS wastewater associated with hydraulic fracturing operations can complicate the use of
available data. Methods used should be suitable for the highly complex matrix of contaminants
encountered with oil and gas wastewater to have confidence in the results of chemical analyses.
Monitoring of surface waters downgradient of discharges, such as screening with a TDS proxy (i.e.,
conductivity), would also help assess the frequency of impacts on receiving waters by hydraulic
fracturing activities (including spills and discharges of wastewater). Such data can also give an
estimation of the severity of those impacts. Other than a few studies in the Marcellus Shale region,
these types of water quality data are lacking. Existing data are also limited regarding legacy effects,
such as accumulation of contaminants in sediments at discharge points, soil accumulation due to
application of de-icing brines or salts from wastewater treatment, and handling of wastewater
treatment residuals.
Assessing longer-term impacts on surface water quality from hydraulic fracturing activities in
general is severely hampered by inadequate data. Bowen etal. (2015) state that available national-
level databases are inadequate for addressing the question of whether there is evidence of national-
level trends in surface water quality (as measured by specific conductivity and chloride) in areas
where unconventional oil and gas production is taking place. Work by the Northeast-Midwest
Institute and the USGS (Betanzo etal.. 2016) was undertaken to explore the types and amounts of
8-74
-------
Chapter 8- Wastewater Disposal and Reuse
data needed to assess whether shale gas development activities contaminate surface water or
groundwater in the Susquehanna River Basin. The focus was on longer-term cumulative impacts
because detection of such impacts requires water quality monitoring. Detection of impacts (in
either surface water or groundwater) requires a systematic monitoring approach that includes
sampling at appropriately selected locations at an adequate frequency and duration and for a suite
of water quality parameters to detect changes over time. Comparison sites without hydraulic
fracturing activity are needed as well. The authors concluded that the data necessary to detect
changes in surface water or groundwater due to hydraulic fracturing activities do not currently
exist for the Susquehanna River Basin.
8.6.4 Conclusions
Oil and gas operations in the United States generated an estimated 2.43 billion gal of wastewater
per day (about 60 million bbls/day) in 2012 (Veil. 2015). This includes wastewater associated with
hydraulic fracturing activities, although what portion of this oil and gas wastewater is attributable
to hydraulic fracturing operations is difficult to estimate. Available information indicates that the
majority of hydraulic fracturing wastewater is injected into Class IID wells regulated under the UIC
Program. In the Marcellus Shale region in Pennsylvania, this option is limited, and the majority of
wastewater is reused (either with or without treatment) for new hydraulic fracturing jobs.
Hydraulic fracturing wastewater may also be treated at a CWT and discharged by the CWT to
surface water or to a POTW. In the western United States, wastewater is used in other ways (e.g.,
livestock watering) if water quality allows. Wastewater is also sent to evaporation ponds for
disposal or stored on-site or in centralized pits or impoundments prior to final disposal or reuse.
Impacts on drinking water resources have resulted from discharges of inadequately treated
wastewater and from leaks, spills, and percolation associated with pits. Other mechanisms for
impacts include improper handling of treatment residuals or pit and tank sludges as well as
leaching and runoff associated with other wastewater management practices. The impacts related
to pits and residuals/sludges affect both surface water and groundwater; unlined pits or those with
compromised liners present a particular concern (see Chapter 7 for additional discussion of spills).
The constituents that have received the greatest attention in the literature include TDS, DBP
precursors (especially bromide), and radium, although hydraulic fracturing wastewater can contain
elevated concentrations of a number of organic and inorganic constituents of concern. Regardless of
the management option utilized, if the wastewater is not thoroughly characterized or sampling is
not conducted for analytes of concern, the severity and frequency of the impacts will be unknown
or unquantified. The nature and volume of wastewater generated through hydraulic fracturing
activities necessitate careful consideration of handling, treatment, and ultimate reuse or disposal to
ensure that water resources are not adversely impacted. There is also a need for reliable and
consistent waste generation data collection and reporting, improved efforts to characterize
wastewater quality (both treated and untreated), and systematic monitoring efforts to be able to
detect impacts on drinking water resources.
8-75
-------
Chapter 8 - Wastewater Disposal and Reuse
This page is intentionally left blank.
8-76
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chapter 9. Identification and Hazard
Evaluation of Chemicals across
the Hydraulic Fracturing Water Cycle
Abstract
This chapter identifies chemicals associated with hydraulic fracturing and provides an overview of the
potential human health effects associated with these chemicals, as well as variables that could affect
chemical occurrence in drinking water. The EPA has identified 1,606 chemicals associated with
hydraulic fracturing, including 1,084 chemicals that are used in hydraulic fracturing fluid and 599
chemicals that have been detected in produced water. There is some uncertainty surrounding this
chemical list, as it does not include a subset of chemicals that are classified as confidential business
information, and because understanding of produced water composition is constrained by limitations of
analytical chemistry as well as site-specific variations in the geochemistry of hydraulically fractured
rock formations.
The EPA used selected federal, state, and international sources of toxicological data to identify toxicity
values that can be used to support risk assessment for these chemicals, including chronic oral reference
values (RfVs) for noncancer effects and oral slope factors (OSFs) for cancer. Chronic oral RfVs or OSFs
were available for 173 (11%) of the total 1,606 chemicals. Health effects associated with chronic oral
exposure to these chemicals include carcinogenicity, neurotoxicity, immune system effects, changes in
body weight, changes in blood chemistry, liver and kidney toxicity, and reproductive and developmental
toxicity.
For the majority of chemicals that lack chronic oral RfVs or OSFs, risk assessors will have to turn
towards other sources of toxicological information that may have greater uncertainty than RfVs and
OSFs, including quantitative structure-activity relationship (QSAR) models or additional data from the
EPA's Aggregated Computational Toxicology Resource (ACToR) database. To understand whether
specific chemicals can affect human health through their presence in drinking water, data on chemical
concentrations in drinking water are needed. In the absence of these data, a preliminary analysis of
relative hazard potential for drinking water resources can be conducted using the multi-criteria decision
analysis (MCDA) approach outlined in this chapter. The MCDA combines data on toxicity, occurrence,
and physicochemical properties for selected subsets of chemicals and was used in this chapter to
highlight several chemicals that may be more likely than others to reach drinking water resources and
present a health hazard.
Overall, while evidence suggests that hydraulic fracturing has the potential to impact human health, the
actual human health implications are not well understood or well documented. Given that chemicals in
hydraulic fracturing fluids and produced water are likely to vary on a regional basis and even between
individual wells, the materials presented in this chapter are best applied for risk assessment and risk
management decision-making at the local level.
9-1
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
This page is intentionally left blank.
9-2
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9. Identification and Hazard Evaluation of Chemicals
across the Hydraulic Fracturing Water Cycle
9.1 Introduction
In this chapter, we present and integrate what is known about chemicals in the hydraulic fracturing
water cycle, and provide an initial assessment of the potential for these chemicals to impact human
health. The discussion is focused on the availability of toxicity values and qualitative assessments
that can be used to inform the risk assessment of these chemicals for oral exposure via drinking
water—in particular, the available noncancer oral reference values (RfVs), cancer oral slope factors
(OSFs), and qualitative cancer classifications.1'2,3 Public health impacts will depend upon both the
inherent toxicity of these chemicals and the potential for human exposure. We highlight several
field studies that have detected hydraulic fracturing-related chemicals in drinking water resources,
and discuss properties of chemicals related to environmental fate and transport that could affect
their potential impact on drinking water resources. To the extent information was available to do
so, knowledge of toxicological and chemical properties was combined to illustrate a preliminary
analysis of the relative hazard that these chemicals could pose to drinking water resources. The
data are presented in this chapter as follows:
Section 9.2 provides a brief background on public health concerns surrounding hydraulic fracturing
and unconventional oil and gas extraction, which have been highlighted in several recent studies.
Section 9.3 discusses how information sources were used to create a list of chemicals used in or
detected in various stages of the hydraulic fracturing water cycle. The consolidated chemical list
includes chemicals reportedly added to hydraulic fracturing fluids in the chemical mixing stage, as
well as fracturing fluid chemicals, formation chemicals, or their reaction products that may be
carried in produced water.
Section 9.4 provides an overview of the methods that were used for gathering information on
toxicity and physicochemical properties for all chemicals identified in Section 9.3, and outlines the
number of chemicals that had available data on these properties. For toxicological data, the primary
focus is on chronic oral RfVs, OSFs, and qualitative cancer classifications from selected data sources
that met the EPA's criteria for inclusion in this assessment This section also discusses other
1A reference value (RfV] is an estimate of an exposure for a given duration to the human population (including
susceptible subgroups] that is likely to be without an appreciable risk of adverse health effects over a lifetime. RfV is a
generic term not specific to a given route of exposure ("U.S. EPA, 2011f). In the context of this report, the term RfV refers to
reference values for non-cancer effects occurring via the oral route of exposure and for chronic durations, except where
noted.
2 An oral slope factor (OSF] is an upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a
lifetime oral exposure to an agent. This estimate, usually expressed in units of proportion (of a population] affected per
mg/kg day, is generally reserved for use in the low dose region of the dose response relationship, that is, for exposures
corresponding to risks less than 1 in 100 (U.S. EPA, 2011f).
3 Qualitative cancer classifications are a system used for the hazard identification of probable carcinogens, in which
human data, animal data, and other supporting evidence are combined to characterize the weight of evidence (WOE]
regarding the potential of an agent to cause cancer in humans.
9-3
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
potential sources of toxicity information: the use of quantitative structure-activity relationship
(QSAR) modeling to estimate chemical toxicity, as well as the availability of toxicological
information on the EPA's Aggregated Computational Toxicology Resource (ACToR) database. A
brief description of other potential tools and approaches that may be used by stakeholders for site-
specific evaluation of chemical hazards, but are not used in this report, is provided in Appendix G.
Section 9.5 describes the potential hazards of subsets of chemicals identified as being of interest in
previous chapters of this report This includes chemicals in hydraulic fracturing fluid (Chapter 5);
organic chemicals, inorganic chemicals, and pesticides detected in produced water (Chapter 7);
stray gas, such as methane (Chapter 6); and disinfection byproducts (DBPs) formed from
constituents of hydraulic fracturing fluid wastewaters (Chapter 8). We discuss instances in which
these chemicals have been detected in drinking water resources in areas of hydraulic fracturing
activity, and provide an overview of the available toxicological information for these chemicals.
Section 9.6 uses a multi-criteria decision analysis (MCDA) framework to provide a preliminary
analysis of the potential hazards of chemicals used in hydraulic fracturing fluids or detected in
produced water. The MCDA framework is used to integrate data on chemical toxicity, occurrence,
and physicochemical properties. In this context, occurrence and physicochemical properties are
used as metrics to estimate the likelihood that a chemical will reach and impact drinking water
resources. Chemicals considered in these hazard evaluations include a subset of chemicals
identified in the EPA FracFocus 1.0 project database, as well as a subset of organic chemicals that
have been detected in produced water.
This chapter is not a human health risk assessment. As shown in Text Box 9-1, risk assessment
consists of four basic steps: hazard identification, dose-response assessment, exposure assessment,
and risk characterization. This chapter provides an overview of hazard identification and dose-
response assessment for these chemicals, but lacks information to fully characterize exposure and
risk. In Section 9.5, we highlight instances in which these chemicals have been detected in drinking
water resources, but these data are only available for a small number of chemicals. The MCDA
approach in Section 9.6 provides a method for integrating data on toxicity and exposure potential,
but should be considered only as a preliminary analysis, and should not be used in place of local
data on chemical exposure.
This chapter is focused on potential human health hazards of chemicals for the oral route of
exposure (drinking water); therefore, the toxicological properties and physicochemical ranking
metrics described herein do not necessarily apply to other routes of exposure that may occur with
these chemicals, such as inhalation or dermal exposure. We additionally note that this analysis is
focused on individual chemicals, rather than mixtures of chemicals used as additives.
In general, characterizing chemicals and their properties on a national scale is challenging and the
use and occurrence of chemicals is likely to differ between geological basins and possibly on a well-
to-well basis (Chapters 5 and 7). Therefore, for the protection of human health at the local level,
chemical hazard evaluations are best conducted on a regional or site-specific scale. This level of
analysis is outside the scope of this report; however, the methods of hazard evaluation presented
9-4
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
here can also be applied on a regional or site-specific scale in order to identify chemicals that may
present a potential human health hazard.
Text Box 9-1. Applying Toxicological Data for Human Health Risk Assessment.
Understanding potential human health impacts requires knowledge not only of the inherent toxicity of the
chemicals found in contaminated environmental media, but also of the potential for exposure to these
chemicals. The process of evaluating the nature and probability of such impacts is known as human health
risk assessment. Overall, human health risk assessment includes four basic steps (U.S. EPA. 2016a):
1. Hazard identification: Examining whether a chemical has the potential to cause harm to humans and/or
ecological systems, and if so, under what circumstances.
2. Dose-response assessment: Examining the numerical relationship between exposure and effects.
3. Exposure assessment: Examining what is known about the frequency, timing, and levels of contact with a
chemical.
4. Risk characterization: Examining how well the data support conclusions about the nature and extent of
risk from exposure to a chemical. Information from the hazard identification, dose-response assessment, and
exposure assessment are summarized and integrated into quantitative and qualitative expressions of risk.
The RfVs and OSFs compiled by the EPA in this study pertain to the first two steps of human health risk
assessment: identifying chemicals that have the potential to affect human health (hazard identification), and
characterizing the exposure levels at which those effects occur (dose-response assessment). These toxicity
values may be used in combination with site-specific chemical exposure information (exposure assessment)
in order to evaluate potential human health risks (risk characterization). Qualitative cancer classifications
characterize the weight of evidence regarding the potential for a chemical to cause cancer, and therefore
provide additional information that can be used for hazard identification.
Toxicity information spans a wide range with respect to extent, quality and reliability. The RfVs, OSFs, and
qualitative cancer classifications compiled in this study are those identified by the EPA as being of the highest
quality and reliability, per the criteria discussed in this chapter. The QSAR-based toxicity estimates discussed
in this chapter are considered to be lower on the continuum of quality and reliability, but may provide useful
information pertaining to hazard identification and dose-response assessment when a chemical does not
have an RfV or OSF available. The EPA's ACToR database provides an aggregation of a wide range of
toxicological data that may also be useful for supporting the risk assessment of these chemicals. This chapter
provides information on whether a chemical has data available from ACToR; however, it is beyond the scope
of this report to evaluate the quality and reliability of data for these chemicals within ACToR, or to provide
guidance on how the data within ACToR should be used to support human health risk assessment.
9.2 Overview: Hydraulic Fracturing and Potential Impacts on Human Health
As discussed in the previous chapters of this assessment, a variety of chemicals are associated with
the hydraulic fracturing water cycle. Chemicals are added to hydraulic fracturing fluids at the
chemical mixing stage (Chapter 5), and then injected into the well (Chapter 6). These chemical
additives may return to the surface in produced water, along with chemicals from the formation
(Chapter 7). The chemicals in produced water may persist in wastewater effluents, with some
constituents contributing to the formation of disinfection byproducts in treated wastewater
(Chapter 8). Through events such as large volume spills (Figure 9-1), mechanical integrity failures,
9-5
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
hydraulic fracturing directly into groundwater resources, or discharge of inadequately treated
hydraulic fracturing wastewater, there are specific instances in which these chemicals have been
demonstrated to enter drinking water resources. Thus, there is potential for human exposure to
these chemicals, and the potential for adverse human health effects resulting from exposure.
Hydraulic fracturing-
related spill or release
chemicals penetrating soil *
layer above groundwater \
Volatilization
^vv,, .at* |,
Sorption
Surface Water
Unsaturated
Soil
Groundwater
chemical
groundwater
plume
Sorption
Transformation
Dissolution
Dispersion
Schematic of the Fate and Transport Processes
Governing Potential Impacts of Spills
and Releases to Drinking
Water Resources
Figure 9-1. Fate and transport schematic for a hydraulic fracturing-related spill or release.
Multiple authors have noted with the recent increase in hydraulic fracturing operations there may
be an increasing potential for significant public health and environmental impacts (Goldstein et al.
2014: Finkel etal. 2013: Korfmacher et aL 2013: Weinhold. 20121. These concerns have been
highlighted in several recent studies. An epidemiological study in Colorado demonstrated
residential proximity of pregnant mothers to natural gas wells is associated with an increased
incidence of congenital heart defects, and, to a lesser extent, neural tube malformations (Mckenzie
et al.. 20141. A similar study in Pennsylvania found pregnant mothers living closer to
unconventional natural gas wells were more likely to have infants that were small for gestational
age, with lower birth weights compared to infants from mothers living farther from wells fStacy et
al.. 2015). Residential proximity to natural gas wells in the Marcellus Shale is associated with an
increase the number of self-reported health symptoms, particularly upper respiratory and dermal
symptoms (Rabinowitz et al., 20151. chronic rhinosinusitis, migraine headache, and fatigue
symptoms (Tustin et al.. 20161. Laboratory studies have found that endocrine disrupting activity
measured using in vitro bioassays may be elevated in surface and groundwater at known hydraulic
fracturing spill sites (Kassotis et al.. 2014) and in surface water downstream from a hydraulic
9-6
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
fracturing wastewater injection facility (Kassotis etal.. 20161. Although none of these studies
demonstrate a direct effect of hydraulic fracturing activity on human health, and none of the
epidemiological studies provided measures of individual or population level exposures or
differentiated between drinking water contamination and other potential routes of exposure (e.g.,
air pollution), all are suggestive of a relationship between unconventional oil and gas development
and adverse health outcomes.
Previous chapters of this report have identified cases in which contamination of drinking water
resources could clearly be linked to hydraulic fracturing activity. For example, equipment failure
and human error have led to spills of hydraulic fracturing fluids across the country and have
affected the quality of drinking water resources fU.S. EPA. 2015m: Brantley etal.. 2014: COGCC.
2014: Gradient. 20131. Other studies highlighted in previous chapters provide indirect evidence
hydraulic fracturing activity has contaminated surface water or groundwater. For example, two
recent studies in the Marcellus Shale detected known hydraulic fracturing-related chemicals in
nearby groundwater wells, and used multiple lines of evidence to link the origin of these chemicals
to hydraulic fracturing activity fDrollette etal.. 2015: Llewellyn etal.. 20151.
There have also been documented impacts on ecological receptors. In Knox County, Kentucky,
retention pits holding hydraulic fracturing flowback fluids overflowed into Acorn Fork Creek during
the development of four natural gas wells, causing a decrease in pH and increase in conductivity.1
Organics and metals including iron and aluminum formed precipitates in the stream, and fish and
aquatic invertebrates were killed or displaced in a 2.7 km length of the stream affected by the
release (Papoulias and Velasco. 2013). A field report from the Pennsylvania Department of
Environmental Protection (PADEP) described a leak in an overland pipe carrying a mixture of
flowback and freshwater between two impoundments that impacted a 0.6 km length of a stream, in
which 168 fish and 6 salamanders were killed (PADEP. 2009b).
In some instances, chemical concentrations in drinking water resources impacted or potentially
impacted by hydraulic fracturing activity exceeded their respective primary or secondary
maximum contaminant level (MCL), or health advisory levels provided by the EPA's National
Primary Drinking Water Regulations (NPDWRs) and Drinking Water Standards and Health
Advisories (DWSHA) tables (U.S. EPA. 2012b). indicating that these chemicals are present at levels
that may impact human health.2 Examples will be discussed in Section 9.5. These studies generally
did not indicate the contaminated water was used directly for human consumption, so it is not clear
that people are being exposed to these chemicals at these levels. Nevertheless, these studies
indicate that hydraulic fracturing activity may contribute to the entry of chemicals into drinking
water resources at potentially harmful levels.
1 "Flowback" refers to fluids containing predominantly hydraulic fracturing fluid that return from a well to the surface.
Flowback is a type of produced water. See Chapter 7 for more details.
2 Maximum contaminant level (MCL]: The highest level of a contaminant that is allowed in drinking water. MCLs are set as
close to the maximum contaminant level goal (MCLG] as feasible using the best available analytical and treatment
technologies and taking cost into consideration. MCLs are enforceable standards. The MCLG is a non-enforceable health
benchmark goal which is set at a level at which no known or anticipated adverse effect on the health of persons is
expected to occur and which allows an adequate margin of safety fU.S. EPA. 2012bl
9-7
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Risk assessment and risk management decisions will be informed by scientific information on the
toxicity of chemicals in hydraulic fracturing fluid and wastewater. The U.S. House of
Representatives' Committee on Energy and Commerce Minority Staff released a report in 2011
noting that more than 650 products (i.e., chemical mixtures) used in hydraulic fracturing contain 29
chemicals that are either known or possible human carcinogens or are currently regulated under
the Safe Drinking Water Act (House of Representatives. 20111. More recently, several studies have
performed a reconnaissance of toxicity and/or physicochemical property data for specific subsets
of chemicals used in hydraulic fracturing fluids fElliottetal.. 2016: Wattenberg et al.. 2015:
Stringfellow et al.. 2014: Colborn etal.. 2011). and have provided discussion on the hazards
inherent to these chemicals. In all cases, authors reported toxicity data was not available for many
of the chemicals assessed in these studies, with some studies indicating significant data gaps. For
instance, Wattenberg etal. f20151 evaluated 168 chemicals commonly used in hydraulic fracturing
fluids in North Dakota, and reported that 59% did not have chronic toxicity data available, and 35%
did not have acute toxicity data available. Elliott etal. (2016) performed a systematic evaluation of
reproductive and developmental toxicity for 1021 chemicals used in hydraulic fracturing fluids or
detected in wastewater, and found this toxicity information was lacking for 76% of these chemicals.
Overall, while combined evidence suggests hydraulic fracturing has the potential to impact human
health via contamination of drinking water resources, the actual public health impacts are not well
understood and not well documented. Available information indicates there are many chemicals
within the hydraulic fracturing water cycle that are known to be hazardous to human health, as
well as hundreds of chemicals for which toxicological data is limited or unavailable.
In this chapter, our primary goal is to evaluate the availability of toxicity data for a list of chemicals
used in hydraulic fracturing fluids or present in produced water, focusing primarily on toxicity
values from sources that meet the criteria for inclusion in this assessment, and to highlight
chemicals that may pose human health hazards.
9.3 Identification of Chemicals Associated with the Hydraulic Fracturing Water
Cycle
As the initial step towards evaluating the hazards of chemicals in the hydraulic fracturing water
cycle, the EPA compiled a list of chemicals used in or released by hydraulic fracturing operations
across the country.1 This section describes the compilation of that list This consolidated list
includes a total of 1,606 chemicals, and can be broken down into two sublists: (1) a list of chemicals
used in hydraulic fracturing fluids, and (2) a list of chemicals detected in produced water from
hydraulically fractured wells (Text Box 9-2).
This list demonstrates the range and variety of chemicals that are associated with the hydraulic
fracturing industry. These chemicals should not be considered unique to the hydraulic fracturing
1 We use the word "chemical" to refer to any individual chemical or chemical substance that has been assigned a CASRN
(Chemical Abstracts Service Registry Number]. A CASRN is a unique identifier for a chemical substance, which can be a
single chemical (e.g., hydrochloric acid, CASRN 7647-01-0] or a mixture of chemicals (e.g., hydrotreated light petroleum
distillates (CASRN 64742-47-8], a complex mixtures of C9 to C16 hydrocarbons]. For simplicity, we refer to both pure
chemicals and chemical substances that are mixtures, which have a single CASRN, as "chemicals."
9-8
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
industry; many of the chemicals used in hydraulic fracturing fluids are widely used industrial
chemicals, and many of the chemicals in produced water are naturally occurring. Although this list
represents the best information available to the EPA at the time of the assessment, it should not be
considered comprehensive. It is likely that, as industry practices change, chemicals may be used or
detected that are not included on these lists. Some additional limitations to this chemical list are
described in the subsections below.
Text Box 9-2. The EPA's List of Chemicals Identified in Hydraulic Fracturing Fluids and/or
Produced Water.
This chemical list progressed through multiple iterations as the EPA's hydraulic fracturing study was
developed, culminating in the list of 1,606 chemicals presented in this report.
The first iteration of this chemical list was published in the interim progress report (U.S. EPA. 2012hl and
included 1,026 chemicals that were identified from ten sources of information. Seven of these information
sources were documents from federal and state governmental units—including the EPA (U.S. EPA. 2011a. e,
2004a: Material Safety Data Sheets 1 the U.S. House of Representatives (House of Representatives. 20111 the
New York State Department of Environmental Conservation (NYSDEC. 20111 and the Pennsylvania
Department of Environmental Protection fPA DEP. 2010a1—which obtained data directly from industry. This
includes a list of chemicals provided directly to the EPA by nine well operating companies, representing
chemicals used in hydraulic fracturing fluids between 2005 and 2009, and a list of chemicals detected by
these companies in produced water from 81 wells. The remaining three sources are as follows: a technical
report prepared by the Gas Technology Institute for the Marcellus Shale Coalition, which is a drilling industry
trade group (Hayes. 20091 a peer-reviewed journal article by Colborn et al. (20111 and the FracFocus
Chemical Disclosure Registry, which is a national hydraulic fracturing chemical registry developed by the
Ground Water Protection Council and the Interstate Oil and Gas Compact Commission fGWPC. 20121
In the external review draft of the EPA's hydraulic fracturing study report (U.S. EPA. 2015dl this chemical list
was updated to 1,173 chemicals. The updated chemical list includes the 1,026 chemicals published in the
progress report, along with additional chemicals that were identified in the EPA FracFocus 1.0 report fU.S.
EPA. 2015al
For the final version of this assessment, the list has again been updated to include additional chemicals in
produced water, which were identified from 18 additional literature sources. The final list includes a total of
1,606 chemicals that have been reported as used in hydraulic fracturing fluids or detected in produced water.
The complete list of sources used to compile the final chemical list is provided in Appendix Table H-l. To the
extent possible, after chemicals were identified from the sources in Table H-l, the EPA verified the identity of
the chemicals used in hydraulic fracturing fluids and detected in produced water of hydraulically fractured
wells as described in Appendix Section H.l.
9.3.1 Chemicals Used in Hydraulic Fracturing Fluids
Of the 1,606 total chemicals, the EPA identified 1,084 chemicals as being used in hydraulic
fracturing fluids. This list was originally introduced in Chapter 5 of this assessment (Section 5.4),
which describes some of the chemical classes and their purpose, and identifies the most frequently
used chemicals. This list of 1,084 chemicals is shown in Appendix Table H-2.
9-9
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Although a total of 8 sources were used to identify the list of chemicals used in hydraulic fracturing
fluids, only one source—the EPA analyses based on disclosures submitted to FracFocus—had
sufficient information for estimating the frequency with which these chemicals were used (Section
5.4, Text Box 5-1).1 Of the 1,084 chemicals, 688 were identified in the EPA FracFocus 1.0 report
(U.S. EPA. 2015a).2 Frequency of use for individual chemicals ranged from low (480 chemicals on
the list were reported in less than 1% of disclosures nationally) to very high (methanol was
reported in 73% of disclosures nationally).
As discussed in Chapter 5, this list provides valuable information on the chemicals used in hydraulic
fracturing fluids, but should not be considered complete. For example, in the analysis of the
disclosures submitted to the FracFocus 1.0 registry, the EPA was only able to assign standardized
chemical names to 65% of ingredient records. The remaining 35% of ingredient records did not
have valid CASRNs and were excluded from the analysis because they could not be assigned a
standardized chemical name (U.S. EPA. 2015a). In a more recent analysis of data reported to the
FracFocus registry through April 2015, Konschnik and Davalu (2016) found that 80% of chemicals
had valid CASRN. That analysis identified an additional 263 CASRNs that are not on the EPA's list of
chemicals used in hydraulic fracturing fluids (Davalu and Konschnik. 2016).
Industry use of CBI is another factor that likely limits the completeness of this chemical list and
introduces uncertainty. For example, companies submitting to FracFocus 1.0 were not required to
disclose chemicals claimed as CBI. EPA determined that approximately 70% of the disclosures
submitted to FracFocus 1.0 contain at least one CBI chemical, and for those disclosures, the average
number of CBI chemicals per disclosure was five. Overall, 11% of ingredients were reported to
FracFocus 1.0 as CBI (U.S. EPA. 2015a). Konschnik and Davalu (2016) report a 5.6% increase in the
number of CBI ingredients, as well as an increase in the number of disclosures reporting the use of
at least one CBI ingredient (Section 5.4; Text Box 5-2).
Although FracFocus disclosures do not provide the name or CASRN of CBI chemicals, the chemical
family is sometimes provided. The EPA determined that 79% of CBI ingredient records submitted
to FracFocus 1.0 had enough information to partially define the chemical and assign it to a chemical
family fU.S. EPA. 2015al. This resulted in the designation of 448 standardized chemical families to
which these chemicals could be assigned. The most common standardized chemical families for CBI
ingredients were oxyalkylated alcohol (4.7% of CBI ingredient records), petroleum distillates (4.0%
of CBI ingredient records), and quaternary ammonium compounds (3.6% of CBI ingredient
records) fU.S. EPA. 2015al (Appendix Table B-l). These standardized chemical family designations
are not discussed further in this chapter, but may be useful for site-specific risk assessment, as they
1 The FracFocus frequency of use data presented in this chapter is based on 35,957 FracFocus disclosures that were
deduplicated, within the study time period (January 1,2011 to February 28,2013], and with ingredients that have a valid
CASRN. In the interest of including as many chemicals as possible, this analysis includes chemicals that do not have valid
concentration data. The 692 chemicals includes 16 chemicals that are listed as being used as proppants.
2 EPA analyses based on disclosures submitted to FracFocus identified 692 unique CASRN. Of these 692, we determined
that 4 chemicals are listed under two different CASRN (indicated in the footnote of Appendix Table H-2]. Frequency of use
data is therefore available for 688 chemicals that were included on EPA's list of chemicals in hydraulic fracturing fluids.
9-10
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
may provide insight into potential physicochemical properties and toxicity of CBI chemicals used at
a particular site.
9.3.2 Chemicals Detected in Produced Water
Of the 1,606 total chemicals, the EPA identified 599 as having been detected in produced water.
Included among these chemicals are naturally occurring organic compounds, metals, radionuclides,
industrial chemicals, and pesticides. These chemicals were originally introduced in Chapter 7 of this
assessment, and were compiled from a total of 21 sources. Seventy-seven of the total 599 chemicals
in produced water were also identified by at least one of the sources in Appendix H as being used in
hydraulic fracturing fluid. However, the EPA used different sets of sources to identify chemicals
used in hydraulic fracturing fluids versus those detected in produced water, and there is not a
matched comparison between the chemicals used in hydraulic fracturing fluids and returned in
produced water at each particular well. Therefore, it is difficult to draw direct comparisons
between these two chemical lists, or to use these lists to draw conclusions on the persistence of
chemicals in produced water from hydraulically fractured wells. The list of 599 chemicals identified
in produced water is shown in Appendix Table H-4.
Although this list provides useful information on the chemical composition of produced water, it is
not likely that the data sources were able to capture all of the chemicals present. Chemicals and
their metabolites may go undetected in produced water because they were not targeted in the
analytical protocols, they were below the limit of detection, or because no standard analytical
method exists. Additionally, as discussed in Chapter 7, the composition and concentration of
chemicals in produced water will differ depending upon factors like the geology of the formation,
the chemicals used for hydraulic fracturing, and the amount of time that has elapsed since hydraulic
fracturing. There is therefore expected to be a high degree of local and temporal variation in these
chemicals, and there was not sufficient information to determine the frequency with which these
chemicals were detected on a national basis.
Concentration data in produced water are available for 175 of these 599 chemicals (Appendix E),
including inorganic contributors to salinity (Appendix Tables E-4 and E-5), metals (Appendix
Tables E-6 and E-7), radioactive constituents (Appendix Table E-8), and organic constituents
(Appendix Tables E-9, E-ll, E-12, and E-13). The remaining chemicals were detected in produced
water, but concentration was not reported. For these chemicals with concentration data, the
measured concentrations spanned several orders of magnitude. For instance, for organic chemicals
in produced water from the Marcellus shale formation (Appendix Table E-ll), average or median
measured concentrations ranged from 2.7 |ig/L for N-nitrosodiphenylamine to 400 |ig/L for carbon
disulfide.
9.4 Toxicological and Physicochemical Properties of Hydraulic Fracturing
Chemicals
As the next step towards evaluating the hazards of chemicals in the hydraulic fracturing water
cycle, toxicological and physicochemical data were collected as available for each of the chemicals
identified in Appendix H. This section describes the compilation of these data, and discusses the
9-11
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
extent to which toxicological and physicochemical property data are available for this list of
chemicals.
The primary focus of the toxicological analysis in this chapter is on the availability of chronic oral
RfVs and OSFs from sources that met the EPA's criteria for inclusion in this study. Qualitative cancer
classifications were also identified from these sources when available. This is not intended to be an
exhaustive compilation of toxicity values for this chemical list. Rather, it is intended to be a
reconnaissance of high-quality toxicological information that met the EPA's criteria for inclusion in
this study. If a source of RfVs, OSFs, or qualitative cancer classifications was not included here, that
only means that it did not meet the criteria for the purposes of the EPA's study, which are described
in this chapter in Section 9.4.1.
Section 9.4.1 describes the criteria used to identify and select RfVs, OSFs, and qualitative cancer
classifications, and describes the availability of these toxicological data for the chemicals on the
EPA's list of hydraulic fracturing-related chemicals. The next two sections describe additional
sources of toxicological information, which may be useful for hazard evaluation when chronic oral
RfVs and OSFs are not available: Section 9.4.2 describes the use of a QSAR model to estimate
chronic oral toxicity, and Section 9.4.3 describes the availability of additional toxicological
information on the EPA's ACToR database. Section 9.4.4 describes other available software tools
and approaches that may be used by stakeholders for site-specific risk assessment, but are not
utilized in this report. Section 9.4.5 discusses the methods used in this report to generate
physicochemical property data, and presents the availability of physicochemical property data for
the chemicals on the EPA's list A brief overview of the toxicity values discussed in the chapter is
presented in Text Box 9-3.
As a resource that can be used to support risk assessment at hydraulic fracturing sites, all of the
selected RfVs, OSFs, qualitative cancer classifications, QSAR-based toxicity estimates, and
physicochemical property data described in this chapter will be compiled into an electronic
database that will be publicly accessible via the EPA's website. Additionally, the EPA's compilation
of toxicity data for this chemical list has been discussed in two recent manuscripts, both of which
focused on the list of 1,173 chemicals that was presented in the external review draft of the EPA's
hydraulic fracturing study report (U.S. EPA. 2015d). Yostetal. (2016b) describes the compilation of
RfVs and OSFs for the list of 1,173 chemicals. Yostetal. (2016a) describes the use of a QSAR model
to estimate toxicity for the list of 1,173 chemicals.
Text Box 9-3. Toxicity Values for Hydraulic Fracturing-Related Chemicals.
Here we provide a brief description of the toxicity values that are presented in this chapter, and how they
should be interpreted and used to evaluate chemical hazards. Formal definitions of these terms are footnoted
in the chapter and can also be found in the glossary (Appendix J).
Reference value (RfV): RfVs are health-protective values, which describe the dose of a chemical that is likely
to be without an appreciable risk of adverse health effects. In general, lower RfVs indicate greater toxicity;
however, comparison of RfVs among a set of chemicals requires careful consideration. RfVs are developed by
considering the full database of epidemiological and experimental studies available for a particular chemical.
(Text Box 9-3 is continued on the following page.)
9-12
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Text Box 9-3 (continued). Toxicity Values for Hydraulic Fracturing-Related Chemicals.
These data are used to identify the critical effect, which is the first adverse effect, or its known precursor, that
occurs as the dose rate increases (U.S. EPA. 2011fl. An RfV is then derived by starting with a quantitative
point of departure (POD), which is the toxicological dose-response point that marks the beginning of a low-
dose extrapolation for the critical effect, and applying uncertainty factors (UFs) to derive a value for the
protection of human health. UFs are applied to account for 5 areas of uncertainty: (1) intraspecies variability;
(2) interspecies uncertainty; (3) extrapolation from a subchronic study; (4) extrapolating from a no-
observed-adverse-effect level (NOAEL); and (5) deficiencies in the database. A UF of 1, 3, or 10 can be applied
for any of these areas of uncertainty depending upon the amount and/or type data available, up to a
maximum total UF of 3,000 (U.S. EPA. 20021 Thus, a chemical with a low RfV may reflect high uncertainty in
the value, and not necessarily the toxicity of the chemical. Chemicals with a lower total UF generally have
more reliable and robust health effect information.
Oral slope factor (OSF): An OSF is a measure of the increased cancer risk from a lifetime oral exposure to an
agent. Higher OSFs indicate greater carcinogenic potency. As with RfVs, OSFs are developed by considering
the full database of epidemiological and experimental studies for a particular chemical, and evaluating the
increase in cancer incidence as dose rate increases. OSFs should be considered in conjunction with qualitative
cancer classifications, which characterize the weight of evidence regarding the agent's potential to cause
cancer in humans.
No-observed-adverse-effect level (NOAEL): NOAEL is defined as the highest exposure level at which there
are no biologically significant increases in the frequency or severity of adverse effect between the exposed
population and its appropriate control; some effects may be produced at this level, but they are not
considered adverse or precursors of adverse effects (U.S. EPA. 2011fl.
Lowest-observed-adverse-effect level (LOAEL): LOAEL is defined as the lowest exposure level at which
there are biologically significant increases in the frequency or severity of adverse effects between the
exposed population and its appropriate control group (U.S. EPA. 2011fl. Lower LOAELs indicate greater
toxicity.
Maximum contaminant level (MCL): MCLs are the highest level of a contaminant that is allowed in drinking
water. MCLs are set as close to the maximum contaminant level goal (MCLG) as feasible using the best
available analytical and treatment technologies and taking cost into consideration. MCLs are enforceable
standards. The MCLG is a non-enforceable health benchmark goal which is set at a level at which no known or
anticipated adverse effect on the health of persons is expected to occur and which allows an adequate margin
of safety (U.S. EPA. 2012b). Whereas RfVs, LOAELs, and NOAELs are expressed in terms of dose (mg/kg-day),
MCLs are expressed in terms of the concentration of an agent in water (|ig/L).
9.4.1 Reference Values (RfVs), Oral Slope Factors (OSFs), and Qualitative Cancer
Classifications
For the purpose of this study, the EPA's primary goal was to identify high quality toxicity values
that met the criteria for inclusion in this study, and that could be used by stakeholders to support
the risk assessment of hydraulic fracturing chemicals (Text Box 9-1). Briefly, the sources of RfVs,
OSFs, and qualitative cancer classifications selected by the EPA for the purposes of this chapter met
the following key criteria:
9-13
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
1. The body or organization generating or producing the peer-reviewed RfVs, peer-reviewed
OSFs, or peer-reviewed qualitative assessment must be a governmental or
intergovernmental body.
2. The data source must include peer-reviewed RfVs, peer-reviewed OSFs, or peer reviewed
qualitative assessments.
3. The RfVs, OSFs, or qualitative assessments must be based on peer-reviewed scientific data.
4. The RfVs, OSFs, or qualitative assessments must be focused on protection of the general
public.
5. The body generating the RfVs, OSFs, or qualitative assessments must be free of conflicts of
interest with respect to the chemicals for which it derives reference values or qualitative
assessments.
These five criteria were developed by the EPA specifically for the purpose of this assessment, and
are similar to the EPA Office of Solid Waste and Emergency Response (OSWER) recommendations
for selecting toxicity values in conducting site-specific risk assessments (Regional Tier 3 Toxicity
Value Workgroup. 2013: U.S. EPA. 2003a. 19891.1 The OSWER directives provide recommendations
on the appropriate sources of toxicity values and toxicological information that should be
considered in risk assessments, and were designed to recognize toxicity values that were developed
using the best available scientific information. In addition, these directives outline references to
various resources that provide guidance on the approaches and issues considered in deriving
toxicity values. This type of information can be especially important in cases in which multiple
sources of toxicity values need to be considered or evaluated, or in which a value needs to be
developed. More detail on these criteria for selection and inclusion of data sources, as well as the
full list of data sources that were considered for this study, are available in Appendix G.
Table 9-1 shows the data sources that met these five criteria for the selection of toxicological
information. The federal databases of RfVs or OSFs that met these criteria are the EPA's Integrated
Risk Information System (IRIS) database, the EPA's Provisional Peer-Reviewed Toxicity Value
(PPRTV) database, the EPA's Human Health Benchmarks for Pesticides (HHBP) database, and the
Agency for Toxic Substances and Disease Registry (ATSDR) database. IRIS and PPRTV also provide
qualitative cancer classifications. One state source of RfVs and OSFs, the California Environmental
Protection Agency (CalEPA) Toxicity Criteria Database, met the criteria for inclusion.2 One
intergovernmental source of RfVs, the World Health Organization (WHO) International Programme
on Chemical Safety (IPCS) Concise International Chemical Assessment Documents (CICAD), met the
criteria for inclusion. The International Agency for Research on Cancer (IARC) and U.S. National
Toxicology Program (NTP) Report on Carcinogens (RoC) also met the criteria and were used as
additional sources for qualitative cancer classifications.
1 OSWER changed its name to the Office of Land and Emergency Management (OLEM], effective December 15, 2015.
2 State RfVs and OSFs are also publicly available from Alabama, Texas, Hawaii, and Florida, but they did not meet the
criteria for consideration as sources for RfVs and OSFs in this report. See Appendix G for details.
9-14
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-1. Sources of selected RfVs, OSFs, and qualitative cancer classifications.
Type of toxicological
Information
Data source
Website
RfVs, OSFs, and
qualitative cancer
classifications
EPA Integrated Risk Information System (IRIS)
database
http://cfpub. eDa.gov/ncea/iris/index.cf
m?fuseaction=iris.showSubstanceList
RfVs, OSFs, and
qualitative cancer
classifications
EPA Provisional Peer-Reviewed Toxicity Value
(PPRTV) database
http://hhpprtv. ornl.gov/index. html
RfVs, OSFs
EPA Human Health Benchmarks for Pesticides
(HHBP) database
http://iaspub.epa.gov/apex/pesticides/
f?p=HHBP:home
RfVs
Agency for Toxic Substances and Disease
Registry (ATSDR) Minimum Risk Levels
http://www.atsd r. cdc.gov/toxprofi les/i
ndex.asp#bookmark05
RfVs, OSFs
California Environmental Protection Agency
(CalEPA) Toxicity Criteria Database
http://oehha.ca.gov/tcdb/index.asp
RfVs
World Health Organization (WHO)
International Programme on Chemical Safety
(IPCS) Concise International Chemical
Assessment Documents (CICAD)
http://www.who.int/ipcs/publications/
cicad/en/
Qualitative cancer
classifications
National Toxicology Program (NTP) 13th
Report on Carcinogens (RoC)
https://ntp.niehs.nih.gov/pubhealth/
roc/
Qualitative cancer
classifications
International Agency for Research on Cancer
(IARC) Monographs
http://monographs.iarc.fr/
In addition to the sources in Table 9-1, we also consulted the NPDWRs and DWSHA tables (U.S. EPA.
2014a) to determine whether the chemicals on this list are regulated as drinking water
contaminants. NPDWRs provide a list of MCLs, which are legally enforceable standards on the
concentration of a substance that is allowed in drinking water under the Safe Drinking Water Act. In
this chapter, MCL values are referenced as a means of comparison with reported concentration data
where appropriate, and are reported in Appendix G and are compiled on the EPA's electronic
database for the hydraulic fracturing study.
As noted above, this chapter focuses on the presentation and use of chronic RfVs. Chronic RfVs
account for the potential that chemical exposure may be continuous, in low concentration, and over
a longer duration. In the absence of reliable information on the potential duration of chemical
exposure, this is a conservative assumption for the protection of human health. Chronic RfVs are
also lower than less-than-chronic RfVs (e.g., acute, intermediate, or subchronic toxicity values), and
are therefore more health protective. For these reasons, chronic RfVs are generally preferred as the
default by risk assessors when conducting site-specific risk assessments (U.S. EPA. 1989) and when
developing regional screening levels (U.S. EPA. 2016b). In contrast, acute RfVs are more applicable
for single exposures and/or exposures of limited frequency to high concentration and shorter
9-15
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
durations (e.g., emergencies). However, the availability of less-than chronic RfVs are also presented
for the sake of completeness.
Some chemicals had chronic oral RfVs or OSFs available from more than one of the sources in Table
9-1. For these chemicals, we selected a single value for use in this chapter by applying a
modification of the EPA OSWER Directives 9285.7-53 and 9285.7-86 tiered hierarchy of toxicity
values fU.S. EPA. 2003a). A single RfV and/or OSF was selected from the sources in this order:
HHBP (pesticides only), IRIS, PPRTV, ATSDR, and then other available values. The RfVs considered
from these sources included chronic oral reference doses (RfDs) from the IRIS, PPRTV, and HHBP
programs; chronic oral minimal risk levels (MRLs) from ATSDR; oral maximum allowable daily
levels (MADLs) from CalEPA; and tolerable daily intakes (TDIs) from CICAD.1'2'3'4'5
Of the 1,606 chemicals identified by the EPA, 173 (11%) have federal, state, or international chronic
oral RfVs and/or OSFs from sources listed in Table 9-1. Chronic oral RfVs and/or OSFs from the
selected sources are lacking for the remaining 1,433 (89%) chemicals that the EPA has identified as
associated with hydraulic fracturing. All available chronic oral RfVs and OSFs from the sources
listed in Table 9-1 are tabulated in Appendix G. Chronic oral RfVs and OSFs for chemicals used in
hydraulic fracturing fluids are listed in Appendix Tables G-la through G-lc, and chronic oral RfVs
and OSFs for chemicals reported in hydraulic fracturing flowback or produced water are listed in
Appendix Tables G-2a through G-2c. The EPA's IRIS database was the most abundant source of
these toxicity values.
Overall, when chemicals in hydraulic fracturing fluid and chemicals in produced water are
considered separately, the availability of chronic RfVs and OSFs can be summarized as follows:
• For the 1,084 chemicals used in hydraulic fracturing fluid, chronic oral RfVs or OSFs from
at least one of the selected federal, state, and international sources were available for 98
chemicals (9%). From the US federal sources alone, chronic oral RfVs were available for 81
chemicals (7%), and OSFs were available for 15 chemicals (1%).
1 The OSWER hierarchy indicates that sources should be used in this order: IRIS, PPRTV, and then other values. In this
report, this hierarchy was followed, but HHBP values were used in lieu of an IRIS value for a few chemicals that are
pesticides.
2 An RfD is an estimate (with uncertainty spanning perhaps an order of magnitude] of a daily oral exposure to the human
population (including sensitive subgroups] that is likely to be without an appreciable risk of deleterious effects during a
lifetime. It can be derived from a NOAEL, LOAEL, or benchmark dose, with uncertainty factors generally applied to reflect
limitations of the data used. Generally used in the EPA's non-cancer health assessments fU.S. EPA. 2011FI. This estimate is
expressed in terms of mg/kg-day.
3 An MRL is an estimate of daily human exposure to a hazardous substance at or below which the substance is unlikely to
pose a measurable risk of harmful (adverse], non-cancerous effects. MRLs are calculated for a route of exposure
(inhalation or oral] over a specified time period (acute, intermediate, or chronic]. MRLs should not be used as predictors
of harmful (adverse] health effects (ATSDR. 2016], Chronic MRL: Duration of exposure is 365 days or longer. This
estimate is expressed in terms of mg/kg-day.
4 An MADL is the maximum allowable daily level of a reproductive toxicant at which the chemical would have no
observable adverse reproductive effect, assuming exposure at 1,000 times that level fOEHHA. 20121. This estimate is
expressed in terms of ng/day.
5 A TDI is an estimate of the intake of a substance, expressed on a body mass basis, to which an individual in a (sub]
population may be exposed daily over its lifetime without appreciable health risk ("WHO. 20151 This estimate is
expressed in terms of mg/kg-day.
9-16
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
• For the 599 chemicals reported in produced water, chronic oral RfVs or OSFs from at least
one of the selected federal, state, and international sources were available for 120
chemicals (20%). From the US federal sources alone, chronic oral RfVs were available for
97 chemicals (16%), and OSFs were available for 30 chemicals (5%).
In addition to these chronic values, some of the chemicals also have less-than-chronic oral RfVs
available from the sources listed in Table 9-1. Subchronic, acute, or intermediate oral RfVs were
identified for 103 chemicals on the consolidated list, including 60 chemicals used in hydraulic
fracturing fluid (Appendix Table G-ld), and 73 chemicals reported in produced water (Appendix
Table G-2d). The majority of these chemicals also had chronic oral RfVs available, although there
were 10 chemicals that had less-than-chronic oral RfVs but lacked a chronic oral RfV. All of these
less-than-chronic RfVs were found on the PPRTV, ATSDR, or HHBP databases. As stated above,
chronic values more protective of human health than less-than-chronic values, and are generally
preferred for risk assessment These less-than-chronic values are therefore not discussed further in
this report, but are provided in Appendix G as supporting information.
Of the 1,606 chemicals identified by EPA, 207 (13%) had a qualitative cancer classification available
from at least one of the sources listed in Table 9-1, which include IRIS, PPRTV, IARC, and RoC. These
classifications are based on the weight-of-evidence (WOE) that a chemical causes cancer in humans.
Of these 207 chemicals:
• 21 were reported by at least one source to be a known carcinogen in humans.
• 66 were reported by at least one source to be a probable or possible carcinogen in
humans. These chemicals have been demonstrated to be carcinogenic in animal models,
but have limited or insufficient data to adequately assess carcinogenicity in humans.
• 117 were reported to be not classifiable as to carcinogenicity in humans. These chemicals
have been evaluated by at least one of these sources for their potential to cause cancer, but
had inadequate evidence from human exposure and animal studies to assess carcinogenic
potential.
• 3 were reported as not likely to be a human carcinogen.
The complete list of chemicals with qualitative cancer classifications are shown in Appendix Table
G-le (chemicals in hydraulic fracturing fluids) and Appendix Table G-2e (chemicals in produced
water).
9.4.2 Estimating Toxicity Using Quantitative Structure Activity Relationship (QSAR)
Modeling
Because the majority of chemicals identified in this report do not have RfVs and/or OSFs from the
selected sources, it is likely that risk assessors at the local and regional level may turn to alternative
sources of toxicological information. One potential resource is QSAR modeling software, which is
able to provide estimates or predictions of toxicity based on chemical structure. A key advantage to
QSAR models is that they are able to rapidly and inexpensively estimate toxicity values for
chemicals. A disadvantage is that QSAR estimates may be of higher uncertainty and less reliable
9-17
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
than values generated using traditional toxicological methods. However, because they increase the
available pool of toxicity information, QSAR estimates may be a useful resource for risk assessors
that are faced with evaluating potential exposures to data-poor chemicals.
A recent study by Yostetal. (2016a) used TOPKAT (Toxicity Prediction by Komputer Assisted
Technology) QSAR software to estimate toxicity for the EPA's list of chemicals used in hydraulic
fracturing fluids or detected in produced water, and evaluated how effectively these toxicity
estimates could be used to rank chemicals based on toxicity. The chemical list examined in this
study is the list of 1,173 chemicals published in the external review draft of the EPA's hydraulic
fracturing study report (U.S. EPA. 2015d) (Text Box 9-2), so the full list of 1,606 chemicals was not
assessed using the QSAR model. TOPKAT is commercially available QSAR software that is able to
estimate the rat chronic oral lowest-observed-adverse-effect level (LOAEL), which is the LOAEL
measured in a rat model following chronic oral exposure to a chemical.1
The authors of this study used TOPKAT to generate rat chronic oral LOAEL estimates for EPA's list
of chemicals, and assigned qualitative confidence scores (high, medium, or low) to each estimate
based on parameters reported by the model. The authors then examined a list of 48 chemicals that
had both a high-confidence TOPKAT LOAEL estimate and a chronic oral reference dose (RfD) from
EPA's IRIS database. The authors ranked these 48 chemicals from most toxic to least toxic based on
either TOPKAT LOAEL estimate or on IRIS chronic oral RfD, and then used Spearman rank
correlation to examine the similarity between these chemical rankings.
Of the 1,173 hydraulic fracturing chemicals, TOPKAT was able to generate toxicity estimates for
515 (44%) of the chemicals, including 453 chemicals that are used in hydraulic fracturing fluids,
and 86 chemicals that have been detected in produced water. The authors found a strong and
statistically significant correlation between chemical rankings based on high-confidence TOPKAT
LOAEL estimates and on IRIS chronic oral RfDs, indicating that high-confidence TOPKAT LOAEL
estimates can effectively be used to rank chemicals based on toxicity when experimentally derived
toxicity values are not available. Overall, TOPKAT LOAEL estimates were available for 417
chemicals in this study that lack chronic oral RfVs or OSFs from the sources identified by EPA. Of
these, 389 were found to be high-confidence estimates.
When available, the high-confidence TOPKAT LOAEL estimates from Yostetal. f2016bl are
discussed in this chapter as an additional resource that can be used to rank chemicals based on
toxicity. Low- or medium-confidence TOPKAT LOAEL estimates are not shown in this chapter, as
the use of these values for chemical ranking has not been validated.
1 LOAEL is defined as the lowest exposure level at which there are biologically significant increases in the frequency or
severity of adverse effects between the exposed population and its appropriate control group following chronic (lifetime]
exposure.
9-18
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.4.3 Chemical Data Available from EPA's Aggregated Computations Toxicology Resource
(ACToR) Database
An additional tool for obtaining information focused on toxicology and risk assessment is the EPA's
ACToR database.1 ACToR is a large data warehouse developed by the EPA to consolidate large and
disparate amounts of public data on chemicals, including data on chemical identity, structure,
physicochemical properties, in vitro assay results, and in vitro toxicology data. The primary goals of
ACToR are to make information on chemical health effects and exposure potential readily
accessible, to characterize chemical toxicological data gaps, and to provide a resource for model
building to address data gaps in environmental risk information Hudson etal.. 20121.
ACToR contains data on over 500,000 chemicals from over 2,500 data sources, covering many
domains including hazard, exposure, risk assessment, risk management, and use. Data sources and
collections in ACToR include the US EPA, National Institutes of Health (NIH), the Centers for Disease
Control and Prevention (CDC), US Food and Drug Administration (FDA), State Agencies, the
European Chemicals Agency (ECHA), corresponding government agencies in Canada (e.g., Health
Canada), Europe and Japan, the World Health Organization (WHO), and non-governmental
organizations (NGOs). Data within ACToR ranges from the federal RfVs and OSFs discussed in
Section 9.3.1, which have undergone extensive peer review, to other toxicity values and study and
test results that have undergone little to no peer review.
ACToR organizes these data into several levels of "assays" and "assay categories," which serve to
classify data sets according to the nature of the data. For instance, the "Hazard" assay category
includes all data that are associated directly or indirectly with toxicology experiments. The "Risk
Management" assay category includes regulatory and non-regulatory risk management
benchmarks. Considering the diversity and overlapping nature of the data resources within ACToR,
a single data set may fall into multiple assay categories (Tudson etal.. 2012).
We searched the ACToR database for information related to the list of 1,606 hydraulic fracturing-
related chemicals. Specifically, we searched within the "Hazard" and "Risk Management" assay
categories of ACToR. Results of the query were then filtered to include the assays that are most
relevant to chemical exposure via drinking water. These assays were assigned into the following
nine data classes: carcinogenicity, dose response values, drinking water criteria, genotoxicity or
mutagenicity, hazard identification, LOAEL/NOAEL, RfV, OSF, and water quality criteria.2
Of the 1,606 chemicals, it was found that 735 (46%) have some data available within these data
classes on ACToR, with the total number of data points found for individual chemicals ranging from
1 to 243. Figure 9-2 shows the percentage of the total 1,606 chemicals that had data available in
each of the nine ACToR data classes, and indicates the fraction of those chemicals that also had a
chronic oral RfV or OSF available from at least one of the selected sources in Table 9-1. As can be
seen in Figure 9-2, 37% of the chemicals had some information on hazard identification, 25% had
1 The ACToR database, including the full list of data collections and assays, is available at: http://actor.epa.gov.
2 NOAEL is defined as the highest exposure level at which there are no biologically significant increases in the frequency
or severity of adverse effect between the exposed population and its appropriate control; some effects may be produced
at this level, but they are not considered adverse or precursors of adverse effects. Source: U.S. EPA f2011fl.
9-19
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
information on carcinogenicity, and 24% had a LOAEL or NOAEL identified. A LOAEL and/or
NOAEL identified from a well conducted dose-response study are often considered the minimum
data needed for RfV derivation (U.S. EPA. 2002).
Focusing on the 1,433 chemicals that lacked a chronic RfV and/or OSF from the selected sources
described in Section 9.3.1, 567 (40%) had available data within at least one of these data classes on
ACToR. Thus, ACToR has a significant amount of potentially useful data on chemical hazards,
including for some data-poor chemicals, and might help to fill data gaps in the ongoing effort to
understand potential hazards of hydraulic fracturing chemicals.
It is outside the scope of this assessment to evaluate the quality and reliability of data within ACToR
that has not already undergone peer review. Therefore, with the exception of data from the sources
listed in Table 9-1, data from ACToR was not considered in the hazard evaluation presented in this
chapter. However, as a potential resource for risk assessors, the tables in this chapter indicate
whether a chemical had data available on ACToR.
Hazard Identification
Carcinogenicity
LOAEL or NOAEL
Genotoxicity or Mutagenicity
Water Quality Criteria
RfV (Oral or Inhalation)
Cancer Slope Factor (Oral or Inhalation)
Drinking Water Criteria
Dose Response Values
5 10 15 20 25 30 35 40
Percentage of hydraulic fracturirig-related chemicals with data on ACToR
I Has RfV or OSF from selected sources (Table 9-1)
Doesn't have RfV or OSF from selected sources
Figure 9-2. Percentage of hydraulic fracturing-related chemicals (out of 1,606 total) with at
least one data point in each ACToR data class.
9.4.4 Additional Tools for Hazard Evaluation
In addition to the methods and approaches utilized in this chapter, there are other potential tools
and approaches that could be used by stakeholders to prioritize and estimate toxicity of chemicals
that have a limited toxicity database. We briefly describe three such approaches in Appendix G
(Section G.4): the Threshold of Toxicological Concern (TTC) approach, the Organisation for
9-20
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Economic Co-operation and Development (OECD) QSAR Toolbox, and the application of data from
high throughput screening (HTS) assays. Toxicity predictions from these additional data sources
can be either quantitative or qualitative, and may be used to fill and address gaps related to risk
assessment.
Although these additional tools may be potentially useful for the evaluation of chemical hazards,
they currently have limited utility in this chapter, and are not discussed further. The TTC approach
requires an estimate of human intake, which is challenging for hydraulic fracturing-related
chemicals, since the potential for human exposure is generally not clear. The OECD QSAR Toolbox is
potentially useful for qualitative assessment, and may be useful for quantitative toxicity assessment
as its human health hazard and repeated dose toxicity databases expand. HTS assays are an
emerging technology, and the potential application of these data for human health risk assessment
is not well understood. These tools would be more appropriately applied by stakeholders on a site-
specific basis, as preliminary steps to identify potential chemicals of concern.
9.4.5 Physicochemical Properties
As presented in Chapter 5, EPI Suite™ software was used to generate data on the physicochemical
properties of the hydraulic fracturing-related chemicals identified by EPA. EPI Suite provides an
estimation of physicochemical properties based upon chemical structure, and will additionally
provide experimentally measured values for these properties when they are available for a given
chemical. For more details on this software and on the use of physicochemical properties for fate
and transport estimation, see Chapter 5.
From the total list of 1,606 chemicals associated with hydraulic fracturing, EPI Suite was able to
generate data on physicochemical properties for 917 (57%) of the chemicals (Appendix H). This
includes 455 chemicals that are reported in hydraulic fracturing fluids, 521 chemicals that have
been reported in produced water, and 59 chemicals that were both used in hydraulic fracturing
fluids and reported in produced water. The remaining 689 chemicals on EPA's total list lacked the
structural information necessary to generate estimates.
In addition to EPI Suite, two other software programs were consulted to generate physicochemical
property data for EPA's list of hydraulic fracturing-related chemicals. QikProp (Schrodinger.
20121 and LeadScope (Leadscope Inc.. 20121 are commercial products designed primarily as drug
development and screening tools, which are able to estimate properties related to chemical fate and
transport as well as pharmacokinetics. Properties generated by QikProp and LeadScope are
generally more relevant to drug development than to environmental assessment The properties
generated by QikProp and LeadScope were not used in the analysis presented in this report, but
will be compiled on the electronic database for EPA's hydraulic fracturing study.
9.4.6 Summary of Available Toxicological and Physicochemical Information for Hydraulic
Fracturing Chemicals
Figure 9-3 summarizes the toxicological and physicochemical information that is available for the
list of hydraulic fracturing chemicals identified by EPA in this study. This figure also summarizes
9-21
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
the availability of data on the occurrence of these chemicals in hydraulic fracturing fluids
(frequency of use) or in produced water (measured concentrations).
Availability of toxicological, physicochemical, and occurrence data
for 1,606 chemicals
1,084 chemicals used in
hydraulic fracturing fluids
599 chemicals detected in
produced water
98 chemicals (9%)
J
RfVs and OSFs
120 chemicals (20%)
454 chemicals (42%)
f ">
TOPKAT LOAELs
119 chemicals (20%)
550 chemicals (51%)
ACToR database
^ -J
259 chemicals (43%)
455 chemicals (42%)
EPI Suite
521 chemicals (87%)
688 chemicals (63%) with
frequency of use data
(FracFocus 1.0)
Occurrence data
175 chemicals (29%) with
measured concentration
data
Figure 9-3. Overall representation of the selected toxicological, physicochemical, and
occurrence data available for the 1,606 hydraulic fracturing-related chemicals identified by
the EPA.
Overall, there is a clear paucity of chronic oral RfVs and OSFs for this list of chemicals, indicating
that the majority of chemicals associated with hydraulic fracturing activity have not undergone
significant toxicological assessment. QSAR-based toxicity estimates (TOPKAT LOAELs) were
availabl e for a larger number of these chemicals, and were often available for chemicals that lack
chronic oral RfVs and OSFs. EPA's ACToR database offers additional toxicological data that may be
useful for the hazard evaluation of these chemicals, although the quality and reliability of the data
for these chemicals within ACToR was not evaluated here.
9.5 Hazard Identification of Hydraulic Fracturing Chemicals
This section focuses on the hazard identification of subsets of chemicals that were identified as
being of particular interest in previous chapters of this report, or which otherwise may be of
particular interest to risk assessors. For these chemicals, we summarize what is known about
events that may lead to the entry of these chemicals into drinking water resources. We provide
examples of recent studies that have reported these chemicals in drinking water resources,
9-22
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
including examples in which these chemicals have been reported at concentrations exceeding
MCLs. We then summarize the available toxicological data for these chemicals, including chronic
oral RfVs, OSFs, cancer classifications, QSAR-based toxicity estimates (TOPKAT LOAELs), and the
availability of relevant toxicological information from EPA's ACToR database—and indicate which
chemicals are regulated by EPA as drinking water contaminates.
We focused on the following subsets of chemicals:
I. Chemicals used in hydraulic fracturing fluids (Chapter 5)
6. Organic chemicals that may be returned to the surface in produced water, including
naturally occurring hydrocarbons such as BTEX (Chapter 7)
7. Inorganic chemicals that may be returned to the surface in produced water, including
metals, inorganic ions, and technologically enhanced naturally occurring radioactive
material (TENORM) (Chapter 7)
8. Methane in stray gas, which has been reported in drinking water resources in areas of
hydraulic fracturing activity (Chapter 6)
9. Disinfection byproducts (DBPs) that may be formed from wastewater constituents
(Chapter 8)
10. Banned chemicals reported in produced water, specifically organochlorine pesticides and
polychlorinated biphenyls (PCBs).
II. Chemicals on EPA's consolidated list that were reported in both hydraulic fracturing fluids
and produced water
The hazard identification for these subsets of chemicals is presented below.
9.5.1 Chemicals Used in Hydraulic Fracturing Fluids
Chapter 5 provided an overview of chemicals that are used in hydraulic fracturing fluids. These
chemicals have the potential to enter drinking water resources through events such as spills of
hydraulic fracturing fluids, injection of hydraulic fracturing fluids directly into groundwater, and
leakoff of fluids into the formation. These chemicals may also persist in produced water, and may
enter drinking water resources through spills or releases of produced water or inadequately
treated wastewater.
Several recent field studies have detected chemicals that are commonly used in hydraulic fracturing
fluids in groundwater near hydraulically fractured wells. In some cases, the origin of the chemicals
could be clearly linked to hydraulic fracturing activity. For example, in Killdeer, North Dakota
(Section 6.2.2.1), evidence strongly suggests a well blowout during hydraulic fracturing led to the
contamination of a drinking water aquifer with tert-butyl alcohol, a degradation product of tert-
butyl hydroperoxide used in hydraulic fracturing fluids at that site (U.S. EPA. 2015il. In
groundwater monitoring wells in the Pavillion Field in Wyoming, Digiulio and Tackson (20161
reported detections of organic chemicals used in hydraulic fracturing fluids at that site, including 2-
9-23
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
butoxyethanol, naphthalene, 1,2,4-trimethylbenzene, diethylene glycol, methanol, ethanol, and
isopropanol, likely as a result of shallow hydraulic fracturing in that region.
Other studies provide indirect evidence that chemical contaminants originated from hydraulic
fracturing activity. For example, in the Marcellus Shale in Pennsylvania, Llewellyn et al. (2015)
detected trace levels of 2-butoxyethanol in water wells near several hydraulically fractured wells,
with multiple lines of evidence suggesting that the chemical originated from a surface spill or leak
related to hydraulic fracturing activity. In northeastern Pennsylvania, Drollette etal. f20151 found
trace concentrations of known constituents of hydraulic fracturing fluid in private residential
groundwater wells, including di(2-ethylhexyl) phthalate, with evidence suggesting that the
chemicals originated from known surface spills of hydraulic fracturing fluids. In the Barnett Shale,
Texas, a survey of water quality in public and residential wells reported chemicals that are known
to be used in hydraulic fracturing fluids, including methanol, ethanol, isopropanol, and propargyl
alcohol, but it was not clear whether these chemicals originated from hydraulic fracturing activity
or from other potential sources fHildenbrand et al.. 20151.
Table 9-2 shows the list of chemicals that were reported in at least 10% of disclosures nationally in
the EPA FracFocus 1.0 project database (excluding water, quartz, and sodium chloride), and shows
the noncancer toxicity data (chronic oral RfVs and TOPKAT LOAEL estimates) and ACToR data
available for these chemicals.1 Cancer information is provided in Table 9-3. Nine (26%) of these 34
chemicals have a chronic oral RfV available from at least one of the sources in Table 9-1. Chronic
oral RfVs ranged from 0.002 mg/kg-day (propargyl alcohol) to 2 mg/kg-day (methanol and
ethylene glycol). Critical effects for these chemicals include kidney/renal toxicity, hepatotoxicity,
developmental toxicity (extra cervical ribs), reproductive toxicity, neurotoxicity, and decreased
terminal body weight Only one of these chemicals, sodium chlorite, is regulated in drinking water
under the NPDWRs.
Of the 25 chemicals that lack chronic oral RfVs, 11 have high-confidence TOPKAT LOAEL estimates
available. Of these, methenamine (~14% of disclosures) had the lowest TOPKAT LOAEL estimate,
and choline chloride (~15% of disclosures) had the second lowest. All but five of these chemicals
had at least some relevant toxicological data available on EPA's ACToR database.
1 The analysis of the FracFocus 1.0 project database presented in this chapter did not exclude chemicals that lacked valid
concentration data, in order to present a more inclusive analysis of the potential toxicity of chemicals used in hydraulic
fracturing fluids. The chemical list and percent disclosures listed for each chemical is therefore slightly different that
those shown in Chapter 5 (Table 5-3], which excluded chemicals lacking valid concentration data.
9-24
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-2. Chemicals reported to FracFocus 1.0 from January 1, 2011 to February 28, 2013 in
10% or more disclosures, with the percent of disclosures for which each chemical is reported.
Chronic oral RfVs, TOPKAT LOAEL estimates, and availability of ACToR data are shown when
available.
Chemicals are ordered in the table, from high to low, based on their number of disclosures in the EPA FracFocus
1.0 project database. Water, quartz, and sodium chloride are excluded from this analysis. Asterisk (*) indicates
chemicals that are regulated as drinking water contaminants under the NPDWRs.
Chemical Name
CASRN
% of
Disclo-
sures3
Chronic oral RfVb
QSAR
ACTOR
RfV
(mg/kg
-day)
Source
of RfV
Critical effect0
TOPKAT
LOAELd
(mg/kg)
# of
data
points6
Methanol
67-56-1
73%
2
IRIS
Extra cervical ribs
122
Distillates,
petroleum,
hydrotreated light
64742-47-8
67%
4
Hydrochloric acid
7647-01-0
66%
50
Ethylene glycol
107-21-1
47%
2
IRIS
Kidney toxicity
130
102
Isopropanol
67-63-0
46%
81.4
26
Diammonium
peroxydisulfate
7727-54-0
44%
11
Guar gum
9000-30-0
39%
2
Sodium hydroxide
1310-73-2
39%
26
Propargyl alcohol
107-19-7
33%
0.002
IRIS
Renal and
hepatotoxicity
42
Glutaraldehyde
111-30-8
33%
398
13
Ethanol
64-17-5
31%
59.2
182
Potassium hydroxide
1310-58-3
31%
21
Acetic acid
64-19-7
25%
183
35
Citric acid
77-92-9
24%
55.8
25
2-Butoxyethanol
111-76-2
23%
0.1
IRIS
Hemosiderin
deposition in the
liver
707
44
Solvent naphtha,
petroleum, heavy
arom.
64742-94-5
21%
5
Naphthalene
91-20-3
19%
0.02
IRIS
Decreased terminal
body weight
67.5
157
9-25
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
% of
Disclo-
sures3
Chronic oral RfVb
QSAR
ACTOR
RfV
(mg/kg
-day)
Source
of RfV
Critical effect0
TOPKAT
LOAELd
(mg/kg)
# of
data
points6
2,2-Dibromo-3-
nitrilopropionamide
10222-01-2
16%
52.4
Choline chloride
67-48-1
15%
20.8
24
Phenol-formaldehyde
resin
9003-35-4
14%
Carbonic acid,
dipotassium salt
584-08-7
14%
137
3
Methenamine
100-97-0
14%
12.3
15
Thiourea, polymer
with formaldehyde
and 1-
phenylethanone
68527-49-1
13%
1,2,4-
Trimethylbenzene
95-63-6
13%
0.01
IRIS
Decreased pain
sensitivity
91.5
71
Polyethylene glycol
25322-68-3
13%
5
Polyethylene glycol
nonylphenyl ether
9016-45-9
13%
4
Quaternary
ammonium
compounds, benzyl-
C12-16-alkyldimethyl,
chlorides
68424-85-1
12%
0.44
HHBP
Decreased body
weight and weight
gain
3
Poly(oxy-l,2-
ethanediyl)-
nonylphenyl-hydroxy
branched
127087-87-0
12%
Ammonium chloride
12125-02-9
12%
18
Formic acid
64-18-6
11%
0.9
PPRTV
Reproductive
toxicity
72
Tetrakis(hydroxy-
methyl)
phosphonium sulfate
55566-30-8
11%
148
3
Sodium chlorite*
7758-19-2
11%
0.03
IRIS
Neuro-
developmental
effects
66
9-26
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
% of
Disclo-
sures3
Chronic oral RfVb
QSAR
ACToR
RfV
(mg/kg
-day)
Source
of RfV
Critical effect0
TOPKAT
LOAELd
(mg/kg)
# of
data
points6
Alcohols, C12-14,
ethoxylated
propoxylated
68439-51-0
11%
1450
Sodium persulfate
7775-27-1
10%
6
CASRN = Chemical Abstract Service Registry Number; RfV = Reference value; IRIS = Integrated Risk Information System; PPRTV =
Provisional Peer-Reviewed Toxicity Value; HHBP = Human Health Benchmarks for Pesticides; QSAR = Quantitative structure-
activity relationship; TOPKAT = Toxicity Prediction by Komputer Assisted Technology; ACToR = EPA's Aggregated Computational
Toxicology Online Resource
a The FracFocus frequency of use data presented in this chapter is based on 35,957 FracFocus disclosures that were
deduplicated, within the study time period (January 1, 2011 to February 28, 2013), and with ingredients that have a valid
CASRN.
b Reference value (RfV): An estimate of an exposure for a given duration to the human population (including susceptible
subgroups) that is likely to be without an appreciable risk of adverse health effects over a lifetime. RfVs considered in this
analysis include chronic oral reference doses (RfDs) from IRIS, PPRTV, and HHBP; chronic oral minimal risk levels (MRLs) from
ATSDR; maximum allowable daily levels (MADLs) from CalEPA; and tolerable daily intake (TDI) from CICAD. See Section 9.4.1.
c Critical effect: The first adverse effect, or its known precursor, that occurs to the most sensitive species as the dose rate of an
agent increases.
d TOPKAT LOAEL: The LOAEL is the lowest exposure level at which there are biologically significant increases in frequency or
severity of adverse effects between the exposed population and its appropriate control group. TOPKAT LOAELs were predicted
using a QSAR-based software model, as described in Section 9.4.2. Values are rounded to 3 significant figures.
0 Indicates the total number of data points available for a chemical in the relevant data classes on EPA's ACToR database, as
described in Section 9.4.3.
Table 9-3 shows the chemicals reported in at least 10% of disclosures nationally in the EPA
FracFocus 1.0 project database that are considered by at least one of the sources in Table 9-1 to be
known, probable, or possible human carcinogens. Ethanol is classified as a "carcinogenic to
humans" (Group 1) by IARC. Naphthalene is classified by IARC as "possibly carcinogenic to humans"
(Group 2B), and is classified by RoC as "reasonably anticipated to be a human carcinogen," while
IRIS classifies naphthalene as having inadequate data to assess carcinogenic potential. Neither
chemical has an available OSF.
9-27
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-3. List of OSFs and qualitative cancer classifications available for all carcinogenic
chemicals reported to FracFocus 1.0 from January 1, 2011 to February 28, 2013 in 10% or
more disclosures.
Includes all chemicals from Table 9-2 that are classified as known, probable, or possible human carcinogens by at
least one of the sources in Table 9-1.
Chemical Name
CASRN
OSFa
Qualitative cancer classification
OSF(per
mg/kg-
day)
Source
of OSF
IRISb
PPRTV0
IARCd
RoCe
Ethanol
64-17-5
1
Naphthalene
91-20-3
"Data are inadequate to
assess human
carcinogenic potential"
2B
RAHC
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; PPRTV = Provisional Peer
Reviewed Toxicity Values; IARC = International Agency for Research on Cancer Monographs; RoC = National Toxicology Program
13th Report on Carcinogens
a Oral slope factor (OSF): An upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a lifetime
oral exposure to an agent. This estimate, usually expressed in units of proportion (of a population) affected per mg/kg-day, is
generally reserved for use in the low dose region of the dose response relationship, that is, for exposures corresponding to risks
less than 1 in 100. OSFs considered in this analysis include values from IRIS, PPRTV, HHBP, and CalEPA. See Section 9.4.1.
b IRIS assessments use EPA's 1986,1996,1999, or 2005 guidelines to establish descriptors for summarizing the weight of
evidence as to whether a contaminant is or may be carcinogenic. See glossary in Appendix G for details.
c PPRTV assessments use EPA's 1999 guidelines to establish descriptors for summarizing the weight of evidence as to whether a
contaminant is or may be carcinogenic. See glossary in Appendix G for details.
d The IARC summarizes the weight of evidence as to whether a contaminant is or may be carcinogenic using five weight of
evidence classifications: Group 1: Carcinogenic to humans; Group 2A: Probably carcinogenic to humans; Group 2B: Possibly
carcinogenic to humans; Group 3: Not classifiable as to its carcinogenicity to humans; Group 4: Probably not carcinogenic to
humans. See glossary in Appendix G for details.
aThe listing criteria in the 13th RoC Document are: Known = Known to be a human carcinogen; RAHC = Reasonably anticipated
to be a human carcinogen.
In addition to evaluating chemicals that are frequently used in hydraulic fracturing fluids, we also
evaluated the availability of toxicological data for subsets of chemicals that are used less frequently
on a national basis (Figure 9-4). For this analysis, we binned the chemicals according to frequency
of use as identified from the EPA FracFocus 1.0 project database (>10% of disclosures, 5-10% of
disclosures, 1-5% of disclosures, <1% of disclosures, or unknown frequency of use), and evaluated
the percentage of chemicals within each bin that have available chronic oral RfVs or OSFs, TOPKAT
LOAEL estimates, and relevant data on ACToR. This analysis demonstrates that the availability of
chronic oral RfVs and OSFs is low across all of these subsets of chemicals. Proportionately, the
availability of chronic oral RfVs, OSFs, and data on ACToR is slightly higher for chemicals that are
used in >10% of disclosures, compared to chemicals that are used less frequently.
Of the chemicals on the EPA's list that had frequency of use data available from the EPA FracFocus
1.0 project database, the majority were used in <1% of disclosures (n=480), suggesting that
potential exposure to these chemicals is more likely to be a local issue rather than a national issue.
Given that the analysis of the EPA FracFocus 1.0 project database presented in this chapter was
9-28
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
based on 35,957 disclosures, a chemical used in <1% of wells nationally could still be used in
several hundred wells. Chemicals used infrequently on a national basis could still be used more
frequently within certain areas or counties, increasing the potential for local exposure to that
chemical.
¦ Has chronic oral RfV or OSF ¦ Has TOPKAT LOAEL estimate ¦ Has relevant ACToR data
100%
90%
>10% of disclosures 5-10% of disclosures 1-5% of disclosures <1% of disclosures Unknown frequency
(n=37) (n=36) (n=135) (n=480) ofuse(n=396)
Figure 9-4. Availability of toxicity data (chronic oral RfVs/OSFs, TOPKAT LOAEL estimates, and
relevant data on ACToR) for subsets of chemicals used at various frequencies in hydraulic
fracturing fluids, as determined based on the number of disclosures in the EPA FracFocus 1.0
project database.
As described in Chapter 5, many of the chemicals used in hydraulic fracturing fluids can be
classified as chemical mixtures. Among the most common chemical mixtures on EPA's list of
chemicals are petroleum distillates (i.e., hydrocarbon solvents), which are complex mixtures of
petroleum hydrocarbons.1 Two of the most frequently used chemicals in Table 9-2 are petroleum
distillates. (Petroleum) hydrotreated light distillates is a mixture of hydrocarbons having carbon
numbers predominantly in the range of C9 through CI6, and was reported as used in 67% of
disclosures in the EPA FracFocus 1.0 project database. Heavy aromatic (petroleum) solvent
naphtha is a mixture consisting predominantly of aromatic hydrocarbons in carbon fraction range
of C9 through C16, and was reported as used in 21% of disclosures in the EPA FracFocus 1.0 project
database. These petroleum distillates lack chronic oral RfVs or OSFs, and have little information
available in ACToR. However, a methodology that describes the toxicity and derivation of surrogate
1 Total petroleum hydrocarbons (TPH] is a term used to describe a large family of several hundred chemical compounds
that originally come from crude oil. TPH is a mixture of chemicals, but they are all made mainly from hydrogen and
carbon, called hydrocarbons. TPH are divided into groups of petroleum hydrocarbons that act alike in soil or water. These
groups are called petroleum hydrocarbon fractions. Each hydrocarbon fraction contains many individual chemicals. Some
chemicals that may be found in TPH are hexane, jet fuels, mineral oils, benzene, toluene, xylenes, naphthalene, and
fluorene, as well as other petroleum products and gasoline components. Source: ATSDR f 20111.
9-29
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
toxicity values for such mixtures was developed by the Total Petroleum Hydrocarbon Criteria
Working Group (TPHCWG) fEdwards etal.. 19971. This indicater/surrogate approach uses a
combination of toxicity data and existing RfVs on individual compounds and fraction-specific
mixtures. Examples of compounds present in each fraction include: toluene, ethylbenzene, and
styrene (C5-C8) and isopropylbenzene (cumene), naphthalene, fluorene, pyrene, and
methylnaphthalene (C9-C16). No data was available for consideration for C>16. Applying their
methodology, the TPHCWG derived surrogate aliphatic and aromatic oral toxicity values for
fractions in the C5-C8, C9-C16, and C17-C35 ranges. For aromatics, the toxicity ranking was C9-C16
and C17-C35 > C5-C8; and for aliphatics, the toxicity ranking was C9-C16 > C17-C35 > C5-C8. As
reviewed by the TPHCWG, compounds above C20 are likely not volatile or soluble in groundwater
and will remain at the release site and compounds above C35 are typically not likely to be
bioavailable by the oral route of exposure. These surrogate toxicity values are not included in EPA's
analysis in this report, but this methodology might be useful for risk assessors at sites where these
petroleum distillates are used.
We additionally note that several of the frequently used chemicals in Table 9-2 are designated as
being "generally recognized as safe" (GRAS) for use in food additives or food contact substances by
the U.S. Food and Drug Administration (FDA). This includes hydrochloric acid, guar gum, sodium
hydroxide, sodium chloride, potassium hydroxide, acetic acid, citric acid, choline chloride, carbonic
acid dipotassium salt, ammonium chloride, and formic acid. Overall, 103 chemicals on EPA's list of
chemicals used in hydraulic fracturing fluids have GRAS designations by the FDA. GRAS chemicals
may be used by hydraulic fracturing industry operators in an effort to avoid more hazardous
chemicals and minimizes concern in the public perception (Loveless etal.. 2011). However, GRAS
determinations are often specific to certain conditions as expressed in the FDA GRAS Notification
Database and therefore do not indicate that the same chemical is safe for use in hydraulic fracturing
fluids. For instance, formic acid is considered GRAS for specific use in paper food packaging
materials fU.S. FDA. 20161. but has a chronic oral RfD of 0.9 mg/kg-day based on reproductive
effects (U.S. EPA. 2010b). For human health risk assessment in areas of hydraulic fracturing activity,
hazard and dose-response relations for these chemicals need to be assessed in the context of the
use and levels that are likely to be encountered in an appropriate exposure scenario.
9.5.2 Organic Chemicals in Produced Water
Chapter 7 discussed the detection of volatile and semi-volatile organic chemicals in produced
water. Many of these chemicals, including the BTEX chemicals and related hydrocarbons, occur
naturally in hydrocarbon formations and are characteristic of produced water from oil and gas
production wells in both conventional and unconventional reservoirs. Some of these chemicals
have anthropogenic origins, such as di(2-ethylhexyl) phthalate, which does not occur naturally but
has known use in hydraulic fracturing fluids. Naphthalene is an example of a chemical that may
occur naturally in hydrocarbon formations but is also used frequently in hydraulic fracturing fluids
(19% of disclosures in the EPA FracFocus 1.0 project database; Table 9-2). These chemicals have
the potential to enter drinking water resources through events such as spills of produced water,
mechanical integrity failures, infiltration into groundwater from produced water storage pits, and
persistence in inadequately treated wastewater.
9-30
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Several recent field studies have reported these organic constituents in surface water and
groundwater in areas of hydraulic fracturing activity. For example, the BTEX chemicals, diesel-
range organics, gasoline-range organics, and naphthalene were detected in groundwater
monitoring wells in Pavillion Field, Wyoming, likely as a result of legacy contamination from leaking
unlined production fluid storage pits (Digiulio and Tackson. 2016). BTEX chemicals were also found
to be elevated above their respective MCLs following spills by the oil and gas industry in Colorado,
and were reduced to lower concentrations following remediation fGross et al.. 20131. Ferrar etal.
f20131 reported mean concentrations of the BTEX chemicals in effluent from a centralized waste
treatment (CWT) facility in Pennsylvania ranged from about 2 to 46 |ig/L, with significantly lower
concentrations observed after oil and gas well operators were asked to stop discharging waste at
this facility (Text Box 8-1). In a survey of 500 private and public water supply wells overlying and
adjacent to the Barnett Shale in Texas, Hildenbrand etal. f20151 reported that benzene
concentrations exceeded their MCL in all 34 wells where benzene was detected, while toluene,
ethylbenzene, and xylenes were prevalent at trace levels; the authors noted that BTEX detections
occurred at a high rate in an area that houses a large number of underground injection wells for
drilling waste disposal, but it was not clear that these chemicals originated from hydraulic
fracturing activity or from another potential source.
As there were a large number of organic chemicals identified on EPA's list, this section focuses on
the toxicological evaluation of those organic chemicals that had measured concentration data
available in Appendix E and had at least some toxicity data available from the sources in Table 9-1,
TOPKAT, or ACToR (69 chemicals total).1 There were an additional 46 organic chemicals that had
measured concentration data in Chapter 7 or Appendix E that did not have any toxicity data
available. Organic chemicals that lacked concentration data and are not discussed here.
For this subset of 69 organic chemicals, noncancer toxicity values (chronic oral RfVs and high
confidence TOPKAT LOAEL estimates) and ACToR data availability are shown in Table 9-4, and
cancer information (OSFs and qualitative cancer classifications) are shown in Table 9-5. Chronic
oral RfVs were available for 31 of these chemicals, and ranged from 0.001 mg/kg-day (pyridine) to
0.9 mg/kg-day (acetone). Critical effects for these chemicals include kidney/renal toxicity,
hepatotoxicity, neurotoxicity, reproductive toxicity (decreased maternal weight gain),
developmental toxicity (decreased offspring body weight, fetal toxicity), and decreased terminal
body weight. Six of the chemicals in Table 9-4 are regulated as drinking water contaminants under
the NPDWRs: the BTEX chemicals (benzene, ethylbenzene, toluene, xylenes), benzo(a)pyrene, and
di(2-ethylhexyl) phthalate.
Of the 38 chemicals in Table 9-4 that lack chronic oral RfVs, 10 have high-confidence TOPKAT
LOAEL estimates available. Several of these had similarly low LOAEL estimates:
benzo(g,h,i)perylene, indeno(l,2,3-cd)pyrene, benzo(b)fluoranthene, benzo(k)fluoranthene,
benzo(a)pyrene, dibenz(a,h)anthracene, and N-nitrosodiphenylamine. Notably, 33 of the chemicals
1 Note that chemical names presented in this chapter and in Appendix H sometimes differ from the chemical names
presented with the concentration data in Appendix E. This is because Appendix E uses the chemical names provided by
the original sources of chemical data, while this chapter and Appendix H use chemical names that were verified by EPA
during the curation of the chemical list. See Appendix H for details on the curation of the chemical list.
9-31
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
in Table 9-4 were added to EPA's chemical list after the release of the external review draft (Text
Box 9-2), and therefore were not included in the QSAR analysis (Section 9.4.2).
Table 9-4. List of a subset of organic chemicals that have been detected in produced water,
with respective chronic oral RfVs, TOPKAT LOAEL estimates, and availability of ACToR data
shown when available.
Includes organic chemicals that were identified on the EPA's list of chemicals in produced water (Appendix H) that
have measured concentration data available in Appendix E and have at least some toxicity data available from the
sources consulted by the EPA. Chemicals are ordered in the table from most toxic to least toxic based on chronic
oral RfV. Chemicals without RfVs were ordered based on TOPKAT LOAEL, and then by number of data points on
ACToR. indicates chemicals that are regulated as drinking water contaminants under the NPDWRs.
Chemical Name
CASRN
Chronic oral RfVa
QSAR
estimate
ACToR
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect15
TOPKAT
LOAEL0
(mg/day)
# of
data
points'1
Pyridine
110-86-1
0.001
IRIS
Increased liver weight
69.5
114
Benzidine
92-87-5
0.003
IRIS
Brain cell vacuolization;
liver cell alterations in
females
127
2,4-Dichlorophenol
120-83-2
0.003
IRIS
Decreased delayed
hypersensitivity response
122
Benzene*
71-43-2
0.004
IRIS
Decreased lymphocyte
count
77.6
238
2-Methylnaphthalene
91-57-6
0.004
IRIS
Pulmonary alveolar
proteinosis
103
52
1,3,5-Trimethylbenzene
108-67-8
0.01
IRIS
Decreased pain sensitivity
63
76
1,2,4-Trimethylbenzene
95-63-6
0.01
IRIS
Decreased pain sensitivity
91.5
71
Chloroform
67-66-3
0.01
IRIS
Moderate/marked fatty
cyst formation in the liver
and elevated serum
glutamic pyruvic
transaminase (SGPT)
47.1
221
Naphthalene
91-20-3
0.02
IRIS
Decreased mean terminal
body weight in males
67.5
157
2,4-Dimethylphenol
105-67-9
0.02
IRIS
Clinical signs (lethargy,
prostration, and ataxia)
and hematological
changes
112
88
Di(2-ethylhexyl)
phthalate*
117-81-7
0.02
IRIS
Increased relative liver
weight
4040
229
9-32
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Chronic oral RfVa
QSAR
estimate
ACTOR
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect15
TOPKAT
LOAELc
(mg/day)
# of
data
points'1
Pyrene
129-00-0
0.03
IRIS
Kidney effects (renal
tubular pathology,
decreased kidney
weights)
36.1
129
1,4-Dioxane
123-91-1
0.03
IRIS
Liver and kidney toxicity
207
148
Fluoranthene
206-44-0
0.04
IRIS
Nephropathy, increased
liver weights,
hematological alterations,
and clinical effects
44.6
103
Fluorene
86-73-7
0.04
IRIS
Decreased RBC, packed
cell volume and
hemoglobin
95.1
120
m-Cresol
108-39-4
0.05
IRIS
Decreased body weights
and neurotoxicity
123
103
o-Cresol
95-48-7
0.05
IRIS
Decreased body weights
and neurotoxicity
229
94
Toluene*
108-88-3
0.08
IRIS
Increased kidney weight
163
188
Diphenylamine
122-39-4
0.1
HHBP
Alterations in clinical
chemistry; increased
kidney, liver, and spleen
weights
30.8
86
Carbon disulfide
75-15-0
0.1
IRIS
Fetal toxicity/
malformations
126
89
Benzyl alcohol
100-51-6
0.1
PPRTV
Effects on survival,
growth, and tissue
histopathology
210
45
Ethylbenzene*
100-41-4
0.1
IRIS
Liver and kidney toxicity
226
207
Cumene
98-82-8
0.1
IRIS
Increased average kidney
weight in female rats
246
101
Acetophenone
98-86-2
0.1
IRIS
General toxicity
274
58
Dibutyl phthalate
84-74-2
0.1
IRIS
Increased mortality
2090
143
Xylenes*
1330-20-7
0.2
IRIS
Decreased body weight,
increased mortality
110
174
Benzyl butyl phthalate
85-68-7
0.2
IRIS
Significantly increased
liver-to-body weight and
liver-to-brain weight
ratios
194
Phenol
108-95-2
0.3
IRIS
Decreased maternal
weight gain
134
170
9-33
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Chronic oral RfVa
QSAR
estimate
ACTOR
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect15
TOPKAT
LOAELc
(mg/day)
# of
data
points'1
Biphenyl
92-52-4
0.5
IRIS
Renal papillary
mineralization in male
F344 rats
103
Caprolactam
105-60-2
0.5
IRIS
Reduced offspring body
weight
39
Acetone
67-64-1
0.9
IRIS
Nephropathy
119
79
Benzo(g,h,i)perylene
191-24-2
29.1
68
lndeno(l,2,3-cd)pyrene
193-39-5
38.6
111
Dibenz(a,h)anthracene
53-70-3
38.9
96
Benzo(b)fluoranthene
205-99-2
39
121
Benzo(k)fluoranthene
207-08-9
39
118
N-Nitrosodiphenylamine
86-30-6
39.4
99
Benzo(a)pyrene*
50-32-8
43
184
Phenanthrene
85-01-8
61.3
69
p-Cresol
106-44-5
95.5
98
Dioctyl phthalate
117-84-0
4740
61
Caffeine
58-08-2
134
Benz(a)anthracene
56-55-3
122
Chrysene
218-01-9
114
2-Mercaptobenzothiazole
149-30-4
95
1,2-Diphenylhydrazine
122-66-7
83
Dimethyl phthalate
131-11-3
79
N-Nitroso-N-
methylethylamine
10595-95-6
42
4-(l,l,3,3-
Tetramethylbutyl)phenol
140-66-9
30
p-Tert-butylphenol
98-54-4
27
2,6-Di-tert-butylphenol
128-39-2
22
Dimethylphenol
1300-71-6
17
2-Ethylhexyl diphenyl
phosphate (Octicizer)
1241-94-7
14
2,5-Cyclohexadiene-l,4-
dione
106-51-4
12
Cholesterol
57-88-5
11
9-34
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Chronic oral RfVa
QSAR
estimate
ACToR
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect15
TOPKAT
LOAEL0
(mg/day)
# of
data
points'1
Benzothiazole
95-16-9
10
Octadecanoic acid
57-11-4
9
Butanoic acid, butyl ester
109-21-7
9
Tetradecanoic acid
544-63-8
7
Triphenyl phosphate
115-86-6
7
Dodecanoic acid
143-07-7
6
Drometrizole
2440-22-4
6
3-(4-Methoxyphenyl)-2-
ethylhexylester-2-
propenoic acid
5466-77-3
6
2,6-Bis(dimethylethyl)-
2,5-cyclohexadiene-l,4-
dione
719-22-2
3
Diphenylmethane
101-81-5
3
Isopropyl myristate
110-27-0
2
2-[2-[4-(l,l,3,3-
tetramethylbutyl)phen-
oxy]ethoxy]-ethanol
2315-61-9
2
Sterane
50-24-8
1
3-(4-Methoxyphenyl)-2-
propenoic acid
830-09-1
1
CASRN = Chemical Abstract Service Registry Number; RfV = Reference value; IRIS = Integrated Risk Information System; PPRTV =
Provisional Peer-Reviewed Toxicity Value; HHBP = Human Health Benchmarks for Pesticides; QSAR = Quantitative structure-
activity relationship; TOPKAT = Toxicity Prediction by Komputer Assisted Technology; ACToR = EPA's Aggregated Computational
Toxicology Online Resource
a Reference value (RfV): An estimate of an exposure for a given duration to the human population (including susceptible
subgroups) that is likely to be without an appreciable risk of adverse health effects over a lifetime. RfVs considered in this
analysis include chronic oral reference doses (RfDs) from IRIS, PPRTV, and HHBP; chronic oral minimal risk levels (MRLs) from
ATSDR; maximum allowable daily levels (MADLs) from CalEPA; and tolerable daily intake (TDI) from CICAD. See Section 9.4.1.
b Critical effect: The first adverse effect, or its known precursor, that occurs to the most sensitive species as the dose rate of an
agent increases.
cTOPKAT LOAEL: The LOAEL is the lowest exposure level at which there are biologically significant increases in frequency or
severity of adverse effects between the exposed population and its appropriate control group. TOPKAT LOAELs were predicted
using a QSAR-based software model, as described in Section 9.4.2. Values are rounded to 3 significant figures.
d Indicates the total number of data points available for a chemical in the relevant data classes on EPA's ACToR database, as
described in Section 9.4.3.
Of the organic chemicals in produced water listed in Table 9-4,17 have available OSFs and 23 are
classified as known, probable, or possible carcinogens (Table 9-5). Benzidine and benzene were
both classified as human carcinogens by IRIS, IARC, and RoC, with benzidine being the most potent
9-35
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
carcinogen listed in Table 9-5 (OSF of 230 per mg/kg-day). Benzo(a)pyrine is classified as a human
carcinogen by IARC, and as a probable human carcinogen by IRIS. The remaining chemicals were
classified as probable or possible human carcinogens.
Table 9-5. List of OSFs and qualitative cancer classifications available for a subset of organic
chemicals that have been reported in produced water.
Includes organic chemicals that were identified on EPA's list of chemicals in produced water (Appendix H) that
have measured concentration data available in Chapter 7 or Appendix E (Table 9-4) and are classified as known,
probable, or possible carcinogens. Chemicals that had OSFs available are ordered in this table from most potent
(highest OSF) to least potent (lowest OSF).
Chemical Name
CASRN
OSFs3
Qualitative Cancer Classifications
OSF(per
mg/kg-
day)
Source
of OSF
IRIS b
PPRTVc
IARC d
RoCe
Benzidine
92-87-5
230
IRIS
A (Human
carcinogen)
1
Known
N-Nitroso-N-
methylethylamine
10595-95-6
22
IRIS
B2 (Probable
human
carcinogen)
2B
Benzo(a)pyrene
50-32-8
7.3
IRIS
B2 (Probable
human
carcinogen)
1
RAHC
Dibenz(a,h)anthra-
cene
53-70-3
4.1
CalEPA
2A
RAHC
lndeno(l,2,3-
cd)pyrene
193-39-5
1.2
CalEPA
2B
RAHC
Benzo(b)fluoran-
thene
205-99-2
1.2
CalEPA
2B
RAHC
Benzo(k)fluoranthene
207-08-9
1.2
CalEPA
2B
RAHC
1,2-
Diphenylhydrazine
122-66-7
0.8
IRIS
B2 (Probable
human
carcinogen)
RAHC
Benz(a)anthracene
56-55-3
0.7
PPRTV
B2 (Probable
human
carcinogen)
2B
RAHC
Chrysene
218-01-9
0.12
CalEPA
B2 (Probable
human
carcinogen)
2B
1,4-Dioxane
123-91-1
0.1
IRIS
"Likely to be
carcinogenic to
humans"
2B
RAHC
9-36
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
OSFsa
Qualitative Cancer Classifications
OSF(per
mg/kg-
day)
Source
of OSF
IRIS b
PPRTVc
IARC d
RoCe
Benzene
71-43-2
0.015-
0.055
IRIS
A (Human
carcinogen)
1
Known
Chloroform
67-66-3
0.019
CalEPA
B2 (Probable
human
carcinogen)
2B
RAHC
Di(2-ethylhexyl)
phthalate
117-81-7
0.014
IRIS
B2 (Probable
human
carcinogen)
2B
RAHC
Ethylbenzene
100-41-4
0.011
CalEPA
D (Not
classifiable as to
human
carcinogenicity)
2B
Biphenyl
92-52-4
0.008
IRIS
"Suggestive
evidence of
carcinogenic
potential"
N-Nitrosodiphenyl-
amine
86-30-6
0.0049
IRIS
B2 (Probable
human
carcinogen)
3
Naphthalene
91-20-3
"Data are
inadequate to
assess human
carcinogenic
potential"
2B
RAHC
Cumene
98-82-8
D (Not
classifiable as to
human
carcinogenicity)
2B
RAHC
2-Mercaptobenzo-
thiazole
149-30-4
2A
m-Cresol
108-39-4
C (Possible
human
carcinogen)
o-Cresol
95-48-7
C (Possible
human
carcinogen)
"Data are
inadequate for
the assessment
of human
carcinogenic
potential"
9-37
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
OSFsa
Qualitative Cancer Classifications
OSF(per
mg/kg-
day)
Source
of OSF
IRIS b
PPRTVc
IARC d
RoCe
Benzyl butyl
phthalate
85-68-7
C (Possible
human
carcinogen)
3
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; PPRTV = Provisional Peer
Reviewed Toxicity Values; HHBP = Human Health Benchmarks for Pesticides; CalEPA = California Environmental Protection
Agency; IARC = International Agency for Research on Cancer Monographs; RoC = National Toxicology Program 13th Report on
Carcinogens
a Oral slope factor (OSF): An upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a lifetime
oral exposure to an agent. This estimate, usually expressed in units of proportion (of a population) affected per mg/kg-day, is
generally reserved for use in the low dose region of the dose response relationship, that is, for exposures corresponding to risks
less than 1 in 100. OSFs considered in this analysis include values from IRIS, PPRTV, HHBP, and CalEPA. See Section 9.4.1.
b IRIS assessments use EPA's 1986,1996,1999, or 2005 guidelines to establish descriptors for summarizing the weight of
evidence as to whether a contaminant is or may be carcinogenic. See glossary in Appendix G for details.
c PPRTV assessments use EPA's 1999 guidelines to establish descriptors for summarizing the weight of evidence as to whether a
contaminant is or may be carcinogenic. See glossary in Appendix G for details.
d The IARC summarizes the weight of evidence as to whether a contaminant is or may be carcinogenic using five weight of
evidence classifications: Group 1: Carcinogenic to humans; Group 2A: Probably carcinogenic to humans; Group 2B: Possibly
carcinogenic to humans; Group 3: Not classifiable as to its carcinogenicity to humans; Group 4: Probably not carcinogenic to
humans. See glossary in Appendix G for details.
0 The listing criteria in the 13th RoC Document are: Known = Known to be a human carcinogen; RAHC = Reasonably anticipated
to be a human carcinogen.
9.5.3 Inorganic Chemicals and TENORM in Produced Water
Chapter 7 discussed the detection of inorganic constituents such as metals, inorganic ions, and
TENORM in produced water. Examples include barium, cadmium, chromium, copper, lead,
manganese, nickel, zinc, and radium. In general, these chemicals are naturally occurring, and are
characteristic of produced water from both conventional and unconventional reservoirs. These
chemicals have the potential to enter drinking water resources through events such as spills of
produced water, mechanical integrity failures, infiltration into groundwater from produced water
storage pits, and persistence in inadequately treated wastewater.
The entry of inorganic constituents of produced water into drinking water resources has been
documented in numerous studies. In Pennsylvania, elevated levels of barium and strontium have
been observed in CWT effluent ("PA PEP, 2015a], with effluent concentrations dropping after oil and
gas well operators were asked to stop discharging waste at this facility (see Text Box 8-1 for details
on temporal trends in wastewater management in Pennsylvania). Likewise, effluent concentrations
at two publicly owned treatment words (POTWs) that had accepted Marcellus wastewater were
found to have lower concentrations of bromide, chloride, barium, strontium, and sulfate after oil
and gas well operators were asked to stop discharging waste at this facility in May 2011 (Ferrar et
al., 2013"). Effluents from POTWs and CWTs that handle Marcellus Shale wastewater have been
found to have levels of radium-226 and radium-228 that exceed the MCL for radium and are
9-38
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
significantly higher than typical background levels of radium in river water (PA DEP. 2015b).
Radium-226 and radium-228 have been demonstrated to accumulate in sediments near the outfalls
of CWTs and of POTWs that handle oil and gas wastewater from CWTs fPA DEP. 2015b: Warner et
al.. 2013a], and in sediments receiving effluent from landfills that accept oil and gas wastes (PA
DEP. 2015b). In West Virginia, water samples collected downstream of a hydraulic fracturing
wastewater injection facility had elevated specific conductance and total dissolved solids, elevated
bromide, chloride, sodium, barium, strontium, and lithium concentrations, and different strontium
isotope ratios compared to those found in upstream, background waters fAkob etal.. 20161. In a
survey of 500 groundwater wells overlying and adjacentto the Barnett Shale in Texas, Hildenbrand
etal. (2015) reported a variety of metals and anions that are known produced water constituents at
concentrations that sometimes exceeded primary or secondary MCLs, health advisory levels, or
other suggested levels as provided in the EPA Drinking Water Standards, although it was not clear
that these chemicals originated from nearby hydraulic fracturing activity or from other potential
sources.
For the inorganic chemicals that were identified in produced water on EPA's chemical list,
noncancer toxicity values (chronic oral RfVs) and ACToR data availability for these chemicals are
shown in Table 9-6, and cancer information (OSFs and qualitative cancer classifications) are shown
in Table 9-7. As shown in Table 9-6, chronic oral RfVs were available for 26 of these chemicals,
ranging from 0.00002 mg/kg-day (phosphorus) to 1.6 mg/kg-day (nitrate). Critical effects for these
metals include neurotoxicity, developmental and liver toxicity, hyperpigmentation and keratosis of
the skin, and decrements in blood copper status and enzyme activity. Nineteen of the inorganic
chemicals in Table 9-6 are regulated as drinking water contaminants under the NPDWR.
All but one of these inorganic chemicals had at least some relevant data available on EPA's ACToR
database. However, none of the inorganic chemicals have TOPKAT LOAEL estimates available, as
this QSAR model is only able to generate estimates for organic chemicals (Section 9.4.2).
Table 9-6. List of inorganics and TENORM reported in produced water, and respective chronic
oral RfVs and OSFs when available.
Includes inorganic chemicals that were identified on EPA's list of chemicals in produced water (Appendix H).
Chemicals are ordered from most toxic to least toxic based on chronic oral RfV. Chemicals without chronic oral
RfVs were ordered in terms of the number of data points on ACToR. indicates chemicals are regulated as drinking
water contaminants under the NPDWR.
Chemical Name
CASRN
Chronic oral RfVsa
ACToR
RfV
(mg/kg-
day)
Source
of RfV
Critical effect15
# of
data
points0
Phosphorus
7723-14-0
0.00002
IRIS
Parturition mortality;
forelimb hair loss
113
Vanadium
7440-62-2
0.00007
PPRTV
Kidney histopathology
76
Arsenic*
7440-38-2
0.0003
IRIS
Hyperpigmentation and
vascular complications
243
9-39
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Chronic oral RfVsa
ACTOR
RfV
(mg/kg-
day)
Source
of RfV
Critical effect15
# of
data
points0
Cobalt
7440-48-4
0.0003
PPRTV
Decreased iodine uptake
76
Antimony*
7440-36-0
0.0004
IRIS
Hematological;
alterations in glucose
and cholesterol
163
Cadmium*
7440-43-9
0.0005
IRIS
Proteinuria
230
Beryllium*
7440-41-7
0.002
IRIS
Intestinal lesions
186
Mercury
7439-97-6
0.002
CICAD
Renal toxicity
177
Lithium
7439-93-2
0.002
PPRTV
Adverse effects in
multiple organ systems
43
Chromium (VI)
18540-29-9
0.003
IRIS
None reported
120
Selenium*
7782-49-2
0.005
IRIS
Clinical selenosis
232
Silver
7440-22-4
0.005
IRIS
Argyria
120
Molybdenum
7439-98-7
0.005
IRIS
Increased uric acid levels
73
Iodine
7553-56-2
0.01
CICAD
27
Nitrite*
14797-65-0
0.1
IRIS
Methemoglobinemia
109
Chlorine
7782-50-5
0.1
IRIS
No adverse effect level
116
Manganese
7439-96-5
0.14
IRIS
Central nervous system
(CNS) effects
128
Barium*
7440-39-3
0.2
IRIS
Nephropathy
167
Boron
7440-42-8
0.2
IRIS
Decreased fetal weight
(developmental)
93
Zinc
7440-66-6
0.3
IRIS
Decreases in erythrocyte
Cu, Zn-superoxide
dismutase (ESOD)
activity in humans
163
Lead*
7439-92-1
0.5 ng/dayd
CalEPA
Reproductive Toxicity
168
Strontium
7440-24-6
0.6
IRIS
Rachitic bone
67
Iron
7439-89-6
0.7
PPRTV
Adverse gastrointestinal
effects
73
Aluminum
7429-90-5
1
PPRTV
Neurotoxicity
88
Chromium (III)
16065-83-1
1.5
IRIS
No effects observed
71
9-40
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Chronic oral RfVsa
ACTOR
RfV
(mg/kg-
day)
Source
of RfV
Critical effect15
# of
data
points0
Nitrate*
14797-55-8
1.6
IRIS
Clinical signs of
methemoglobinemia in
excess of 10%
130
Nickel
7440-02-0
181
Copper*
7440-50-8
163
Thallium*
7440-28-0
136
Chromium
7440-47-3
125
Uranium-238*
7440-61-1
100
Ammonia
7664-41-7
90
Zirconium
7440-67-7
55
Alpha particle*
12587-46-1
55
Fluoride*
16984-48-8
53
Radium*
7440-14-4
52
Beta particle*
12587-47-2
51
Magnesium
7439-95-4
40
Tin
7440-31-5
40
Chloride
16887-00-6
32
Sodium
7440-23-5
31
Sulfate
14808-79-8
27
Potassium
7440-09-7
25
Titanium
7440-32-6
25
Calcium
7440-70-2
24
Radium-226*
13982-63-3
13
Radium-228*
15262-20-1
11
Sulfide
18496-25-8
11
Caesium
7440-46-2
7
Caesium-137
10045-97-3
6
Silicon
7440-21-3
5
Rubidium
7440-17-7
5
9-41
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Chronic oral RfVsa
ACToR
RfV
(mg/kg-
day)
Source
of RfV
Critical effect15
# of
data
points0
Bromide
24959-67-9
2
Sulfite
14265-45-3
1
Uranium-235*
15117-96-1
1
Octasulfur
10544-50-0
CASRN = Chemical Abstract Service Registry Number; RfV = Reference value; IRIS = Integrated Risk Information System; PPRTV =
Provisional Peer-Reviewed Toxicity Value; CalEPA = California Environmental Protection Agency; CICAD = Concise International
Chemical Assessment Documents; ACToR = EPA's Aggregated Computational Toxicology Online Resource
a Reference value (RfV): An estimate of an exposure for a given duration to the human population (including susceptible
subgroups) that is likely to be without an appreciable risk of adverse health effects over a lifetime. RfVs considered in this
analysis include chronic oral reference doses (RfDs) from IRIS, PPRTV, and HHBP; chronic oral minimal risk levels (MRLs) from
ATSDR; maximum allowable daily levels (MADLs) from CalEPA; and tolerable daily intake (TDI) from CICAD. See Section 9.4.1.
b Critical effect: The first adverse effect, or its known precursor, that occurs to the most sensitive species as the dose rate of an
agent increases.
c Indicates the total number of data points available for a chemical in the relevant data classes on EPA's ACToR database, as
described in Section 9.4.3.
d CalEPA MADLs are in units of [ag/day, while all other chronic oral RfVs in this table are in units of mg/kg-day.
OSFs were available for 4 of the inorganic chemicals reported in produced water, and 14 are
classified as known or probable carcinogens (Table 9-7). OSFs ranged from 15 per mg/kg-day for
cadmium to 0.0085 mg/kg-day for lead. Chromium (VI), arsenic, alpha particle, beta particle,
radium-226, and radium-288 are all classified as known human carcinogens by all sources
reporting in this table. Beryllium and cadmium are both classified as known human carcinogens by
IARC and NTP, and as probable human carcinogens by EPA. Lead, cobalt, nickel, nitrate, and nitrite
are classified by these sources as possible or probable human carcinogens.
Table 9-7. List of qualitative cancer classifications available for inorganics and NORM that
were reported in produced water.
Includes inorganic chemicals that were identified on EPA's list of chemicals in produced water (Appendix H) that
classified as known, probable, or possible carcinogens by at least one of the sources in Table 9-1. Chemicals that
had OSFs available are ordered in this table from most potent (highest OSF) to least potent (lowest OSF).
Chemical Name
CASRN
OSFa
Qualitative Cancer Classifications
OSF (per
mg/kg-
day)
Source
of OSF
IRISb
PPRTV0
IARCd
RoCe
Cadmium
7440-43-9
15
CalEPA
B1 (Probable
human
carcinogen)
1
Known
Arsenic
7440-38-2
1.5
IRIS
A (Human
carcinogen)
1
Known
9-42
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
OSFa
Qualitative Cancer Classifications
OSF (per
mg/kg-
day)
Source
of OSF
IRISb
PPRTV0
IARCd
RoCe
Chromium (VI)
18540-29-9
0.5
CalEPA
Inhaled: A
(Human
carcinogen)
Oral: D (Not
classifiable as to
human
carcinogenicity)
1
Known
Lead
7439-92-1
0.0085
CalEPA
B2 (Probable
human
carcinogen)
2B
RAHC
Alpha particle
12587-46-1
1
Beryllium
7440-41-7
B1 (Probable
human
carcinogen)
1
Known
Beta particle
12587-47-2
1
Radium
7440-14-4
1
Radium-226
13982-63-3
1
Radium-228
15262-20-1
1
Cobalt
7440-48-4
Likely to be
carcinogenic to
humans
2B
Nickel
7440-02-0
2B
RAHC
Nitrate
14797-55-8
2A
Nitrite
14797-65-0
2A
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; PPRTV = Provisional Peer
Reviewed Toxicity Values; HHBP = Human Health Benchmarks for Pesticides; CalEPA = California Environmental Protection
Agency; IARC = International Agency for Research on Cancer Monographs; RoC = National Toxicology Program 13th Report on
Carcinogens
a Oral slope factor (OSF): An upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a lifetime
oral exposure to an agent. This estimate, usually expressed in units of proportion (of a population) affected per mg/kg-day, is
generally reserved for use in the low dose region of the dose response relationship, that is, for exposures corresponding to risks
less than 1 in 100. OSFs considered in this analysis include values from IRIS, PPRTV, HHBP, and CalEPA. See Section 9.4.1.
b IRIS assessments use EPA's 1986,1996,1999, or 2005 guidelines to establish descriptors for summarizing the weight of
evidence as to whether a contaminant is or may be carcinogenic. See glossary in Appendix G for details.
c PPRTV assessments use EPA's 1999 guidelines to establish descriptors for summarizing the weight of evidence as to whether a
contaminant is or may be carcinogenic. See glossary in Appendix G for details.
d The IARC summarizes the weight of evidence as to whether a contaminant is or may be carcinogenic using five weight of
evidence classifications: Group 1: Carcinogenic to humans; Group 2A: Probably carcinogenic to humans; Group 2B: Possibly
carcinogenic to humans; Group 3: Not classifiable as to its carcinogenicity to humans; Group 4: Probably not carcinogenic to
humans. See glossary in Appendix G for details.
aThe listing criteria in the 13th RoC Document are: Known = Known to be a human carcinogen; RAHC = Reasonably anticipated
to be a human carcinogen.
9-43
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.5.4 Organochlorine Pesticides and Polychlorinated Biphenyls (PCBs) in Produced Water
EPA's list of chemicals detected in produced water includes several chemicals that have been
banned from commercial use: specifically, organochlorine pesticides and Aroclor 1248, which is a
commercial PCB mixture. These chemicals were reported by two of the sources used to compile
EPA's chemical list (Appendix H): a technical report prepared by the Gas Technology Institute for
the Marcellus Shale Coalition (MSC), which is a drilling industry trade group fHaves. 20091: and a
report by the New York State Department of Environmental Conservation (NYSDEC), which
referenced the results of the MSC study fNYSDEC. 20111. These chemicals are listed in Table 9-8
along with their respective noncancer toxicity values (chronic oral RfVs and TOPKAT LOAELs) and
availability of relevant toxicological information on ACToR. Cancer information (OSF or qualitative
cancer classification) for these chemicals is listed in Table 9-9.
There is uncertainty about why organochlorine pesticides and PCBs were detected, as they are not
used in hydraulic fracturing fluids and are not naturally occurring. The MSC study stated the
banned substances were detected sporadically and at low concentrations, and suggested they may
have originated from laboratory contamination. The NYSDEC report suggested that the banned
substances may have been introduced to the shale or the water as a result of drilling or fracturing
operations. It is possible that these chemicals were present as legacy contaminants in the source
water used for hydraulic fracturing fluid formulation, or were mobilized from the environment near
the well. Although these chemicals are notable for their high toxicity, the extent to which these
chemicals may be detected in produced water from other hydraulic fracturing sites is not clear.
Chronic oral RfVs for these organochlorine pesticides ranged from 0.000013 mg/kg-day
(Heptachlor epoxide) to 0.0005 mg/kg-day (heptachlor), and were all based on liver toxicity. All of
these pesticides had TOPKAT LOAEL estimates, and all have relevant data available within EPA's
ACToR database.). Heptachlor epoxide, heptachlor, and lindane are regulated as drinking water
contaminants under the NPDWR.
Table 9-8. List of organochlorine pesticides and PCBs that were reported in produced water,
and their respective chronic oral RfVs, TOPKAT LOAEL estimates, and availability of data in
EPA's ACToR database.
Includes banned chemicals that were identified on EPA's list of chemicals in produced water (Appendix H).
Chemicals are ordered from most toxic to least toxic based on chronic oral RfV. Chemicals without chronic oral
RfVs were ordered in terms of the number of data points on ACToR. indicates chemicals that are regulated as
drinking water contaminants under the NPDWRs.
Chemical Name
CASRN
Chronic oral RfVa
QSAR
ACToR
RfV
(mg/kg-
day)
Source
of RfV
Critical effect15
TOPKAT
LOAEL0
(mg/kg)
# of
data
points'1
Heptachlor epoxide*
1024-57-3
0.000013
IRIS
Increased liver-to-body
weight ratio in both
males and females
0.595
168
Aldrin
309-00-2
0.00003
IRIS
Liver toxicity
0.743
166
9-44
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Chronic oral RfVa
QSAR
ACToR
RfV
(mg/kg-
day)
Source
of RfV
Critical effect15
TOPKAT
LOAEL0
(mg/kg)
# of
data
points'1
Dieldrin
60-57-1
0.00005
IRIS
Liver lesions
0.442
167
Lindane*
58-89-9
0.0003
IRIS
Liver and kidney
toxicity
23.9
238
Heptachlor*
76-44-8
0.0005
IRIS
Liver weight increases
in males
0.927
203
beta-
Hexachlorocyclohexane
319-85-7
23.9
88
delta-
Hexachlorocyclohexane
319-86-8
23.9
22
Aroclor 1248
12672-29-6
21.87
35
p,p'-DDE
72-55-9
14.6
103
Endrin aldehyde
7421-93-4
4.09
27
Endosulfan 1
959-98-8
2.27
32
Endosulfan II
33213-65-9
2.27
32
CASRN = Chemical Abstract Service Registry Number; RfV = Reference value; IRIS = Integrated Risk Information System; QSAR =
Quantitative structure-activity relationship; TOPKAT = Toxicity Prediction by Komputer Assisted Technology; ACToR = EPA's
Aggregated Computational Toxicology Online Resource
a Reference value (RfV): An estimate of an exposure for a given duration to the human population (including susceptible
subgroups) that is likely to be without an appreciable risk of adverse health effects over a lifetime. RfVs considered in this
analysis include chronic oral reference doses (RfDs) from IRIS, PPRTV, and HHBP; chronic oral minimal risk levels (MRLs) from
ATSDR; maximum allowable daily levels (MADLs) from CalEPA; and tolerable daily intake (TDI) from CICAD. See Section 9.4.1.
b Critical effect: The first adverse effect, or its known precursor, that occurs to the most sensitive species as the dose rate of an
agent increases.
cTOPKAT lowest-observed-adverse-effect level (LOAEL): The LOAEL is the lowest exposure level at which there are biologically
significant increases in frequency or severity of adverse effects between the exposed population and its appropriate control
group. TOPKAT LOAELs were predicted using a QSAR-based software model, as described in Section 9.4.2.
d Indicates the total number of data points available for a chemical in the relevant data classes on EPA's ACToR database, as
described in Section 9.4.3.
OSFs were available for 7 of the organochlorine pesticides that are classified as known, probable, or
possible human carcinogens (Table 9-9). OSFs ranged from 17 per mg/kg-day (aldrin) to 0.34 per
mg/kg-day (p,p'-DDE). Aldrin, dieldrin, heptachlor epoxide, heptachlor, beta-
hexachlorocyclohexane, and p,p'-DDE are classified as probable or possible carcinogens. Lindane is
classified as a known carcinogen by IARC, and as "reasonably anticipated to be a human
carcinogen" by RoC.
9-45
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-9. List of OSFs and qualitative cancer classifications available for organochlorine
pesticides reported in produced water.
Includes banned chemicals that were identified on EPA's list of chemicals in produced water (Appendix H) that are
classified as known, probable, or possible carcinogens by at least one of the sources in Table 9-1. Chemicals are
ordered in this table from most potent (highest OSF) to least potent (lowest OSF).
Chemical Name
CASRN
OSFa
Qualitative cancer classifications
OSF(per
mg/kg-
day)
Source
of OSF
IRISb
PPRTVC
IARCd
RoCe
Aldrin
309-00-2
17
IRIS
B2 (Probable human
carcinogen)
3
Dieldrin
60-57-1
16
IRIS
B2 (Probable human
carcinogen)
3
Heptachlor
epoxide
1024-57-3
9.1
IRIS
B2 (Probable human
carcinogen)
Heptachlor
76-44-8
4.5
IRIS
B2 (Probable human
carcinogen)
2B
beta-
Hexachlorocyclohe
xane
319-85-7
1.8
IRIS
C (Possible human
carcinogen)
Lindane
58-89-9
1.1
CalEPA
1
RAHC
p,p'-DDE
72-55-9
0.34
IRIS
B2 (Probable human
carcinogen)
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; CalEPA = California
Environmental Protection Agency; IARC = International Agency for Research on Cancer Monographs; RoC = National Toxicology
Program 13th Report on Carcinogens
a Oral slope factor (OSF): An upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a lifetime
oral exposure to an agent. This estimate, usually expressed in units of proportion (of a population) affected per mg/kg-day, is
generally reserved for use in the low dose region of the dose response relationship, that is, for exposures corresponding to risks
less than 1 in 100. OSFs considered in this analysis include values from IRIS, PPRTV, HHBP, and CalEPA. See Section 9.4.1.
b IRIS assessments use EPA's 1986,1996,1999, or 2005 guidelines to establish descriptors for summarizing the weight of
evidence as to whether a contaminant is or may be carcinogenic. See glossary in Appendix G for details.
c PPRTV assessments use EPA's 1999 guidelines to establish descriptors for summarizing the weight of evidence as to whether a
contaminant is or may be carcinogenic. See glossary in Appendix G for details.
d The IARC summarizes the weight of evidence as to whether a contaminant is or may be carcinogenic using five weight of
evidence classifications: Group 1: Carcinogenic to humans; Group 2A: Probably carcinogenic to humans; Group 2B: Possibly
carcinogenic to humans; Group 3: Not classifiable as to its carcinogenicity to humans; Group 4: Probably not carcinogenic to
humans. See glossary in Appendix G for details.
aThe listing criteria in the 13th RoC Document are: Known = Known to be a human carcinogen; RAHC = Reasonably anticipated
to be a human carcinogen.
9.5.5 Methane in Stray Gas
Chapter 6 discussed stray gas as a potential hazard in areas of hydraulic fracturing activity (Text
Box 6-3). Stray gas refers to the phenomenon of natural gas (primarily methane, plus lesser
amounts of ethane) migrating into shallow groundwater, into water wells, or to the surface (e.g.,
cellars, streams, or springs). As discussed in Chapter 6, some studies indicate an association
9-46
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
between hydraulic fracturing activity and elevated methane concentrations in drinking water, while
other studies did not find such a correlation. Potential pathways for migration of stray gas into
aquifers include pathways along production wells with casing and/or cement issues, through
naturally existing fractures, through induced fractures, or via a route that is some combination of
these pathways.
Although ingestion of methane is not considered to be toxic, it has the potential to pose a physical
hazard. Methane can accumulate to explosive levels when allowed to exsolve (degas) from
groundwater in closed environments. High concentrations of methane may also displace oxygen
and act as an asphyxiant (NIOSH. 20001. potentially causing suffocation, loss of consciousness, or
symptoms such as headache and nausea. Methane is not a regulated drinking water contaminant.
Methane does not have an RfV, OSF, or qualitative cancer classification available from any of the
sources consulted by EPA, and did not have a high-confidence TOPKAT LOAEL estimate.
Information on methane is available within the ACToR database.
9.5.6 Disinfection Byproducts (DBPs) Formed from Wastewater Constituents
Some of the inorganic constituents of hydraulic fracturing produced water, including chloride,
bromine, iodine, and ammonium, can contribute to the formation of DBPs during wastewater
treatment (Harkness etal.. 2015: Parker etal.. 20141. The entry of these constituents into drinking
water resources—e.g., as a result of wastewater spills or from the discharge of inadequately treated
hydraulic fracturing wastewater—can result in DBPs in finished drinking water from downstream
drinking water treatment plants (States etal.. 2013). DBPs may also be formed when hydraulic
fracturing produced water is treated at a centralized or publicly owned treatment works, and may
reach drinking water resources when the treated wastewater is discharged to surface water (Hladik
etal.. 20141. Currently, there are no data available on the concentrations of DBPs in finished
drinking water as related to contributions of DBP precursors from hydraulic fracturing wastewater.
Regulated DBPs such as bromate, chlorite, haloacetic acids, and trihalomethanes are a small subset
of the full spectrum of DBPs that include other chlorinated and brominated DBPs as well as
nitrogenous and iodated DBPs. Long term exposure to these DBPs can result in an increased risk of
cancer, anemia, liver and kidney effects, and central nervous system effects. Some of the
unregulated DBPs may be more toxic than their regulated counterparts (Harkness etal.. 2015:
McGuire etal.. 2014: Parker etal.. 2014). In addition, brominated forms of DBPs are considered to
be more cytotoxic, genotoxic, and carcinogenic than chlorinated species based on studies using
rodents, various types of human cells, and a salmonella strain containing human P450 genes
(McGuire etal.. 2014: Parker etal.. 2014: States etal.. 2013: Krasner. 2009: Richardson etal.. 2007).
As with brominated DBPs, there is concern that some iodinated forms of DBPs are more cytotoxic
and genotoxic than chlorinated species fMcGuire etal.. 2014: Parker etal.. 2014: Krasner. 2009:
Richardson etal.. 20071. as evidenced by studies involving rodent research and human cell research
fPlewa etal.. 2010: Plewa and Wagner. 2009: Richardson et al.. 20071. The MCLs (mg/L) for the
regulated DBPs are: 0.01 for bromate, 1.0 for chlorite, 0.06 for haloacetic acid, and 0.08 for total
trihalomethanes.
9-47
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.5.7 Chemicals Detected in Multiple Stages of the Hydraulic Fracturing Water Cycle
As mentioned in Section 9.3 above, there were a total of 77 chemicals on EPA's list that were
identified as being used in hydraulic fracturing fluids and detected in produced water. The presence
of these chemicals within both of these stages of the hydraulic fracturing water cycle may indicate
that these chemicals persist after they are injected into the well. However, this is not necessarily the
case, as some of these chemicals (e.g., BTEX, naphthalene, metals) also occur naturally in oil and gas
reservoirs. Additionally, the EPA's list of chemicals used in hydraulic fracturing fluids and list of
chemicals in produced water were compiled from different sets of sources, and does not provide a
matched comparison between the chemicals used in hydraulic fracturing fluid and the chemicals
present in produced water at a particular site. There may have been other chemicals in present in
produced water that were not detected by these studies due to limitations of analytical chemistry.
Thus, the EPA's composited chemical list cannot reliably be used to draw conclusions on the
persistence of hydraulic fracturing chemicals following well injection.
Of the 77 chemicals identified in both hydraulic fracturing fluids and produced water, 45 have a
chronic oral RfV or OSF available from at least one of the sources in Table 9-1. These 45 chemicals
and their respective toxicity values are shown in Table 9-10, with frequency of use data from the
EPA FracFocus 1.0 project database provided when available. Eleven of these chemicals are
regulated as drinking water contaminants.
Table 9-10. List of 45 chemicals on EPA's list that were used in hydraulic fracturing fluids and
detected in produced water and have an RfV or OSF available.
Frequency of use data from the EPA FracFocus 1.0 project database is provided when available. Chemicals with
available data from the FracFocus 1.0 project database are ordered from high to low based on frequency of use.
Chemicals without frequency of use data are ordered from most toxic to least toxic based on chronic oral RfV.
indicates chemicals that are regulated as drinking water contaminants under the NPDWRs.
Chemical Name
CASRN
% of
Disclo-
sures3
Chronic oral RfVsb
OSFsd
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect0
OSF
(per mg/kg-
day)
Source
of OSF
Methanol
67-56-1
73%
2
IRIS
Extra cervical ribs
Ethylene glycol
107-21-1
47%
2
IRIS
Kidney toxicity
Propargyl alcohol
107-19-7
33%
0.002
IRIS
Renal and
hepatotoxicity
2-Butoxyethanol
111-76-2
23%
0.1
IRIS
Hemosiderin
deposition in liver
(inhalation study)
Naphthalene
91-20-3
19%
0.02
IRIS
Decreased mean
terminal body
weight in males
9-48
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
% of
Disclo-
sures3
Chronic oral RfVsb
OSFsd
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect0
OSF
(per mg/kg-
day)
Source
of OSF
1,2,4-
Trimethylbenzene
95-63-6
13%
0.01
IRIS
Decreased pain
sensitivity
Formic acid
64-18-6
11%
0.9
PPRTV
Reproductive
Toxicity
N,N-
Dimethylformamide
68-12-2
9%
0.1
PPRTV
Increase in ALT
enzyme and liver
weight
Benzyl chloride
100-44-7
6%
0.002
PPRTV
Cardiotoxicity
0.17
IRIS
1,2-Propylene glycol
57-55-6
4%
20
PPRTV
Reduced RBC
counts and
hyperglycemia
Xylenes*
1330-20-7
2%
0.2
IRIS
Decreased body
weight, increased
mortality
D-Limonene
5989-27-5
2%
0.1
CICAD
Increased liver
weight
1-Butanol
71-36-3
1%
0.1
IRIS
Hypoactivity and
ataxia
Toluene*
108-88-3
0.7%
0.08
IRIS
Increased kidney
weight
Bis(2-chloroethyl)
ether
111-44-4
0.7%
1.1
IRIS
2-(2-
Butoxyethoxy)ethan
ol
112-34-5
0.6%
0.03
PPRTV
Changes in red
blood cells (RBC)
1,3,5-
Trimethylbenzene
108-67-8
0.5%
0.01
IRIS
Decreased pain
sensitivity
Cumene
98-82-8
0.5%
0.1
IRIS
Increased average
kidney weight in
female rats
Iron
7439-89-6
0.4%
0.7
PPRTV
Adverse
gastrointestinal
effects
1,2,3-
Trimethylbenzene
526-73-8
0.4%
0.01
IRIS
Decreased pain
sensitivity
9-49
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
% of
Disclo-
sures3
Chronic oral RfVsb
OSFsd
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect0
OSF
(per mg/kg-
day)
Source
of OSF
Phenol
108-95-2
0.4%
0.3
IRIS
Decreased
maternal weight
gain
Ethylbenzene*
100-41-4
0.4%
0.1
IRIS
Liver and kidney
toxicity
0.011
CalEPA
1,4-Dioxane
123-91-1
0.3%
0.03
IRIS
Liver and kidney
toxicity
0.1
IRIS
Acetone
67-64-1
0.2%
0.9
IRIS
Nephropathy
Boron
7440-42-8
0.05%
0.2
IRIS
Decreased fetal
weight
o-Xylene*
95-47-6
0.05%
0.2
ATSDR
Neurotoxicity
Acetophenone
98-86-2
0.04%
0.1
IRIS
General toxicity
Quinoline
91-22-5
0.02%
3
IRIS
Dichloromethane*
75-09-2
0.02%
0.006
IRIS
Hepatic effects
(hepatic
vacuolation, liver
foci)
0.002
IRIS
Trimethylbenzene
25551-13-7
0.01%
0.01
IRIS
Decreased pain
sensitivity
Benzene*
71-43-2
0.01%
0.004
IRIS
Decreased
lymphocyte count
0.015-0.055
IRIS
Bisphenol A
80-05-7
0.01%
0.05
IRIS
Reduced mean
body weight
Aluminum
7429-90-5
0.003%
1
PPRTV
Neurotoxicity
Hydrazine
302-01-2
0.003%
3
IRIS
Chlorobenzene*
108-90-7
0.003%
0.02
IRIS
Histopathologic
changes in liver
Arsenic*
7440-38-2
0.0003
IRIS
Hyperpigmentation
and vascular
complications
1.5
IRIS
Acrolein
107-02-8
0.0005
IRIS
Decreased survival
Chromium (VI)
18540-29-9
0.003
IRIS
None reported
0.5
CalEPA
Tributyl phosphate
126-73-8
0.01
PPRTV
Occasional
salivation
0.009
PPRTV
9-50
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
% of
Disclo-
sures3
Chronic oral RfVsb
OSFsd
RfV
(mg/kg-
day)
Source
of RfV
Critical Effect0
OSF
(per mg/kg-
day)
Source
of OSF
Di(2-ethylhexyl)
phthalate*
117-81-7
0.02
IRIS
Increased relative
liver weight
0.014
IRIS
Chlorine
7782-50-5
0.1
IRIS
No adverse effect
level
p-Xylene*
106-42-3
0.2
ATSDR
Neurotoxicity
Zinc
7440-66-6
0.3
IRIS
Decreases in
erythrocyte Cu, Zn-
superoxide
dismutase (ESOD)
activity in humans
Lead*
7439-92-1
0.5
Hg/daye
CalEPA
Reproductive
toxicity
0.0085
CalEPA
Chromium (III)
16065-83-1
1.5
IRIS
CASRN = Chemical Abstract Service Registry Number; RfV = Reference value; IRIS = Integrated Risk Information System; PPRTV =
Provisional Peer-Reviewed Toxicity Value; HHBP = Human Health Benchmarks for Pesticides; ATSDR = Agency for Toxic
Substances and Disease Registry; CalEPA = California Environmental Protection Agency; CICAD = Concise International Chemical
Assessment Documents; QSAR = Quantitative structure-activity relationship; TOPKAT = Toxicity Prediction by Komputer Assisted
Technology; ACToR = EPA's Aggregated Computational Toxicology Online Resource
a The FracFocus frequency of use data presented in this chapter is based on 35,957 FracFocus disclosures that were
deduplicated, within the study time period (January 1, 2011 to February 28, 2013), and with ingredients that have a valid
CASRN.
b Reference value (RfV): An estimate of an exposure for a given duration to the human population (including susceptible
subgroups) that is likely to be without an appreciable risk of adverse health effects over a lifetime. RfVs considered in this
analysis include chronic oral reference doses (RfDs) from IRIS, PPRTV, and HHBP; chronic oral minimal risk levels (MRLs) from
ATSDR; maximum allowable daily levels (MADLs) from CalEPA; and tolerable daily intake (TDI) from CICAD. See Section 9.4.1.
c Critical effect: The first adverse effect, or its known precursor, that occurs to the most sensitive species as the dose rate of an
agent increases.
d Oral slope factor (OSF): An upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a lifetime
oral exposure to an agent. This estimate, usually expressed in units of proportion (of a population) affected per mg/kg-day, is
generally reserved for use in the low dose region of the dose response relationship, that is, for exposures corresponding to risks
less than 1 in 100. OSFs considered in this analysis include values from IRIS, PPRTV, HHBP, and CalEPA. See Section 9.4.1.
0 CalEPA MADLs are in units of [ag/day, while all other chronic oral RfVs in this table are in units of mg/kg-day.
9.6 Hazard Evaluation of Selected Subsets of Hydraulic Fracturing Chemicals
Using Multi-Criteria Decision Analysis (MCDA): Integrating Toxicity,
Occurrence, and Physicochemical Data
Based on the information presented in Section 9.5, it is clear that there are a variety of chemicals
used in hydraulic fracturing fluids or detected in produced water that are known to be hazardous to
human health. However, there are gaps in our understanding of the potential for human exposure
to these chemicals. Although there are subsurface and surface pathways by which these chemicals
9-51
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
may be introduced into drinking water resources—including spills, leaks, mechanical integrity
failures, intersection of the fracture network with groundwater, or discharge of wastewater, as
described in previous chapters of this report—there are significant limitations associated with the
publicly available data on these potential impacts, and the potential for human exposure has not
been systematically characterized. This makes it difficult to determine which chemicals are of the
greatest concern for human exposure in drinking water, and creates a challenge for hazard
evaluation.
Although exposure assessment data are limited, some of the chemicals identified by EPA have other
data available that might provide preliminary insight into relative hazard potential. This includes
data on toxicity, frequency of use in hydraulic fracturing fluids, detected concentrations in
produced water, and data on physicochemical properties. By integrating these types of data, we can
place the severity of potential impacts (i.e., the toxicity of specific chemicals) into the context of
factors that affect the likelihood of impacts (i.e., frequency of use, environmental fate and
transport).
Multi-criteria decision analysis (MCDA) is one possible approach that can be used to facilitate data
integration. MCDA is a well-established decision support tool, which is used to integrate multiple
lines of evidence to develop an overall ranking or classification (Hristozov etal.. 2014: Mitchell et
al.. 2013b: Huang etal.. 2011: Linkov etal.. 2011). Using MCDA, a problem is approached by
dividing it into smaller criteria that need to be evaluated; the criteria are each analyzed
individually, and then combined to provide an integrated evaluation. This approach is structured
yet flexible, and offers a transparent means of combining information to provide weight of evidence
and insight into a complex problem. MCDA has gained increasing popularity as an environmental
decision-making tool (Huang etal.. 2011). A recent publication by Yost etal. (In Press) described
the use of an MCDA framework to evaluate the hazard potential of chemicals associated with
hydraulic fracturing.
Here, to demonstrate one possible method for exploring the potential hazards of these chemicals,
we use an adaptation of the MCDA framework developed by Yost et al. fin Press! to analyze and
rank selected subsets of chemicals that have data available.1 Chemicals were assigned scores based
on toxicity, occurrence, and physicochemical properties that describe transport in water. These
scores were then combined to develop a relative ranking of chemicals based on hazard potential.
The MCDA scores provide a preliminary evaluation of hazard potential, and serve as a qualitative
metric for making comparison between chemicals when exposure assessment data is limited or
unavailable. This analysis is not intended to define whether or not a chemical will present a human
health hazard or indicate that one chemical is safer than another, and should not be used in place of
1 Yost etal. fin Press) used the MCDA framework to analyze and rank the hazards of chemicals used in hydraulic
fracturing fluids, using data from the FracFocus 1.0 project database as the metric of occurrence. This chapter uses that
same framework for the analysis of chemicals used in hydraulic fracturing fluids. For chemicals detected in produced
water, this chapter modifies the MCDA framework by using measured concentration in produced water as the metric of
occurrence.
9-52
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
site-specific data on chemical exposures. An overview of the MCDA framework and selection of
chemicals for inclusion in the MCDA is described below.
9.6.1 Overview of the MCDA Framework for Hazard Evaluation
The MCDA framework employed in this chapter was designed specifically to fit the scope of EPA's
hydraulic fracturing study (Yost et al.. In Press). A basic schematic of the model is shown in Figure
9-5, and the methods for assigning scores are outlined below. Under the MCDA framework, each
chemical was assigned three scores:
1. A Toxicity Score;
2. An Occurrence Score; and
3. A Physicochemical Properties score.
The three scores were each standardized based on the highest and lowest respective score within
the given subset of chemicals, and then summed to develop a Total Hazard Potential Score for each
chemical. The Total Hazard Potential Scores reflect a relative ranking of each chemical within the
given subset of chemicals, and offer a means of making comparisons between chemicals.
Volatility
Score
Persistence
Score
Mobility
Score
Toxicity Score
(Noricancer or Cancer)
Physicochemical
Properties Score
Occurrence Score
Total Hazard Potential Score
Figure 9-5. Overview of the MCDA framework for hazard evaluation.
Source: Yost et al. (In Press).
9.6.2 Selection of Chemicals for Hazard Evaluation in the MCDA Framework
From the overall list of 1,606 chemicals identified in this assessment, subsets of chemicals were
selected for hazard evaluation in the MCDA framework if they had sufficient data for inclusion,
9-53
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
using an adaptation of the criteria outlined by Yost et al. fin Press). Specifically, chemicals were
selected if they had the following information available:
1. Had a chronic oral RfV or OSF from a US federal source (IRIS, PPRTV, ATSDR, HHBP);
2. Had available data on frequency of use in hydraulic fracturing fluids (data available from
the EPA FracFocus 1.0 project database) or measured concentrations in produced water
(data available from Appendix E)1; and
3. Had data on physicochemical properties available from EPI Suite.
The rationale for applying these criteria is as follows:
1. Federal toxicity values generally undergo more extensive peer review compared to other
sources of toxicity values, and are based on the best available scientific information. For
this reason, EPA generally prefers to apply RfVs and OSFs from US federal sources for
human health risk assessment
2. Data on frequency of use (in hydraulic fracturing fluids) or measured concentration (in
produced water) provide a metric to help assess the likelihood of chemical occurrence in
the hydraulic fracturing water cycle.
3. Information on physicochemical properties enables estimation of the likelihood a
chemical will be transported in water.
Chemicals used in hydraulic fracturing fluids and chemicals detected in produced water were
evaluated separately using the MCDA framework. To explore the different types of toxicity values
identified by EPA, two versions of the MCDA were performed on each of these subsets of chemicals:
a noncancer MCDA, in which the Toxicity Score is calculated using chronic oral RfVs; and a cancer
MCDA, in which the Toxicity Score is calculated using OSFs. For chemicals used in hydraulic
fracturing fluids, the noncancer MCDA was repeated for specific subsets of chemicals used in three
states that have a significant amount of hydraulic fracturing activity: Texas, Pennsylvania, and
North Dakota. Thus, seven iterations of the MCDA were performed: 1-4) noncancer MCDAs for
chemicals used in hydraulic fracturing fluids on a national or state-specific basis, 5) a cancer MCDA
for chemicals used in hydraulic fracturing fluids, 6) a noncancer MCDA for chemicals detected in
produced water, and 7) a cancer MCDA for chemicals detected in produced water.
In total, 42 chemicals used in hydraulic fracturing fluid and 29 chemicals detected in produced
water had sufficient information available for inclusion in noncancer MCDAs (Figure 9-6), while 10
chemicals used in hydraulic fracturing fluid and 7 chemicals detected in produced water had
sufficient information available for inclusion in cancer MCDAs (Figure 9-7).
1 Chemicals in produced water were considered for the MCDA if they had average or median measured concentrations
from any of the tables in Appendix E. Chemicals with only a maximum or minimum concentration listed in Appendix E
were not considered for the MCDA.
9-54
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
1,084 chemicals used in hydraulic fracturing fluid: 599 chemicals detected in produced water:
chemicals with chronic oral RfV
(US federal source)
chemicals with chronic oral RfV
(US federal source)
Chemicals selected for
noncancer MCDA
42
455
chemicals with
physicochemical
property data
688
chemicals with
frequency of use data
(from FracFocus) t
521
chemicals with
physicochemical
property data
175
chemicals with
measured
concentrations
Figure 9-6. The subsets of chemicals selected for hazard evaluation using the noncancer MCDA framework included 42 chemicals
used in hydraulic fracturing fluids and 29 chemicals detected in produced water.
For chemicals used in hydraulic fracturing fluids, subsets of these chemicals were also considered in state-specific analyses for Texas (36 chemicals),
Pennsylvania (20 chemicals), and North Dakota (21 chemicals).
9-55
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
1,084 chemicals used in hydraulic fracturing fluid: 599 chemicals detected in produced water:
chemicals with OSF
(US federal source)
chemicals with
physicochemical
property data
chemicals with
frequency of use data
(from FracFocus) J
chemicals with OSF
(US federal source)
chemicals with
physicochemical
property data
chemicals with
measured
concentrations
Chemicals selected for
cancer MCDA
Figure 9-7. The subsets of chemicals selected for hazard evaluation using the cancer MCDA framework included 10 chemicals used
in hydraulic fracturing fluids, and 7 chemicals detected in produced water.
9-56
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.6.3 Calculation of MCDA Scores
For each iteration of the MCDA, chemicals were assigned scores based on toxicity, occurrence, and
physicochemical properties according to the protocol outlined by Yost et al. (In Press). These scores
were then standardized to the highest and lowest score within the given subset of chemicals, and
then summed to determine a total score and relative ranking for each chemical. The methods used
to assign each score and calculate a total score are outlined below.
9.6.3.1 Toxicity Score (Noncancer MCDA)
For each noncancer MCDA, Toxicity Scores were calculated based on chronic oral RfVs from US
federal sources (IRIS, PPRTV, ATSDR, and HHBP). If a chemical had a chronic oral RfV available
from more than one of these sources, a single value was selected in this order, as described in
Section 9.4: HHBP (pesticides), IRIS, PPRTV, ATSDR. Toxicity Scores for the noncancer MCDA were
then assigned based on a relative ranking. Within each suite of chemicals considered in this analysis
(chemicals used in hydraulic fracturing fluids, or chemicals detected in produced water), RfVs were
ranked based on quartiles, and each chemical was assigned a Toxicity Score of 1 to 4 (Table 9-11).
Chemicals in the lowest quartile received the highest Toxicity Score, as these chemicals have lower
RfVs than other chemicals (i.e., may have lower thresholds for toxicity).
9.6.3.2 Toxicity Score (Cancer MCDA)
For each cancer MCDA, Toxicity Scores were calculated based on OSFs from US federal sources
(IRIS, PPRTV, and HHBP). If a chemical had an OSF available from more than one of these sources, a
single value was selected in this order, as described in Section 9.4: HHBP (pesticides), IRIS, PPRTV.
Toxicity Scores for the cancer MCDA were assigned based on a relative ranking. Within each suite of
chemicals considered in this analysis (chemicals used in hydraulic fracturing fluids, or chemicals
detected in produced water), OSFs were ranked based on quartiles, and each chemical was assigned
a Toxicity Score of 1 to 4 (Table 9-11). Chemicals in the highest quartile received the highest
Toxicity Score, as these chemicals have higher OSFs than other chemicals (i.e., are associated with a
higher increased risk of cancer per unit of exposure).
9.6.3.3 Occurrence Score
For each of the noncancer and cancer MCDAs, an Occurrence Score was calculated based on the
frequency or concentration at which each chemical was reported within the hydraulic fracturing
water cycle. For chemicals used in hydraulic fracturing fluids, the Occurrence Score was based on
the number of well disclosures for each chemical in the EPA FracFocus 1.0 project database. For
chemicals detected in produced water, the Occurrence Score was based on the average or median
measured concentration reported in Appendix E. If an average or median concentration of a
chemical was reported by multiple studies in Appendix E, the highest of these reported average or
median concentrations was used for this calculation. Once a value was determined for each
chemical, Occurrence Scores were then assigned based on a relative ranking. Within each suite of
chemicals considered in this analysis (chemicals used in hydraulic fracturing fluids, or chemicals
9-57
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
detected in produced water), chemical occurrence was ranked based on quartiles, with each
chemical assigned an Occurrence Score of 1 to 4 (Table 9-11).
9.6.3.4 Physicochemical Properties Score
For each of the noncancer and cancer MCDAs, a Physicochemical Properties Score was calculated
based upon inherent physicochemical properties that describe the likelihood that a chemical will be
transported in water. The total Physicochemical Properties Score was calculated as the sum of
three subcriteria scores: a Mobility Score, a Volatility Score, and a Persistence Score. The Mobility
Score was assessed based upon three physicochemical properties that describe chemical solvency
in water: the octanol-water partition coefficient (Kow), the soil adsorption coefficient [Koc), and
aqueous solubility. The Volatility Score was assessed based on the Henry's law constant, which
describes partitioning of a chemical between water and air. The Persistence Score was assessed
based on estimated half-life in water, which describes how long a chemical will remain in water
before it is degraded.
For input into the MCDA, experimentally measured physicochemical property values (provided in
EPI Suite) were used whenever available. Otherwise, estimated values from EPI Suite were used. To
classify these values and assign a score, these numerical values were compared against threshold
values (Table 9-11). Each chemical was assigned a Mobility Score, Volatility Score, and Persistence
Score (each on a scale of 1 to 4), which were then summed to calculate the Physicochemical
Properties Score. The threshold values in Table 9-11 are based upon previously published values
employed by existing exposure assessment models, including the EPA's Design for the Environment
Alternatives Assessment Criteria for Hazard Evaluation (U.S. EPA. 2011b). the EPA's Pollution
Prevention (P2) Framework (U.S. EPA. 2012i). and a peer-reviewed publication by Mitchell et al.
f2013bl More details on the Physicochemical Properties Score calculation are provided in the
Chapter 9 Annex, Section 9.8.1.
9.6.4 Total Hazard Potential Score
Within each iteration of the MCDA, the three criteria scores (Toxicity, Occurrence, Physicochemical
Properties) were each standardized to the dataset by scaling to the highest and lowest respective
score within the given subset of chemicals. The following equation was used:
Sx_final — (Sx — Smin) / (Smax — Smin)
in which Sx is the raw score for a particular chemical, Smax is the highest observed raw score within
the set of chemicals, and Smin is the lowest observed raw score within the set of chemicals. Sx final is
the standardized score for the chemical. Each standardized score (Toxicity, Occurrence, or
Physicochemical Properties) falls on a scale of 0 to 1, and represents a relative ranking within the
given subset of chemicals.
The standardized Toxicity Score, Occurrence Score, and Physicochemical Properties Score were
summed to calculate a Total Hazard Potential Score for each chemical. The Total Hazard Potential
Scores fall on a scale of 0 to 3, with higher scores indicating chemicals that may be more likely to
9-58
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
affect drinking water resources. Examples of the Total Hazard Potential Score calculation can be
found in the Chapter 9 Annex, Section 9.8.2.
Table 9-11. Thresholds used for developing the Toxicity Score, Occurrence Score, and
Physicochemical Properties Score in this MCDA framework.
Adapted from Yost et al. (In Press).
Criteria
Sub-criteria
Value
Score
1
2
3
4
Toxicity
(Noncancer MCDA)
NA
Chronic oral RfV
(mg/kg-day)
>3rd
quartile
>2nd quartile to
<3rd quartile
>lst quartile to
<2nd quartile
lst quartile to
<2nd quartile
>2nd quartile to
<3rd quartile
>3rd
quartile
Occurrence
NA
Frequency of
use (% of
disclosures in
EPA's FracFocus
1.0 project
database)
or
Measured
concentration in
produced water
(Hg/L; Appendix
E)
lst quartile to
<2nd quartile
>2nd quartile to
<3rd quartile
>3rd
quartile
Physico-chemical
Properties
Mobility
Log Kow
>5
>3 to <5
>2 to <3
<2
Log Koc
>4.4
>3.4 to <4.4
>2.4 to <3.4
<2.4
Aqueous
solubility (mg/L)
<0.1
>0.1 to <100
>100 to <1000
>1000
Volatility
Henry's law
constant
>10_1
>10 3 to <101
>10"5 to <10"3
<10"5
Persistence
Half-life in
water (days)
<16
>16 to <60
>60 to <180
>180
9.6.5 MCDA Results
For each iteration of the MCDA, we first present the data used for input into the MCDA, including
data on toxicity, occurrence, and physicochemical properties. We then present the results of each
MCDA, which show a relative ranking of chemicals based on integration of these data. Lastly, we
discuss the key limitations of this MCDA approach, which is intended as a preliminary analysis only.
9-59
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.6.5.1 Results: Noncancer MCDA for Chemicals Used in Hydraulic Fracturing Fluids
A total of 42 chemicals used in hydraulic fracturing fluids were evaluated in a noncancer MCDA
(Table 9-12). Chronic oral RfVs within this suite of chemicals range from 0.001-20 mg/kg-day, with
(E)-crotonaldehyde having the lowest chronic oral RfV and 1,2-propylene glycol having the highest
These RfVs were derived based on health effects including immune system effects, changes in body
weight, changes in blood chemistry, cardiotoxicity, neurotoxicity, liver and kidney toxicity, and
reproductive and developmental toxicity. The total UFs used in the derivation of these chronic oral
RfVs (Table 9-12) reflect varying degrees of confidence surrounding the data sets for these
chemicals. Three of the chemicals with the lowest chronic oral RfVs [(E)-crotonaldehyde, propargyl
alcohol, benzyl chloride] have total UFs of 3000, indicating a relatively large amount of uncertainty
in these values. Comparatively, chemicals such as benzene, acrylamide, and dichloromethane also
have low chronic oral RfVs, but with much less uncertainty reflected in the values.
Figure 9-8 presents the results of a noncancer MCDA for these 42 chemicals in hydraulic fracturing
fluids. Of these 42 chemicals, propargyl alcohol received the highest overall Total Hazard Potential
Score. Propargyl alcohol was reported in 33% of disclosures nationally in the EPA FracFocus 1.0
project database, making it one of the most widely used chemicals that was considered in this
analysis. It has physicochemical properties that are conducive to transport in water, and a low RfV.
Given these properties, propargyl alcohol received the highest overall ranking based on hazard
potential across all of the metrics that were considered in the MCDA.
Several of the other chemicals that received high Occurrence Scores also received among the
highest Total Hazard Potential Scores, including 2-butoxyethanol, naphthalene, 1,2,4-
trimethylbenzene, N,N-dimethylformamide, and formaldehyde (reported in 23%, 19%, 13%, 9%,
and 7% of disclosures, respectively). Methanol, ethylene glycol, and formic acid (73%, 47%, and
11% of disclosures, respectively) received lower Total Hazard Potential Scores as a result of having
higher RfVs. Likewise, didecyldimethylammonium chloride and dodecylbenzenesulfonic acid (8%
and 7% of disclosures, respectively) received lower Total Hazard Potential Scores as a result of
having higher RfVs and more hydrophobic properties.
The other chemicals that received high Toxicity Scores (i.e., had low chronic oral RfVs) received
moderate to high Total Hazard Potential Scores overall. Acrylamide was reported in only 1% of
disclosures, but has physicochemical properties that are very conducive to transport in water, and
therefore received one of the highest overall Total Hazard Potential Scores. 1,2,4-
Trimethylbenzene, benzyl chloride, and epichlorohydrin (13%, 6%, and 1% of disclosures in the
EPA FracFocus 1.0 project database, respectively) scored slightly lower than acrylamide with
regards to physicochemical properties. Other chemicals, including 1,2,3-trimethylbenzene, 1,3,5-
trimethylbenzene, (E)-crotonaldehyde, benzene, dichloromethane, aniline, furfural, and 2-
(Thiocyanomethylthio)benzothiazole, received lower overall scores because they are used more
infrequently (the trimethylbenzenes were reported in <1% of disclosures, and the rest reported in
<0.1% of disclosures).
9-60
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.6.5.2 Results: Noncancer MCDA for Chemicals Used in Hydraulic Fracturing Fluids (State-
specific analysis for Texas, Pennsylvania, and North Dakota)
To investigate the extent of regional differences and examine the applicability of the MCDA model
at the regional scale, we repeated the noncancer MCDA for hydraulic fracturing fluids for subsets of
chemicals used in three representative states that have a significant amount of hydraulic fracturing
activity: Texas, Pennsylvania, and North Dakota. The chemicals used in these state-specific analyses
are subsets of the chemicals used nationally, and are indicated in Table 9-12. Some of the chemicals
considered in the national analysis were not included in the state-specific analyses because they
were not disclosed to FracFocus 1.0 as used in these states.
Results are presented in Figure 9-9 (Texas), Figure 9-10 (Pennsylvania), and Figure 9-11 (North
Dakota). By comparing these results to each other and to the national noncancer MCDA (Figure
9-8), it is evident that there are some regional differences in the Total Hazard Potential Scores,
although many chemicals were commonly used and received similar overall rankings.
Methanol, ethylene glycol, and 2-butoxyethanol were among the most frequently reported
chemicals in all three state-specific analyses, while other chemicals differed distinctly between
states. For instance, propargyl alcohol was frequently reported in Texas (39% of disclosures) and
Pennsylvania (58% of disclosures), but not North Dakota (1% of disclosures). Likewise,
naphthalene was reported frequently in Texas (14% of disclosures) and North Dakota (43% of
disclosures), but not in Pennsylvania (1% of disclosures). The most toxic chemicals (occurring in
the lowest quartile of chronic oral RfVs) common among all three states include propargyl alcohol,
benzyl chloride, acrylamide, and 1,2,4-trimethylbenzene. Other chemicals receiving high Toxicity
Scores in these states include epichlorohydrine (Texas and Pennsylvania), 1,3,5-Trimethylbenzene
(Texas and Pennsylvania), 1,4-dioxane (North Dakota), naphthalene (North Dakota), benzene,
aniline, and 1,2,3-Trimethylbenzene (Texas).
Overall, in Texas, propargyl alcohol received the highest possible Total Hazard Potential Score, with
acrylamide receiving the second highest score. In Pennsylvania, propargyl alcohol also received the
highest possible Total Hazard Potential Score, with 2-butoxyethanol receiving the second highest
score. In North Dakota, 2-butoxyethanol received the highest Total Hazard Potential Score, with
naphthalene receiving the second highest score.
The results of these state-specific MCDAs support the concept presented in Chapter 5 that there is
no single hydraulic fracturing fluid formulation, and that the chemicals of most potential concern
will vary between regions or even between wells.
9-61
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-12. Data on the selected subset of chemicals in hydraulic fracturing fluids used for input into a noncancer MCDA.
Chemicals within the table are ordered from most toxic to least toxic based on chronic oral RfV.
Chemical Name
CASRN
Noncancer toxicity
(chronic oral RfV)a
% disclosures in EPA
FracFocus 1.0 project
database13
Mobility
Volatility
Persist-
ence
RfV
(mg/
kg-day)
Total
UF
Source
National
TX
PA
ND
Log
Kow
Log
Koc
Solu-
bility
(mg/L)
Henry's
Law
Constant
Half-life
in water
(days)
(E)-Crotonaldehyde
123-73-9
0.001
3000
PPRTV
0.06%
0.6
0.254
41480
1.94E-05
15
Propargyl alcohol
107-19-7
0.002
3000
IRIS
33%
39%
58%
1%
-0.38
0.28
935500
1.15E-06
15
Benzyl chloride
100-44-7
0.002
3000
PPRTV
6%
7%
5%
0.80%
2.3
2.649
1030
4.12E-04
15
Acrylamide
79-06-1
0.002
30
IRIS
1%
2%
1%
1%
-0.67
0.755
504000
1.70E-09
15
Benzene
71-43-2
0.004
300
IRIS
0.006%
0.0I-'..
0.20-1.. 0.08'.'-..
2.13
1.75
2000
5.55E-03
37.5
Epichlorohydrin
106-89-8
0.006
1000
PPRTV
1%
0.45
1
50630
3.04E-05
15
Dichloromethane
75-09-2
0.006
30
IRIS
0.02%
1.25
1.44
10950
3.25E-03
37.5
Aniline
62-53-3
0.007
1000
PPRTV
0.02%
0.05-1..
0.9
1.6
20820
2.02E-06
15
1,2,4-
Trimethylbenzene
95-63-6
0.01
300
IRIS
13%
11%
1%
25%
3.63
2.788
79.59
6.16E-03
37.5
1,3,5-
Trimethylbenzene
108-67-8
0.01
300
IRIS
0.5%
0.80%
1%
3.42
2.82
120.3
8.77E-03
37.5
1,2,3-
Trimethylbenzene
526-73-8
0.01
300
IRIS
0.4%
0.80''-..
3.66
2.8
75.03
4.36E-03
37.5
2-(Thiocyanomethyl-
thio)benzothiazole
21564-17-
0
0.01
300
HHBP
0.006%
3.3
3.528
41.67
6.49E-12
37.5
Furfural
98-01-1
0.01
3000
HHBP
0.003%
0.41
0.784
53580
3.77E-06
15
Naphthalene
91-20-3
0.02
3000
IRIS
19%
14%
1%
43%
3.3
2.96
142.1
4.40E-04
37.5
9-62
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Noncancer toxicity
(chronic oral RfV)a
% disclosures in EPA
FracFocus 1.0 project
database13
Mobility
Volatility
Persist-
ence
RfV
(mg/
kg-day)
Total
UF
Source
National
TX
PA
ND
Log
Kow
Log
Koc
Solu-
bility
(mg/L)
Henry's
Law
Constant
Half-life
in water
(days)
Chlorobenzene
108-90-7
0.02
1000
IRIS
0.003%
0.01%
2.84
2.15
400.5
3.11E-03
15
2-(2-Butoxyethoxy)
ethanol
112-34-5
0.03
3000
PPRTV
0.6%
0.40%
4%
0.56
1
71920
7.20E-09
8.67
1,4-Dioxane
123-91-1
0.03
300
IRIS
0.3%
0.50-1.. 0.80'.'-..
-0.27
0.421
213900
4.80E-06
15
1,3-Dichloropropene
542-75-6
0.03
100
IRIS
0.02%
0.0I-1..
1%
2.04
1.82
1994
3.55E-03
37.5
Bisphenol A
80-05-7
0.05
1000
IRIS
0.006%
3.32
4.576
172.7
9.16E-12
37.5
Toluene
108-88-3
0.08
3000
IRIS
0.7%
2.73
2.07
573.1
6.64E-03
15
Ethylenediamine
107-15-3
0.09
100
PPRTV
0.01%
0.027o
-2.04
1.172
1000000
1.73E-09
15
2-Butoxyethanol
111-76-2
0.1
10
IRIS
23%
27%
21%
15%
0.83
0.451
64470
1.60E-06
8.67
N,N-Dimethylform-
amide
68-12-2
0.1
1000
PPRTV
9%
10%
11%
0.60%
-1.01
0
977900
7.39E-08
15
Didecyldimethylam-
monium chloride
7173-51-5
0.1
100
HHBP
8%
7%
12%
0.05%
4.66
5.546
0.9
6.85E-10
15
1-Butanol
71-36-3
0.1
1000
IRIS
1%
2%
0.70-'..
0.88
0.5
76700
8.81E-06
8.67
Cumene
98-82-8
0.1
1000
IRIS
0.5%
0.80%
1%
3.66
2.844
75.03
1.15E-02
15
Ethylbenzene
100-41-4
0.1
1000
IRIS
0.4%
0.50%
0.10%
3.15
2.23
228.6
7.88E-03
15
Acetophenone
98-86-2
0.1
3000
IRIS
0.04%
0.04%
1.58
1.8
4484
1.04E-05
15
Formaldehyde
50-00-0
0.2
100
IRIS
7%
8%
4%
8%
0.35
0
57020
3.37E-07
15
Xylenes
1330-20-7
0.2
1000
IRIS
2%
3%
1%
0.20%
3.2
2.25
207.2
7.18E-03
15
9-63
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Noncancer toxicity
(chronic oral RfV)a
% disclosures in EPA
FracFocus 1.0 project
database13
Mobility
Volatility
Persist-
ence
RfV
(mg/
kg-day)
Total
UF
Source
National
TX
PA
ND
Log
Kow
Log
Koc
Solu-
bility
(mg/L)
Henry's
Law
Constant
Half-life
in water
(days)
o-Xylene
95-47-6
0.2
30
ATSDR
0.05%
0.1'--..
0.80'.'-.. 0.05-1..
3.12
2.25
224.1
5.18E-03
15
Phenol
108-95-2
0.3
300
IRIS
0.4%
1.46
1.9
26160
3.33E-07
15
2-Methyl-l-propanol
78-83-1
0.3
1000
IRIS
0.3%
4%
0.76
0.465
97120
9.78E-06
15
Dodecylbenzenesul-
fonic acid
27176-87-
0
0.5
100
HHBP
7%
10%
2%
8%
4.71
4.066
0.8126
6.27E-08
15
Formic acid
64-18-6
0.9
300
PPRTV
11%
14%
8%
11%
-0.54
0
955200
1.67E-07
8.67
Ethyl acetate
141-78-6
0.9
1000
IRIS
0.4%
0.70%
0.73
0.747
29930
1.34E-04
15
Acetone
67-64-1
0.9
1000
IRIS
0.2%
0.02%
1%
-0.24
0.374
219900
3.50E-05
15
Methanol
67-56-1
2
100
IRIS
73%
80%
69%
54%
-0.77
0.44
1000000
4.55E-06
8.67
Ethylene glycol
107-21-1
2
100
IRIS
47%
60%
35%
37%
-1.36
0
1000000
6.00E-08
8.67
Hexanedioic acid
124-04-9
2
300
PPRTV
0.70%
1%
0.04'.'-..
0.08
1.386
167300
4.71E-12
8.67
Benzoic acid
65-85-0
4
1
IRIS
0.06%
0.10%
1.87
1.5
2493
3.81E-08
15
1,2-Propylene glycol
57-55-6
20
300
PPRTV
4%
4%
8%
8%
-0.92
0.36
811100
1.29E-08
8.67
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; PPRTV = Provisional Peer Reviewed Toxicity Values; ATSDR = Agency for Toxic
Substances and Disease Registry; HHBP = Human Health Benchmarks for Pesticides; K0w = octanol-water partitioning coefficient; K0c = soil adsorption coefficient
a Reference value (RfV): An estimate of an exposure for a given duration to the human population (including susceptible subgroups) that is likely to be without an appreciable
risk of adverse health effects over a lifetime. RfVs considered in the MCDA include chronic oral reference doses (RfD) from IRIS, PPRTV, and HHBP; and chronic oral minimal risk
levels (MRLs) from ATSDR.
bThe FracFocus frequency of use data presented in this chapter is based on 35,957 FracFocus disclosures that were deduplicated, within the study time period (January 1, 2011
to February 28, 2013), and with ingredients that have a valid CASRN.
9-64
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Noncancer MCDA: Chemicals Used in Hydraulic Fracturing Fluids
Toxicity Score (Noncancer) ¦ Occurrence Score Physicochemical Properties Score
Propargyl alcohol
Acrylamide
2-Butoxyethanol
Naphthalene
1.2.4-Tri methy Ibenze ne
N,N-Dimethylformamide
Formaldehyde
Benzyl chloride
Epichlorohydrin
2-(2-Butoxyethoxy)ethanol
Methanol
Ethylene glycol
Formic acid
1-Butanol
1.3.5-T ri methylbenze ne
Aniline
Furfural
1,4-Dioxane
(E)-Crotonaldehyde
Didecyldimethylamrnonium chloride
1.2-Propylene glycol
Hexanedioicacid
Toluene
1,2,3-Tri methylbenze ne
Dichloromethane
Ethylenediamine
Bisphenol A
Benzene
2-(Thiocyanomethylthio)benzothiazole
Dodecylbenzenesulfonic acid
Xylenes
Phenol
2-Methyl-l-propanol
Benzoic acid
1.3-Dichloropropene
Ethyl acetate
Ethylbenzene
Acetone
Acetophenone
Chlorobenzene
Cumene
o-Xylene
0.00 0.50 1.00 1.50 2.00
Total Hazard Potential Score
2.50
3.00
Figure 9-8. Noncancer MCDA results for 42 chemicals used in hydraulic fracturing fluids
(national analysis), showing the Toxicity Score, Occurrence Score, and Physicochemical
Properties Score for each chemical.
Chemicals are ordered from high to iow based on Total Hazard Potential Score. See Section 9.6.4 for details on the
calculation.
9-65
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Noncancer MCDA: Chemicals Used in Hydraulic Fracturing Fluids
(Texas)
Toxicity Score (Noncancer) ¦ Occurrence Score Physicochemical Properties Score
Propargyl alcohol
Acrylamide
2-Butoxyethanol
Naphthalene
1,2,4-Trimethylbenzene
N, IM-Dimethylformamide
Benzyl chloride
Methanol
Ethylene glycol
Formic acid
Formaldehyde
1-Butanol
Aniline
1,4-Dioxane
2-(2-Butoxyethoxy)ethanol
Epichlorohydrin
1,2-Propylene glycol
Toluene
Hexanedioic acid
1,2,3-T ri methy Ibenze ne
1,3,5 -Tri methy I b e n ze ne
Phenol
Ethylenediamine
Bisphenol A
Benzene
Dodecylbenzenesulfonic acid
Didecyldimethylammonium chloride
Xylenes
Ethyl acetate
Ethylbenzene
Benzoic acid
Acetophenone
Chlorobenzene
Cumene
Acetone
o-Xylene
0.00 0.50 1.00 1.50 2.00 2.50 3.00
Total Hazard Potential Score
Figure 9-9. Noncancer MCDA results for 36 chemicals used in hydraulic fracturing fluids in
Texas (state-specific analysis), showing the Toxicity Score, Occurrence Score, and
Physicochemical Properties Score for each chemical.
Chemicals are ordered from high to iow based on Total Hazard Potential Score. See Section 9.6.4 for details on the
calculation.
9-66
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Noncancer MCDA: Chemicals Used in Hydraulic Fraturing Fluids
(Pennsylvania)
Toxicity Score (Noncancer)
¦ Occurrence Score Physicochemical Properties Score
Dodecylbenzenesulfonic acid
Xylenes
¦¦¦
Cumene
0.00
0.50 1.00 1.50 2.00 2.50 3.00
Total Hazard Potential Score
Figure 9-10. Noncancer MCDA results for 20 chemicals used in hydraulic fracturing fluids in
Pennsylvania (state-specific analysis), showing the Toxicity Score, Occurrence Score, and
Physicochemical Properties Score for each chemical.
Chemicals are ordered from high to low based on Total Hazard Potential Score. See Section 9.6.4 for details on the
calculation.
9-67
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Noncancer MCDA: Chemicals Used in Hydraulic Fracturing Fluids
(North Dakota)
Toxicity Score (Noncancer)
Occurrence Score
Physicochemicai Properties Score
2-Butoxyethanol
Naphthalene
Propargyl alcohol
Acrylamide
1,4-Dioxane
Methanol
Ethylene glycol
1,2,4-Tri m ethy I be n ze ne
Formic acid
Formaldehyde
2-Methyl -1-propanol
1-Butanol
Benzyl chloride
1,2-Propylene glycol
IM,IM-Dimethylformamide
Phenol
Acetone
Dodecylbenzenesulfonic acid
Ethylbenzene
Didecyldimethylammonium chloride
Xylenes
0.00 0.50 1.00 1.50 2.00
Total Hazard Potential Score
2.50
3.00
Figure 9-11. Noncancer MCDA results for 21 chemicals used in hydraulic fracturing fluids in
North Dakota (state-specific analysis), showing the Toxicity Score, Occurrence Score, and
Physicochemicai Properties Score for each chemical.
Chemicals are ordered from high to iow based on Total Hazard Potential Score. See Section 9.6.4 for details on the
calculation.
9-68
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.6.5.3 Results: Cancer MCDA for Chemicals Used in Hydraulic Fracturing Fluids
A total of 10 chemicals used in hydraulic fracturing fluids were evaluated in a cancer MCDA (Table
9-13). OSFs for these chemicals ranged from 0.002 to 3 per mg/kg-day, with quinoline having the
highest OSF, and dichloromethane having the lowest Benzene is the only one of these chemicals
that is classified as a known human carcinogen by at least one of the sources in Table 9-1, while the
other chemicals in this subset are classified as probable carcinogens in humans (Appendix Table G-
le).
Figure 9-12 presents the results from the cancer MCDA for chemicals used in hydraulic fracturing
fluids. Of the 10 chemicals that were considered in this analysis, acrylamide received the highest
Total Hazard Potential Score. Acrylamide has an OSF of 0.5 per mg/kg-day, which is one of the
higher OSFs in this suite of chemicals, and has physicochemical properties that are highly conducive
to transport in water. Acrylamide was reported in 1% of disclosures nationally in the EPA
FracFocus 1.0 project database. This nevertheless places acrylamide in the top quartile in terms of
frequency of use, as none of the chemicals within this subset were used with great frequency on a
national basis.
Bis(2-chloroethyl)ether and quinoline, which are the two most potent carcinogens considered in
the analysis and received high Toxicity Score, received the second and third highest Total Hazard
Potential Scores within this suite of chemicals. Bis(2-chloroethyl)ether was reported in 0.7% of
disclosures, while quinoline was reported in 0.02% of disclosures. Both are expected to be readily
transported in water.
In addition to acrylamide, the other two chemicals receiving high Occurrence Scores were benzyl
chloride and epichlorohydrin (6% and 1% of disclosures, respectively). These two chemicals both
received moderate Total Hazard Potential Scores. Benzyl chloride has an OSF of 0.17 per mg/kg-
day, while epichlorohydrine has an OSF of 0.0099 per mg/kg-day. Both received lower
Physicochemical Properties Scores relative to other chemicals in this analysis, due in part to
volatility.
9-69
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-13. Data on the selected subset of chemicals in hydraulic fracturing fluids used for input into a cancer MCDA.
Chemicals within the table are ordered from most potent to least potent based on OSF.
Chemical Name
CASRN
Cancer-specific
toxicity (OSF)a
% disclosures in
EPA FracFocus
1.0 project
database13
Mobility
Volatility
Persistence
OSF(per
mg/kg-
day)
Source
National
Log Kow
Log Koc
Solubility
(mg/L)
Henry's
Law
Constant
Half-life in
water (days)
Quinoline
91-22-5
3
IRIS
0.02%
2.03
3.1
1711
1.67E-06
15
Bis(2-chloroethyl)
ether
111-44-4
1.1
IRIS
0.7%
1.29
1.88
6435
1.70E-05
37.5
Acrylamide
79-06-1
0.5
IRIS
1%
-0.67
0.755
504000
1.70E-09
15
Benzyl chloride
100-44-7
0.17
IRIS
6%
2.3
2.649
1030
4.12E-04
15
1,4-Dioxane
123-91-1
0.1
IRIS
0.3%
-0.27
0.421
213900
4.80E-06
15
Benzene
71-43-2
0.015-
0.055°
IRIS
0.006%
2.13
1.75
2000
5.55E-03
37.5
1,3-Dichloropropene
542-75-6
0.05
IRIS
0.02%
2.04
1.82
1994
3.55E-03
37.5
Epichlorohydrin
106-89-8
0.0099
IRIS
1%
0.45
1
50630
3.04E-05
15
Aniline
62-53-3
0.0057
IRIS
0.02%
0.9
1.6
20820
2.02 E-06
15
Dichloromethane
75-09-2
0.002
IRIS
0.02%
1.25
1.44
10950
3.25E-03
37.5
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; K0w = octanol-water partitioning coefficient; K0c = soil adsorption coefficient
a Oral slope factor (OSF): An upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a lifetime oral exposure to an agent. This estimate, usually
expressed in units of proportion (of a population) affected per mg/kg-day, is generally reserved for use in the low dose region of the dose response relationship, that is, for
exposures corresponding to risks less than 1 in 100. OSFs considered in the MCDA include values from IRIS, PPRTV, and HHBP.
bThe FracFocus frequency of use data presented in this chapter is based on 35,957 FracFocus disclosures that were deduplicated, within the study time period (January 1, 2011
to February 28, 2013), and with ingredients that have a valid CASRN.
c IRIS lists the OSF for benzene as a range from 0.015 to 0.055 per mg/kg-day. For input into the MCDA, we used the high end of this range (0.055 per mg/kg-day).
9-70
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Cancer MCDA: Chemicals Used in Hydraulic Fracturing Fluids
¦ Toxicity Score (Cancer) ¦ Occurrence Score Physicochemical Properties Score
Acrylamide
Bis(2-chloroethyl) ether |
Quinoline
1,4-Dioxane
Benzyl chloride
Aniline |
Epichlorohydrin |
1,3-Dichloropropene
Benzene |
Dichloromethane
0
Figure 9-12. Cancer MCDA results for 10 chemicals used in hydraulic fracturing fluids, showing
the Toxicity Score, Occurrence Score, and Physicochemical Properties Score for each
chemical.
Chemicals are ordered from high to low based on Total Hazard Potential Score. See Section 9.6.4 for details on the
calculation.
9.6.5.4 Results: Noncancer MCDA for Chemicals in Produced Water
A total of 29 chemicals detected in produced water were evaluated in a noncancer MCDA (Table
9-14). Of these 29 chemicals, 13 were also included in the noncancer MCDA for hydraulic fracturing
fluids. Chronic oral RfVs within this suite of chemicals range from 0.001 to 0.9 mg/kg-day, with
pyridine having the lowest chronic oral RfV, and acetone having the highest Chronic oral exposure
to these chemicals may induce a variety of adverse outcomes, including immune system effects,
changes in body weight, changes in blood chemistry, pulmonary toxicity, neurotoxicity, liver and
kidney toxicity, and reproductive and developmental toxicity. The total UFs used in the derivation
of these chronic oral RfVs (Table 9-14) reflect varying degrees of confidence surrounding the data
sets for these chemicals.
Figure 9-13 presents the results of a noncancer MCDA for these 29 chemicals detected in produced
water. Benzene, pyridine, and naphthalene received the highest Total Hazard Potential Scores,
followed by 2-methylnaphthalene. These four chemicals all received high Toxicity Scores and high
Occurrence Scores (with maximum average concentrations of 1500 ng/L, 413 |ig/L, 238 |ig/L, and
1362 |ig/L in Barnett, Marcellus, or Powder River Basin produced water, respectively), but received
moderate to low Physicochemical Property Scores.
Total Hazard Potential Score
9-71
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-14. Data on the selected subset of chemicals detected in produced water used for input into a noncancer MCDA.
Chemicals within the table are ordered from most toxic to least toxic based on chronic oral RfV.
Chemical Name
CASRN
Noncancer toxicity
(chronic oral RfV)a
Occurrence
(concentration in
produced water)b
Mobility
Volatility
Persist-
ence
RfV
(mg/kg-
day)
Total UF
Source
Average or
Median
Cone. (ng/L)
Reference
Log Kow
Log Koc
Solubility
(mg/L)
Henry's
Law
Constant
Half-life
in water
(days)
Pyridine
110-86-1
0.001
1000
IRIS
413
Table E-ll
0.65
1.6
729800
1.10E-05
15
Benzene
71-43-2
0.004
300
IRIS
1500
Table E-13
2.13
1.75
2000
5.55E-03
37.5
2-Methylnaphthalene
91-57-6
0.004
1000
IRIS
1362
Table E-ll
3.86
3.6
41.42
5.18E-04
15
1,2,4-
Trimethylbenzene
95-63-6
0.01
300
IRIS
173
Table E-ll
3.63
2.788
79.59
6.16E-03
37.5
1,3,5-
Trimethylbenzene
108-67-8
0.01
300
IRIS
59
Table E-ll
3.42
2.82
120.3
8.77E-03
37.5
Chloroform
67-66-3
0.01
1000
IRIS
28
Table E-ll
1.97
1.6
2096
3.67E-03
37.5
Tributyl phosphate
126-73-8
0.01
1000
PPRTV
0.26
Table E-12
4
3.371
7.355
1.41E-06
8.67
Naphthalene
91-20-3
0.02
3000
IRIS
238
Table E-ll
3.3
2.96
142.1
4.40E-04
37.5
Di(2-ethylhexyl)
phthalate0
117-81-7
0.02
1000
IRIS
210
Table E-ll
7.6
4.94
0.001132
2.70E-07
15
Chlorobenzened
108-90-7
0.02
1000
IRIS
100
Table E-13
2.84
2.15
400.5
3.11E-03
15
2,4-Dimethylphenol
105-67-9
0.02
3000
IRIS
14.5
Table E-ll
2.3
2.692
4068
9.51E-07
15
Pyrene
129-00-0
0.03
3000
IRIS
13
Table E-ll
4.88
4.9
0.2249
1.19E-05
60
1,4-Dioxane
123-91-1
0.03
300
IRIS
6.5
Table E-ll
-0.27
0.421
213900
4.80E-06
15
Fluorene
86-73-7
0.04
3000
IRIS
8.4
Table E-ll
4.1
3.614
20.13
1.59E-03
15
Fluoranthene
206-44-0
0.04
3000
IRIS
6.1
Table E-ll
5.16
4.8
0.1297
8.86E-06
60
o-Cresole
95-48-7
0.05
1000
IRIS
28.3
Table E-ll
1.95
2.486
9066
1.20E-06
15
Toluene
108-88-3
0.08
3000
IRIS
760
Table E-9
2.73
2.07
573.1
6.64E-03
15
9-72
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Chemical Name
CASRN
Noncancer toxicity
(chronic oral RfV)a
Occurrence
(concentration in
produced water)b
Mobility
Volatility
Persist-
ence
RfV
(mg/kg-
day)
Total UF
Source
Average or
Median
Cone. (ng/L)
Reference
Log Kow
Log Koc
Solubility
(mg/L)
Henry's
Law
Constant
Half-life
in water
(days)
Ethylbenzene
100-41-4
0.1
1000
IRIS
2010
Table E-13
3.15
2.23
228.6
7.88E-03
15
Carbon disulfide
75-15-0
0.1
100
IRIS
400
Table E-ll
1.94
1.337
2928
1.44E-02
15
Cumenef
98-82-8
0.1
1000
IRIS
120
Table E-ll
3.66
2.844
75.03
1.15E-02
15
Benzyl alcohol
100-51-6
0.1
1000
PPRTV
81.5
Table E-ll
1.1
1.1
41050
3.37E-07
15
Dibutyl phthalate8
84-74-2
0.1
1000
IRIS
41
Table E-ll
4.5
3.14
2.351
1.81E-06
8.67
Acetophenone
98-86-2
0.1
3000
IRIS
13
Table E-ll
1.58
1.8
4484
1.04E-05
15
Diphenylamine
122-39-4
0.1
100
HHBP
5.3
Table E-ll
3.5
2.78
63.61
2.69E-06
37.5
Xylenes
1330-20-7
0.2
1000
IRIS
360
Table E-9
3.2
2.25
207.2
7.18E-03
15
Benzyl butyl phthalate
85-68-7
0.2
1000
IRIS
34.3
Table E-ll
4.73
3.72
0.9489
1.26E-06
0.04
Phenol
108-95-2
0.3
300
IRIS
63
Table E-ll
1.46
1.9
26160
3.33E-07
15
Caprolactam
105-60-2
0.5
100
IRIS
0.75
Table E-12
0.66
1.3892
28720
2.53E-08
14508
Acetone
67-64-1
0.9
1000
IRIS
145
Table E-10
-0.24
0.374
219900
3.50E-05
15
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; PPRTV = Provisional Peer Reviewed Toxicity Values; HHBP = Human Health
Benchmarks for Pesticides; K0w = octanol-water partitioning coefficient; K0c = soil adsorption coefficient
a Reference value (RfV): An estimate of an exposure for a given duration to the human population (including susceptible subgroups) that is likely to be without an appreciable
risk of adverse health effects over a lifetime. RfVs considered in the MCDA include chronic oral reference doses (RfD) from IRIS, PPRTV, and HHBP; and chronic oral minimal risk
levels (MRLs) from ATSDR.
b From Appendix E.
c Di(2-ethylhexyl) phthalate is listed under the name bis(2-ethylhexyl) phthalate in Appendix Table E-ll.
d Chlorobenzene is listed under the name chloro-benzene in Appendix Table E-13.
0 o-Cresol is listed under the name 2-methylphenol in Appendix Table E-ll.
f Cumene is listed under the name isopropylbenzene in Appendix Table E-ll.
5 Dibutyl phthalate is listed under the name dibutyl-n-phthalate in Appendix Table E-ll.
9-73
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Noncancer MCDA: Chemicals in Produced Water
Toxicity Score (Noncancer) ¦ Occurrence Score Physicochemical Properties Score
¦1
Benzyl alcohol
Pyrene
Tributyl phosphate
Xvlenes
Phenol ¦
o-Cresol
1,4-Dioxane
Fluoranthene
Cumene
Acetone ~
Dibutyl phthalate
Acetophenone
Caprolactam
Diphenylamine
Fluorene
Benzyl butyl phthalate
0.00 0.50 1.00 1.50 2.00 2.50 3.00
Total Hazard Potential Score
Figure 9-13. Noncancer MCDA results for a subset of 29 chemicals detected in produced
water, showing the Toxicity Score, Occurrence Score, and Physicochemical Properties Score
for each chemical.
Chemicals are ordered from high to low based on Total Hazard Potential Score. See Section 9.6.4 for details on the
calculation.
9-74
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
The other chemicals that received high Toxicity Scores were 1,2,4-trimethylbenzene, 1,3,5-
trimethylbenzene, chloroform, 2,4,-dimethylphenol, tributyl phosphate, di(2-ethylhexyl) phthalate,
and chlorobenzene. These chemicals received moderate Total Hazard Potential Scores, as all were
detected at lower concentrations compared to other chemicals considered in this analysis and are
expected to have moderate transport in water.
The other chemicals that received high Occurrence Scores are ethylbenzene, toluene, xylenes, and
carbon disulfide, which were detected at maximum average concentrations of 2010 ng/L, 760 |ig/L,
360 ng/L, and 400 ng/L in Barnett, Marcellus, or Powder River Basin produced water. These
chemicals received moderate Total Hazard Potential Scores, as all have as all have higher chronic
oral RfVs relative to many of the other chemicals in the hazard evaluation, and are all expected to
have moderate transport in water relative to the other chemicals.
9.6.5.5 Results: Cancer MCDA for Chemicals in Produced Water
A total of 7 chemicals reported in produced water were evaluated in a cancer MCDA (Table 9-15).
OSFs within this suite of chemicals ranged from 7.3 to 0.0049 per mg/kg-day, with benzo(a)pyrene
having the highest OSF and N-nitrosodiphenylamine having the lowest Of these 7 chemicals,
benzene and 1,4-dioxane were also included in the cancer MCDA for chemicals used in hydraulic
fracturing fluids. Benzene and benzo(a)pyrene are both classified by at least one of the sources in
Table 9-1 as a known human carcinogen, while the other chemicals as classified as likely or
probable carcinogens in humans (Appendix G: Tables G-le and G-2e).
Figure 9-14 presents the results of a cancer MCDA for these 7 chemicals in hydraulic fracturing
fluids. Benzene and benzo(a)pyrene tied for highest Total Hazard Potential Scores. Of these,
benzene was detected at the highest average concentrations in produced water (1500 |ig/L in
Power River Basin produced water), while benzo(a)pyrene were detected at lower average
concentrations (6.7 |ig/L in Barnett shale produced water). Benzo(a)pyrine and 1,2-
diphenylhydrazine were the most potent carcinogens within this suite of chemicals and received
high Toxicity Scores.
The other chemical that received a high Occurrence Score was di(2-ethylhexyl) phthalate, which
was detected at an average concentration of 210 |ig/L in Barnett Shale produced water. It received
a moderate Total Hazard Potential Score because it is hydrophobic and not expected to be readily
transported in water.
9-75
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Table 9-15. Data on the selected subset of chemicals detected in produced water used for input into a cancer MCDA.
Chemicals within the table are ordered from most potent to least potent based on OSF.
Chemical Name
CASRN
Cancer-specific
toxicity (OSF)a
Occurrence
(concentration in
produced water)b
Mobility
Volatility
Persist-
ence
OSF(per
mg/kg-
day)
Source
of OSF
Average or
Median
Cone. (pg/L)
Reference
Log Kow
Log Koc
Solubility
(mg/L)
Henry's Law
Constant
Half-life
in water
(days)
Benzo(a)pyrene
50-32-8
7.3
IRIS
6.7
Table E-ll
6.13
5.95
0.01038
4.57E-07
60
1,2-Diphenylhydrazine
122-66-7
0.8
IRIS
4.2
Table E-ll
2.94
2.98
161.9
4.78E-07
28.17
1,4-Dioxane
123-91-1
0.1
IRIS
6.5
Table E-ll
-0.27
0.421
213900
4.80E-06
15
Benzene
71-43-2
0.015-
0.055°
IRIS
1500
Table E-13
2.13
1.75
2000
5.55E-03
37.5
Di(2-ethylhexyl) phthalated
117-81-7
0.014
IRIS
210
Table E-ll
7.6
4.94
0.001132
2.70E-07
15
Tributyl phosphate
126-73-8
0.009
PPRTV
0.26
Table E-12
4
3.371
7.355
1.41E-06
8.67
N-Nitrosodiphenylamine
86-30-6
0.0049
IRIS
8.9
Table E-ll
3.13
3.42
94.85
1.21E-06
37.5
CASRN = Chemical Abstract Service Registry Number; IRIS = Integrated Risk Information System; PPRTV = Provisional Peer-Reviewed Toxicity Values; K0w = octanol-water
partitioning coefficient; K0c = soil adsorption coefficient
a Oral slope factor (OSF): An upper-bound, approximating a 95% confidence limit, on the increased cancer risk from a lifetime oral exposure to an agent. This estimate, usually
expressed in units of proportion (of a population) affected per mg/kg-day, is generally reserved for use in the low dose region of the dose response relationship, that is, for
exposures corresponding to risks less than 1 in 100. OSFs considered in the MCDA include values from IRIS, PPRTV, and HHBP.
b From Appendix E.
c IRIS lists the OSF for benzene as a range from 0.015 to 0.055 per mg/kg-day. For input into the MCDA, we used the high end of this range (0.055 per mg/kg-day).
d Di(2-ethylhexyl) phthalate is listed under the name bis(2-ethylhexyl) phthalate in Appendix Table E-ll.
9-76
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Cancer MCDA: Chemicals in Produced Water
Toxicity Score (Cancer) ¦ Occurrence Score Physicochemical Properties Score
Benzene
Benzo(a)pyrene
1,2-Diphenylhydrazine
1,4-Dioxane
Di(2-ethylhexyl) phthalate
N-Nitrosodiphenylamine
Tributyl phosphate
0 0.5 1 1.5 2 2.5 3
Total Hazard Potential Score
Figure 9-14. Cancer MCDA results for 7 chemicals detected in produced water, showing the
Toxicity Score, Occurrence Score, and Physicochemical Properties Score for each chemical.
Chemicals are ordered from high to low based on Total Hazard Potential Score. See Section 9.6.4 for details on the
calculation.
9.6.6 Limitations and Uncertainty of the MCDA Framework
While this MCDA framework provides a simple and transparent tool for exploring the relative
hazard potential of chemicals in the hydraulic fracturing water cycle, it is intended only as a
preliminary analysis. It is important to acknowledge the limitations of this analysis, as well as the
limitations of the parameters that were used for input in the MCDA.
Chronic oral RfVs and OSFs were selected for the MCDA because they are a primary focus of the
toxicological evaluation presented in this chapter. We were interested in placing these values in the
context of variables that may impact the likelihood of human exposure. These toxicity values were
available for a relatively small fraction of chemicals on EPA's list, which limited the number of
chemicals considered in the MCDA.
The FracFocus 1.0 data used in the MCDA does not represent a complete record of hydraulic
fracturing chemical usage in the United States, as described in more detail in Chapter 5 and in
Section 9.3.1. Frequency of use also does not reflect the volume or concentration of chemical usage,
and therefore is an incomplete metric for potential exposure. The EPA FracFocus 1.0 project
database provides data on the maximum concentration of chemicals in additives and in hydraulic
fracturing fluid, as discussed in Section 5.4, but we elected not to use this data in the MCDA because
reported concentrations for each chemical varied widely between disclosures (see Table 5-5 and
volume estimates in Figure 5-5), making it difficult to determine a chemical concentration to use in
an MCDA. Additionally, many chemicals in the EPA FracFocus 1.0 project database did not have
valid concentration data; for instance, the maximum concentrations of a chemical in additive often
added up to greater than 100%. We therefore elected to focus on frequency of use as a general
metric of chemical occurrence in the hydraulic fracturing water cycle.
9-77
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
The produced water concentrations used in the MCDA are based on the compilation of data
presented in Appendix E. While this data reflects the findings of recent studies, it does not
represent a complete record of chemicals present in produced water, as described in more detail in
Chapter 7 and in Section 9.3.2. Concentrations in produced water also do not necessarily reflect the
concentrations in treated wastewater, drinking water wells, or residuals in soil or sediment
Concentrations of these chemicals in treated wastewater or well water would likely be more dilute
compared to concentrations in produced water. Concentrations in soils or sediments may be higher,
particularly for hydrophobic chemicals.
The physicochemical properties from EPI Suite used in the MCDA are useful for making comparison
across chemicals, but these values are also subject to uncertainty. Many of the values used in the
MCDA were estimated by EPI Suite, and therefore are subject to the inherent limitations of the EPI
Suite model (Section 5.8). Chemical fate and transport will be also influenced by environmental and
site-specific conditions, which are outside the scope of this analysis. For instance, the half-lives used
to develop the Physicochemical Properties Score are estimated values that assume aerobic
conditions, and thus may underestimate the expected half-life under anaerobic conditions (e.g., in a
groundwater contaminant plume). If chemicals are present in a mixture, as inevitably occurs in
hydraulic fracturing fluids and in the subsurface environment, fate and transport will be influenced
by changes in solubility or degradation resulting from interactions with other chemicals.
There are also fundamental limitations with regards to the scope of the MCDA. The chemicals used
in these analyses may not be representative of chemicals at a specific field site. The analysis only
examined organic chemicals, as EPI Suite is not able to estimate physicochemical properties of
inorganic chemicals. Additionally, the physicochemical properties used in the MCDA were chosen
specifically to reflect chemical transport in water, and therefore do not highlight the potential
hazards of hydrophobic or volatile chemicals. Hydrophobic chemicals may serve as long-term
sources of pollution by sorbing to soils or sediments at contaminated sites, and volatile chemicals
may be hazardous when inhaled. This analysis also does not attempt to address bioavailability or
toxicokinetics, which may be influenced by physicochemical properties such as log Kow- For
instance, chemicals with log Kow of 2-4 tend to absorb well through biological membranes, while
chemicals with log Kow > 4 tend not to absorb well, and those with log Kow of 5-7 tend to
bioconcentrate (U.S. EPA. 2012i).
9.6.7 Application of the MCDA Framework for Preliminary Hazard Evaluation
The MCDA framework presented here is intended as a preliminary analysis, and illustrates one
possible method for integrating data to explore potential hazards. By combining multiple lines of
data, we can stratify chemicals according to estimated hazard potential, and gain preliminary
insight into those chemicals that may be of more concern than others to drinking water resources.
Researchers may find this approach useful in their efforts to explore the potential hazards of
chemicals present at specific field sites, particularly in instances when exposure assessment data is
not available. The MCDA framework is flexible, and could be adapted to incorporate site-specific
data on chemical usage, different types of toxicity data, as well as other variables that may be of
interest for risk assessment For instance, rather than focusing on RfVs and OSFs from US federal
9-78
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
sources, one could choose to derive the Toxicity Score using other sources of relevant toxicity
information. Additionally, one could choose to perform this analysis using different
physicochemical property inputs, to highlight chemical interactions with different environmental
media (e.g., hydrophobic or volatile chemicals). Researchers could also choose to apply different
weights to each of the three criteria considered in this analysis (toxicity, occurrence,
physicochemical properties), to reflect expert judgement of each variable's relative importance.
9.7 Synthesis
The overall objective of this chapter was to identify and provide information on the toxicological
properties of chemicals used in hydraulic fracturing and of hydraulic fracturing wastewater
constituents, and to evaluate the potential hazards of these chemicals for drinking water resources.
Toward this end, the EPA developed a list of 1,606 chemicals that are reported to be associated
with hydraulic fracturing, separating them into subsets based on whether they were reported to
have been used in hydraulic fracturing fluids (1,084 chemicals total) or detected in produced water
(599 chemicals total). To evaluate the potential hazards of these chemicals, the EPA compiled
chronic oral RfVs, OSFs, and qualitative cancer classifications from selected federal, state, and
international sources that met the EPA's criteria for consideration in this assessment This
toxicological information was used to conduct an initial identification of the potential human health
hazards associated with several subsets of chemicals identified as being of particular interest in
previous chapters of this report Finally, in order to illustrate how data integration could be used to
explore potential hazards, an MCDA framework was used to evaluate selected subsets of chemicals
based on toxicity, environmental occurrence, and physicochemical properties affecting chemical
transport in water.
9.7.1 Summary of Findings
A major finding of this chapter was that chronic oral RfVs and OSFs were not available for the
majority of chemicals that the EPA has identified as being associated with hydraulic fracturing
activity, indicating that the majority of these chemicals have not undergone significant toxicological
evaluation. Similarly, there have been several recent peer-reviewed studies that have attempted to
gather toxicological information for subsets of chemicals that are used in hydraulic fracturing fluids,
and they have found that many of these chemicals do not have toxicity values available fElliottetal..
2016: Wattenberg etal.. 2015: Stringfellowetal.. 2014: Colborn etal.. 2011). Taken together, this
suggests a potentially significant knowledge gap exists with respect to the scientific community's
understanding of the potential human health impacts of these chemicals. With the limited
availability of toxicity values, risk assessment is difficult, and potential impacts on drinking water
resources may not be assessed adequately. This lack of toxicity values is not unique to the hydraulic
fracturing industry; in fact, it has been estimated that there are tens of thousands of chemicals in
commercial use that have not undergone significant toxicological evaluation Hudson et al.. 20091.
There are a variety of chemicals associated with hydraulic fracturing known to be hazardous to
human health. Chronic oral RfVs or OSFs from the sources considered by the EPA in this assessment
were available for 98 (9%) of the 1,084 chemicals used in hydraulic fracturing fluids, and 120
(20%) of the 599 chemicals detected in hydraulic fracturing produced water. Potential hazards
9-79
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
associated with chronic oral exposure to these chemicals include carcinogenesis, immune system
effects, changes in body weight, changes in blood chemistry, cardiotoxicity, neurotoxicity, liver and
kidney toxicity, and reproductive and developmental toxicity. Methane is not considered to be toxic
when ingested, but may accumulate to explosive levels or act as an asphyxiant DBPs formed during
wastewater treatment can contribute to an increased risk of cancer, anemia, liver and kidney
effects, and central nervous system effects, with brominated forms of DBPs considered to be more
cytotoxic, genotoxic, and carcinogenic than chlorinated species.
To assess the toxicity of chemicals that lack chronic oral RfVs and OSF, risk assessors will need to
turn towards alternative data sources. This chapter explored two alternative data sources that may
provide useful information. QSAR-based toxicity estimates—specifically, rat chronic oral LOAEL
estimates generated using TOPKAT—were available for many of the chemicals that lacked chronic
oral RfVs and OSFs from the sources considered in this assessment, and may be used to rank
chemicals based on toxicity when other data are not available. Additionally, many of these
chemicals have information available on the EPA's ACToR database, which is an online data
warehouse designed to consolidate large and disparate amounts of chemical data. The information
available in the ACToR data warehouse ranges from the selected RfVs and OSFs discussed in this
assessment, which have undergone extensive peer review, to toxicological data that have
undergone little-to-no peer review.
When considering the potential impact of chemicals on drinking water resources and human health,
it is important to consider exposure as well as toxicological properties. As discussed in previous
chapters of this report and highlighted in this chapter, events such as spills, leaks from storage pits,
and discharge of inadequately treated wastewater have led to the entry of hydraulic fracturing-
related chemicals into drinking water resources. In some instances, chemical concentrations in
surface water or groundwater were in exceedance of MCLs, indicating their presence at levels that
could impact human health. While these studies demonstrate the potential entry of these chemicals
into drinking water resources, there is a lack of systematic studies examining actual human
exposures to these chemicals in drinking water as a result of hydraulic fracturing activity.
In the absence of exposure assessment data, the MCDA framework presented in this chapter
provides a preliminary analysis of the relative hazard potential of these chemicals. In this context,
occurrence and physicochemical property data were used as metrics to estimate the likelihood that
a chemical could reach and impact drinking water, and toxicity data was used as a metric for the
potential severity of an impact. This analysis highlighted several chemicals that may be more likely
than others to reach drinking water and create a toxicological hazard. Of the chemicals used in
hydraulic fracturing fluids that were considered in this analysis, chemicals such as propargyl
alcohol stood out as having high potential toxicity, high frequency of use, and physicochemical
properties that are conducive to transport in water. Of the chemicals in produced water, chemicals
such as benzene, pyridine, 2-methylnaphthalene, and naphthalene stood out as having high
potential toxicity, high concentrations in produced water, and physicochemical properties that are
conducive to transport in water.
9-80
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
9.7.2 Factors Affecting the Frequency or Severity of Impacts
There are multiple pieces of information that could be taken into account when evaluating the
frequency and severity of impacts that these chemicals may have on drinking water resources. This
includes knowledge of the chemicals used at a given site, the toxicological and physicochemical
properties of these chemicals, the amount of fluid being used and recovered, the likelihood of
mechanical integrity failures, the likelihood of spills and other unintentional releases, and the
efficiency of chemical removal during wastewater treatment The MCDA presented in this chapter
incorporated parameters that may impact the likelihood of chemical exposure, including frequency
of use, measured concentration, and transport in water, and was used to stratify and rank chemicals
based on relative hazard potential. However, it should be considered only as a preliminary analysis,
and should not be used in place of local data on the concentrations and volumes of chemicals in
areas of hydraulic fracturing activity.
Analysis of the chemicals used in hydraulic fracturing fluids indicated that the majority of chemicals
on the EPA's list are used in <1% of wells nationally (Figure 9-4). Therefore, potential exposure to
the majority of these chemicals is more likely to be a local issue, rather than a national one. Given
that the analysis of the EPA FracFocus 1.0 project database presented in this chapter was based on
35,957 disclosures, a chemical used in <1% of wells nationally could still be used in several
hundred wells. Chemicals used infrequently on a national basis could still be used more frequently
within certain areas or counties, increasing the potential for local exposure to that chemical.
As an example of how an infrequently used chemical could have local impacts, consider (E)-
crotonaldehyde, which had one of the lowest chronic oral RfVs among the chemicals considered in
the noncancer MCDA for hydraulic fracturing chemicals, and was reported in approximately 0.06%
of disclosures in the EPA FracFocus 1.0 project database. If the EPA FracFocus 1.0 project database
is a representative sample of all of the wells across the country, then the likelihood of (E)-
crotonaldehyde contamination on a national scale is limited. However, this in no way diminishes
the likelihood or potential severity of (E)-crotonaldehyde contamination at sites where this
chemical is used.
This is in contrast with frequently used chemicals such as methanol. Methanol was reported in 73%
of wells in the EPA FracFocus 1.0 project database, and was the most frequently used chemical
considered in the noncancer MCDA for chemicals used in hydraulic fracturing fluids. Methanol is
soluble and relatively mobile in water, but has a higher chronic oral RfV compared other chemicals
considered in this analysis. Therefore, methanol may be expected to have a higher exposure
potential on a national basis compared to other chemicals, with a moderate hazard potential due to
its relatively high RfV.
Even if no chemicals were added to hydraulic fracturing fluids, there is still a potential for impacts
from constituents naturally present in the subsurface which could be brought to the surface in
produced water. As described in Section 9.5, many of the naturally occurring chemicals in produced
water—e.g., organic chemicals (e.g., BTEX and related hydrocarbons), metals, anions, and
TENORM—are hazardous to human health and have been reported in drinking water resources as a
result of hydraulic fracturing activity, sometimes at concentrations exceeding MCLs. The
9-81
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
constituents of produced water that contribute to the formation of DBPs, specifically bromide,
chloride, iodine, and ammonium, are naturally occurring and are characteristic of wastewater from
hydraulically fractured wells.
Overall, contamination of drinking water resources depends on site-, chemical-, and fluid-specific
factors fGoldstein etal.. 20141. and the exact mixture and concentrations of chemicals at a site will
depend upon the geology and the chemicals used in the oil and gas extraction processes. Therefore,
potential hazard and risk considerations are best made on a site-specific, well-specific basis.
9.7.3 Uncertainties
There are notable uncertainties in the chemical and toxicological data limiting a comprehensive
assessment of the potential health impacts of hydraulic fracturing on drinking water resources.
For human health risk assessment, a significant data gap is the lack of chronic oral RfVs and OSFs
from sources meeting the EPA's criteria for inclusion in this report. For instance, of the 34
chemicals (excluding water, quartz, and sodium chloride) that were reported in >10% of
disclosures in the EPA FracFocus 1.0 project database, 9 chemicals have chronic oral RfVs available,
and none have OSFs (Table 9-2). Without reliable and peer reviewed toxicity values, comprehensive
hazard evaluation and hazard identification of chemicals is difficult, and the ability to consider the
potential cumulative effects of exposure to chemical mixtures in hydraulic fracturing fluid or
produced water may be limited. Although there are other potential sources of toxicity information
for many of these chemicals, some of it may be limited or of lesser quality. Consequently, potential
impacts on drinking water resources and human health may not be assessed adequately.
An equally significant data gap is the lack of exposure assessment data for drinking water resources
in areas of hydraulic fracturing activity. As discussed in Text Box 9-1, data on exposure potential is
a critical component of the risk assessment process, and is necessary for risk characterization. In
the absence of exposure assessment information, the MCDA framework presented in this chapter
may be useful for exploring the potential hazards of hydraulic fracturing-related chemicals, but
should be considered as a preliminary analysis only. The MCDA presented in this chapter
considered only a small subset of chemicals that had data available, was limited in scope, and may
not be representative of the chemicals that are present at a specific field site. It should be
emphasized that this MCDA framework represents just one method that can be used to integrate
chemical data for hazard evaluation, and is readily adaptable to include different variables, different
weights for the variables, and site-specific considerations.
There is also uncertainty surrounding the EPA's list of chemicals associated with hydraulic
fracturing activity. As discussed in Section 5.4 and Section 9.3.1, there is incomplete information
available on chemicals used in hydraulic fracturing fluids due to industry use of CBI as well as
incomplete reporting of chemical use. For instance, the EPA's analysis of the FracFocus 1.0 project
database found that approximately 11% of ingredients were reported as CBI, and that more than
70% of FracFocus 1.0 disclosures contained at least one CBI ingredient. There may also be regional
limitations in the disclosures submitted to FracFocus 1.0, as 78% of chemical disclosures in came
from five states, and 47% were from Texas fU.S. EPA. 2015al. Despite these limitations, FracFocus
9-82
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
remains the most complete source for tracking hydraulic fracturing chemical usage in the United
States, and therefore was the best available source for the hazard evaluation in this chapter.
Although the sources used to compile the chemical list represented the best available data at the
time of this study, it is possible that some of these chemicals are no longer used at all, and many of
these chemicals may only be used infrequently. Therefore, it may be possible that significantly
fewer than 1,084 chemicals are currently used in abundance. As practices evolve, it is likely that
chemicals are used or will be used that are not included on this chemical list. Having a better
understanding of the chemicals and formulations, including those that are CBI, along with their
frequency of use and volumes, would greatly benefit risk assessment and risk management
decisions.
Additionally, the list of produced water chemicals identified in this chapter is almost certainly
incomplete. As discussed in Chapter 7, chemicals and their metabolites may go undetected because
they were not included in the analytical methodology, or because an analytical methodology was
not available. Chemical analysis of produced water can also be challenging because high levels of
dissolved solids in produced water and wastewater can interfere with chemical detection. As a
result, there are likely chemicals of concern in produced water that have not been detected or
reported, and are not included on the chemical list presented in this report
9.7.4 Conclusions
The EPA identified 1,606 chemicals associated with the hydraulic fracturing water cycle, including
1,084 chemicals used in hydraulic fracturing fluids, and 599 chemicals detected in produced water.
Toxicity-based chronic oral RfVs and/or OSFs from sources meeting selection criteria were not
available for the majority (89%) of the chemicals on this total list Thirty-seven percent of
chemicals on the EPA's list that are used in hydraulic fracturing fluids lack data on their frequency
of use. Current understanding of the chemical composition of produced water is constrained by
analytical chemistry limitations and by the likelihood that chemical composition will vary between
wells. A limited number of studies have detected these chemicals in surface water, groundwater, or
well water near areas of hydraulic fracturing activity, suggesting the potential for human exposure;
however, actual human exposures to these chemicals in drinking water resources has not been well
characterized. Given the large number of chemicals used or detected in various stages of the
hydraulic fracturing water cycle, as well as the large number of hydraulically fractured wells
nationwide, this missing chemical information represents a significant data gap.
While it remains challenging to fully understand the toxicity and potential public health impacts of
these chemicals for drinking water resources, the toxicological data, occurrence data, and
physicochemical data compiled in this report provide a resource for assessing the potential hazards
of chemicals in the hydraulic fracturing water cycle. The MCDA framework presented here
illustrates one method for integrating these data for a preliminary hazard evaluation, which may be
useful when exposure assessment data are not available. While the analysis in this chapter is
constrained to the assessment of chemicals on a national scale, this approach is readily adaptable
for use on a regional or site-specific basis.
9-83
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
This collection of data provides a tool to inform decisions about protection of drinking water
resources. Stakeholders may use these results to prioritize chemicals for hazard assessment or for
determining future research priorities. Industry may use this information to prioritize chemicals for
replacement with less toxic, persistent, and mobile alternatives.
9.8 Annex
9.8.1 Calculation of Physicochemical Property Scores (MCDA Hazard Evaluation)
Section 9.6.3 describes how Physicochemical Properties Scores for the noncancer and cancer
MCDAs were calculated based on three subcriteria which affect the likelihood that a chemical will
be transported in water: mobility, volatility, and persistence. Calculation of these subcriteria scores
was performed as described by Yost et al. fin Press! as follows:
9.8.1.1 Mobility Score
Chemical mobility in water was assessed based upon three physicochemical properties that
describe chemical solvency in water: the octanol-water partition coefficient (Kow), the soil
adsorption coefficient (Koc), and aqueous solubility. Kow describes the partitioning of a chemical
between water and a carbon-based media (octanol), while Koc described the partitioning of a
chemical between water and organic carbon in soil. Kow and Koc are generally represented on a
logarithmic scale. Aqueous solubility is the maximum amount of a chemical that will dissolve in
water in the presence of pure chemical. Chemicals with low Kow, low Koc, or high aqueous solubility
are more likely to solubilize and move with water, and therefore were ranked as having greater
potential to affect drinking water resources.
For input into the MCDA, we used experimentally measured values (provided in EPI Suite)
whenever available. Otherwise, we used the following estimated values from EPI Suite: log Kow
estimated using the KOWWIN™ model, log Koc estimated using the KOCWIN™ Sabljic molecular
connectivity method, and aqueous solubility estimated using the WSKOWWIN™ model. Using the
thresholds designated in Table 9-11, each of these properties was assigned a score of 1-4. The
highest of these three scores (Kow, Koc, or solubility) was designated as the Mobility Score for each
chemical.
9.8.1.2 Volatility Score
Chemical volatility was assessed based on the Henry's law constant, which is the ratio of the
concentration of a chemical in air to the concentration of that chemical in water. Chemicals with
low Henry's law constants are less likely to leave water via volatilization, and were therefore
ranked as having greater potential to affect drinking water resources.
For input into the MCDA, we used experimentally measured values (provided in EPI Suite)
whenever available. Otherwise, we used Henry's Law constants that were estimated using the EPI
Suite HENRYWIN™ model, which generates values using two different methods (group contribution
and bond contribution); the lower of these two estimated values was used as input into the MCDA.
9-84
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
Using the thresholds designated in Table 9-11, the Henry's law constant for each chemical was
assigned a score of 1-4. This value was designated as the Volatility Score for each chemical.
9.8.1.3 Persistence Score
Chemical persistence was assessed based on estimated half-life in water, which describes how long
a chemical will persist in water before it is transformed or degraded. Chemicals with longer half-
lives are more persistent, and were therefore ranked as having greater potential to impact drinking
water resources.
EPI Suite estimates biodegradation time using the BIOWIN™ 3 model, which provides an indication
of a chemical's environmental biodegradation rate in relative terms (e.g., hours, days, weeks, etc.),
assuming aerobic conditions. These BI0WIN3 estimates are converted to numerical half-life values
for use in EPI Suite's Level III Fugacity model. For input into the MCDA, we used the same estimated
half-life in water that is used in the Level III Fugacity model. Using the thresholds designated in
Table 9-11, the half-life in water of each chemical was assigned a score of 1-4. This value was
designated as the Persistence Score for each chemical.
9.8.1.4 Total Physicochemical Properties Score
For each chemical, the Mobility Score, Volatility Score, and Persistence Score (each on a scale of 1 to
4) were summed to calculate a total Physicochemical Properties Score. Higher Physicochemical
Properties Scores indicate chemicals that are more likely to be transported in water, with a
maximum possible score of 12.
9.8.2 Example of MCDA Score Calculation
The methods used for MCDA score calculation are described in Section 9.6.3. For an example of how
the MCDA scores were calculated, consider benzene, which was included in both the noncancer
MCDA (national analysis) and cancer MCDA for chemicals used in hydraulic fracturing fluids. This
demonstrates how MCDA scores were calculated for benzene for these two different analyses.
9.8.2.1 Score Calculation for Benzene in Noncancer MCDA for Hydraulic Fracturing Fluids
• Toxicity Score (Noncancer): Benzene has a chronic oral RfV of 0.004 mg/kg-day (source:
IRIS). Across the 42 chemicals that were considered in the noncancer MCDA (national
analysis), chronic oral RfVs ranged from 0.001 mg/kg-day [(E)-crotonaldehyde] to 20
mg/kg-day (1,2-propylene glycol). The chronic oral RfV of benzene falls in the lowest
(most toxic) quartile of these chemicals, and therefore benzene was assigned a Toxicity
Score of 4. When the results were standardized to the highest Toxicity Score (4) and
lowest Toxicity Score (1) within the set of chemicals, benzene was calculated to have a
final Toxicity Score of 1, as follows:
1 = (4- 1)/ (4- 1)
• Occurrence Score: Benzene was used in 0.006% of wells nationally. For the 42 chemicals
considered in the national noncancer MCDA, frequency of use ranged from 73%
9-85
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
(methanol) to 0.003% (furfural) of wells nationally. Benzene falls in the lowest quartile
with regards to frequency of use, and therefore benzene was assigned an Occurrence Score
of 1. When the results were standardized to the highest Occurrence Score (4) and lowest
Occurrence Score (1) within the set of chemicals, benzene was calculated to have a final
Occurrence Score of 0, as follows:
0 = (1 - 1)/ (4- 1)
• Physiochemical Properties Score: Benzene received a Mobility Score of 4 (log K0w =
2.13; logKoc = 1.75; solubility = 2000 mg/1), a Volatility Score of 2 (Henry's law constant =
0.00555), and a Persistence Score of 2 (half-life in water = 37.5 days). This sums to a Total
Physicochemical Properties Score of 8. Within the 42 chemicals considered in the national
noncancer MCDA, several chemicals received Total Physicochemical Properties Scores of
9, which was the highest observed score. Cumene received a Total Physicochemical
Properties Scores of 6, which was the lowest score. When the results were standardized to
the highest (9) and lowest (6) of these scores, benzene was calculated to have a final Total
Physicochemical Properties Scores of 0.67, as follows:
0.67 = (8-6)/(9-6)
• Total Hazard Potential Score (Noncancer MCDA): For benzene, the Toxicity Score (1),
Occurrence Score (0), and Physicochemical Properties Score (0.67) were summed to
calculate a Total Hazard Potential Score of 1.67. The relative contribution of the three
criteria scores to this total score is depicted as a graphic in Figure 9-8.
9.8.2.2 Score Calculation for Benzene in Cancer MCDA for Hydraulic Fracturing Fluids
• Toxicity Score (Cancer): Benzene has an OSF of 0.055 per mg/kg-day (source: IRIS).
Within the entire set of 10 chemicals that was considered in the cancer MCDA, OSFs
ranged from 3 (quinoline) to 0.002 (dichloromethane) per mg/kg-day. The OSF of benzene
falls in the second quartile of these scores, and therefore was assigned a Toxicity Score of
2. When the results were standardized to the highest Toxicity Score (4) and lowest
Toxicity Score (1) within the set of chemicals, benzene was calculated to have a final
Toxicity Score of 0.33, as follows:
0.33 = (2 - 1)/ (4- 1)
• Occurrence Score: As described in the noncancer MCDA above, benzene was used in
0.006% of wells nationally. This was the lowest frequency of use among the 10 chemicals
that were considered in the cancer MCDA, with benzyl chloride (used in 6% of wells)
having the highest Benzene therefore falls in the lowest quartile with regards to
frequency of use, and was assigned an Occurrence Score of 1. When the results were
standardized to the highest Occurrence Score (4) and lowest Occurrence Score (1) within
the set of chemicals, benzene was calculated to have a final Occurrence Score of 0, as
follows:
9-86
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
0 = (1 - 1)/ (4- 1)
• Physiochemical Properties Score: As described in the noncancer MCDA above, benzene
received a Total Physicochemical Properties Score of 8. Within the 10 chemicals that were
considered in the cancer MCDA, all chemicals either received a Total Physicochemical
Properties Score of 8 or 9. When the results were standardized to these high and low
scores, benzene was calculated to have a final Total Physicochemical Properties Scores of
0 as follows:
0 = (8 - 8) / (9 - 8)
• Total Hazard Potential Score (Cancer MCDA): The Toxicity Score (0.33), Occurrence
Score (0), and Physicochemical Properties Score (0) were summed to calculate a Total
Hazard Potential Score of 0.33. The relative contribution of the three criteria scores to this
total score is depicted as a graphic in Figure 9-12.
9-87
-------
Chapter 9 - Identification and Hazard Evaluation of Chemicals across the Hydraulic Fracturing Water Cycle
This page is intentionally left blank.
9-88
-------
Chapter 10 - Synthesis
Chapter 10. Synthesis
10-1
-------
Chapter 10 - Synthesis
This page is intentionally left blank.
10-2
-------
Chapter 10 - Synthesis
10. Synthesis
Introduction
The goals of this report were to assess the potential for activities in the hydraulic fracturing water
cycle to impact the quality or quantity of drinking water resources, and to identify factors affecting
the frequency or severity of those impacts. Overall, we conclude activities in the hydraulic
fracturing water cycle can impact drinking water resources under some circumstances. Impacts can
range in frequency and severity, depending on the combination of hydraulic fracturing water cycle
activities and local- or regional-scale factors. The following combinations of activities and factors
are more likely than others to result in more frequent or more severe impacts:
• Water withdrawals for hydraulic fracturing in times or areas of low water availability,
particularly in areas with limited or declining groundwater resources;
• Spills during the management of hydraulic fracturing fluids and chemicals or produced
water that result in large volumes or high concentrations of chemicals reaching
groundwater resources;
• Injection of hydraulic fracturing fluids into wells with inadequate mechanical integrity,
allowing gases or liquids to move to groundwater resources;
• Injection of hydraulic fracturing fluids directly into groundwater resources;
• Discharge of inadequately treated hydraulic fracturing wastewater to surface water
resources; and
• Disposal or storage of hydraulic fracturing wastewater in unlined pits, resulting in
contamination of groundwater resources.
These conclusions are based on cases of identified impacts and other data, information, and
analyses presented in this report. Cases of impacts were identified for all stages of the hydraulic
fracturing water cycle. Identified impacts generally occurred near hydraulically fractured oil and
gas production wells and ranged in severity, from temporary changes in water quality to
contamination making private drinking water wells unusable. The inherent characteristics of
groundwater resources make them more vulnerable to impacts from activities in the hydraulic
fracturing water cycle compared to surface water.
We see the identification of factors affecting the frequency or severity of impacts, and uncertainties
and data gaps in this report as particularly useful for decision makers. Factors often can be
managed, changed, or used to identify areas for specific monitoring or modification of practices.
Thus, in the short-term, information on factors can help decision makers reduce current
vulnerabilities of drinking water resources to activities in the hydraulic fracturing water cycle. In
the longer term, reducing the uncertainties and filling the data gaps could enhance science-based
decisions to protect drinking water resources in the future.
The purpose of this chapter is to synthesize for decision makers the information on factors,
uncertainties, and data gaps presented in this assessment In Section 10.2, we focus on factors
10-3
-------
Chapter 10 - Synthesis
increasing or decreasing the frequency or severity of impacts at each stage of the hydraulic
fracturing water cycle. In Section 10. 3, we discuss major uncertainties and data gaps identified in
this assessment Finally, in Section 10.4, we discuss potential uses for this assessment
10.1 Factors Affecting the Frequency or Severity of Impacts
10.1.1 Water Acquisition
Groundwater and surface water resources serve as both sources of water for hydraulic fracturing
and public and private drinking water supplies. Thus, water withdrawals for hydraulic fracturing
can impact the quantity or quality of drinking water resources under certain circumstances. Since,
by definition, every water withdrawal affects water quantity, we focused in this assessment not on
all water withdrawals per se, but rather on those with the potential to limit the availability of
drinking water or alter its quality. Whether a withdrawal has this potential depends upon a
combination of factors at the local scale. Factors can either increase or decrease the frequency or
severity of impacts. In this section on water acquisition, we combine our discussion of frequency
and severity because all of the factors we discuss in this section affect both frequency and severity
in a similar fashion (i.e., either increase both frequency and severity, or decrease both frequency
and severity).
10.1.1.1 Frequency and Severity
The local balance between water withdrawals and water availability is the most important factor
determining whether water acquisition impacts are likely to occur or be severe. Impacts are more
likely to be frequent or severe where or when hydraulic fracturing water withdrawals are relatively
high and water availability is low. In contrast, the same amount of water withdrawn can have a
negligible effect if withdrawn in an area of—or at a time of—higher water availability. For this
reason, it is important not to focus solely on the amount withdrawn, but the balance between water
withdrawals and availability in place and time.
For this assessment, we developed county-level estimates of water use (i.e., water withdrawals) for
hydraulic fracturing, which were then compared to an index of readily available fresh water. This
readily available fresh water index included unappropriated surface water and groundwater, and
appropriated water potentially available for purchase (Tidwell etal.. 2013) (Text Box 4-2).1 In the
majority of counties where hydraulic fracturing takes place, hydraulic fracturing water use was less
than 1% of this index of readily available fresh water. We did find, however, a small number of
counties with higher percentages. There were 45 counties out of the almost 400 surveyed where
hydraulic fracturing water use was above 10% of the index. Of these counties, 35 exceeded 30%,
and 17 of these counties had hydraulic fracturing water use exceeding the index. All of the counties
in this latter category are located in Texas.
1 In the western United States, water is generally allocated by the principle of prior appropriation—that is, first in time of
use is first in right. New development must use unappropriated water or purchase appropriated water from vested users.
In the index of readily available fresh water, it was assumed 5% of appropriated irrigated water could be purchased. See
Text Box 4-2 for more details about this analysis.
10-4
-------
Chapter 10 - Synthesis
This does not mean impacts to drinking water quantities occurred or will occur in these counties,
nor does it mean that impacts did not or will not occur in counties with relatively low percentages.
To truly determine whether impacts occurred, water withdrawals and availability need to be
compared at the scale of the drinking water resource. For instance, groundwater withdrawals for
hydraulic fracturing could affect water levels in nearby private water wells. As a national
assessment, we could not often examine impacts at this local scale, although we did cite studies of
local impacts where available. Nevertheless, our county level assessment does point to places
where the potential for impacts is higher. This information may be useful to focus efforts on
reducing the fresh water demand of hydraulic fracturing.
Beyond our county level assessment, we conclude that declining groundwater resources are
particularly vulnerable to water quantity and quality impacts from withdrawals. Groundwater
recharge rates can be low, and groundwater withdrawals are exceeding recharge in areas of the
country (Konikow. 2013). When withdrawals exceed recharge, the result is declining water levels.
For this reason, water levels in some aquifers in the United States have declined substantially over
the last century fKonikow. 20131. Although irrigated agriculture is often the dominant user of
groundwater, hydraulic fracturing withdrawals now also contribute to declining groundwater
levels in some areas (e.g., southern Texas; Steadman etal.. 2015: Scanlon et al.. 2014b) Cumulative
groundwater withdrawals can also impact water quality by mobilizing chemicals, such as uranium,
from naturally occurring sources in the surrounding rock into the groundwater (DeSimone etal..
2014).
In certain instances, state and local governments have encouraged or mandated the use of surface
water in place of groundwater, as evidenced in both Louisiana and North Dakota. In 2008, the state
of Louisiana asked oil and gas companies to switch from groundwater to surface water to mitigate
stress on the Carrizo-Wilcox aquifer, a critical source of drinking water in the region. Likewise, the
state of North Dakota requested the oil industry obtain water from the Missouri river system, and
not from stressed groundwater sources. By contrast, surface water availability is limited in other
regions and cannot provide an alternative source of water (e.g., western Texas).
Among surface water sources, small streams are particularly vulnerable to impacts. This is the case
across the country, even in the eastern United States where surface water is generally more
plentiful. An EPA study of the Susquehanna River Basin in northeastern Pennsylvania found that the
smallest streams (with less than 10 mi2 of contributing area-i.e., the watershed area drained by the
stream) would be the most likely to be impacted from water withdrawals in the absence of
protective passby flows; see discussion below and U.S. EPA f2015el.1 While the amount of
contributing area varies by geographic location due to differences in runoff, the finding that the
smallest streams are the most vulnerable to withdrawals holds across all landscapes.
Not only does water availability vary from one location to another, but it can also vary temporally at
a given location, often due to variations in precipitation. Because of this dynamic, long-term or
seasonal drought can increase the frequency or severity of impacts from withdrawals by decreasing
water availability. The EPA study of the Susquehanna River Basin found even larger streams (up to
1 Passby flows are low stream flow thresholds below which withdrawals are not allowed.
10-5
-------
Chapter 10 - Synthesis
600 mi2 of contributing area) would be vulnerable to impacts at times of drought, again absent
passby flows fU.S. EPA. 2015el. Dry conditions can also stress groundwater supplies by
simultaneously increasing water demand (e.g., irrigation water demand increases in dry
conditions) while also decreasing groundwater recharge. Much of the western United States has
experienced extended periods of drought over the last decade. Climate change is likely to
exacerbate these conditions in certain locations (Meixner etal.. 2016).
Conversely, there are factors that can reduce the frequency or severity of impacts. Reuse of
hydraulic fracturing wastewater (i.e., produced water managed for reuse, treatment and discharge,
or disposal), for example, can reduce demands on fresh water resources.1 Reuse does not appear to
be driven by water scarcity, but rather by the cost of disposal. Operators are likely to dispose of
wastewater when it is less expensive than reuse. For instance, greater reuse of wastewater occurs
in the Marcellus Shale in Pennsylvania than in the Barnett Shale in Texas, even though water
availability is generally higher in the Marcellus region (Figure 10-1). The general lack of disposal
wells in Pennsylvania means disposing of wastewater requires trucking to Ohio or other locations
with disposal wells. Because of this expense, operators reuse substantial proportions of their
wastewater, in contrast to the Barnett Shale where disposal wells are readily available.
The reuse of wastewater to offset fresh water use in hydraulic fracturing is often limited by the
amount of wastewater available. The volume of produced water from a single well can be relatively
small compared to the volume needed to fracture a well (Figure 10-la). This means produced water
would need to be aggregated from multiple wells to equal the volume needed to hydraulically
fracture an additional well. For instance, it would take 10 wells to make enough water to fracture
an 11th well if, as has been shown in the Marcellus Shale in Pennsylvania, produced water volumes
are 10% of injected volumes (Figure 10-la). Thus, reuse is a factor that can reduce fresh water
demand, but not eliminate it in most cases. Nevertheless, even a marginal decline in fresh water
demand can make a difference in the frequency or severity of impacts.
The use of brackish groundwater is also a factor reducing fresh water demand, in some cases to a
much greater degree than reuse. In the Permian Basin in western Texas, for instance, brackish
water makes up 30 to 80% of water used for hydraulic fracturing, and 20% in the Eagle Ford Shale
in southern Texas (Nicot etal.. 2012). Our county level estimates suggest brackish water availability
could entirely meet current hydraulic fracturing water demand in Texas and many other locations.2
In 35 counties nationally, hydraulic fracturing water use equaled or exceeded 30% of an index of
fresh water availability; when brackish water and wastewater were considered in addition to fresh
water availability, only two counties equaled or exceeded 30% (Text Box 4-2).
1 Hydraulic fracturing wastewater is produced water that is managed using practices that include, but are not limited to,
reuse in subsequent hydraulic fracturing operations, treatment and discharge, and injection into disposal wells. The term
is being used in this study as a general description of certain waters and is not intended to constitute a term of art for
legal or regulatory purposes (see Chapter 8 and Appendix J, the Glossary, for more detail].
2 Brackish water for the purposes of this analysis ranged from 3,000 to 10,000 ppm of total dissolved solids (TDS], and
from 50 to 2,500 ft (15-760 m] below the surface fTidwell et al.. 20131 (See Text Box 4-2 for more details.]
10-6
-------
Chapter 10 - Synthesis
Marcellus Shale,
Susquehanna River Basin
4.1-4.6 million
gallons injected1"
420,000-1.3 million gallons
produced11
90%d*
79%a
Surface Water ¦ Groundwater
Reused hydraulic fracturing wastewater
Reuse in hydraulic fracturing
¦ Class II well
'Less than approximately 1% is treated at facilities that are
either permitted to discharge to surface water or whose
discharge status is uncertain.
Most of the injected fluid stays in the subsurface; produced
water volumes over 10 years are approximately 10-30% of
the injected fluid volume.c
Barnett Shale, Texas
3.9-4.5 million
gallons injected'
3.9-4.5 million gallons
produced6
Surface Water ¦ Groundwater
Reused hydraulic fracturing wastewater
Reuse in hydraulic fracturing
¦ Class II well
Produced water volumes over three years are
approximately the same as the injected fluid volume.^
Figure 10-1. Water budgets representative of practices in (top) the Marcellus Shale in the
Susquehanna River Basin in Pennsylvania and (bottom) the Barnett Shale in Texas.
Pie size and arrow thickness represent the relative volume of water as it flows through the hydraulic fracturing
water cycle. Water budgets illustrative of typical water management practices in the Marcellus Shale in the
Susquehanna River Basin between approximately 2008 and 2013 and the Barnett Shale in Texas between
approximately 2011 and 2013, They do not represent any specific well. Sources for the top figure (a) Tables 4-1 and
4-2 (SRBC, 2016)—note, surface water, groundwater, and reuse values of 92%, 8%, and 16% in table normalized to
79%, 7%, 14%, respectively, for this chart (this was done to represent reuse on the same chart as surface water
and groundwater—in the original tabular values, reuse is expressed as a percentage of total water used, and
surface water and groundwater are expressed in percentages relative to each other); (b) Appendix Table B-5 (U.S.
EPA, 2015a); (c) Table 7-2 (Ziemkiewicz et al,, 2014)—note: produced water volumes estimated from percentages
applied to volumes injected, and value from the West Virginia portion of the Marcellus Shale used in this chart
since it was the longest term measurement of produced water volumes; (d) Figure 8-4 (PA PEP, 2015a) and Table
8-6 (Ma et al., 2014; Shaffer et al,, 2013). Sources for the bottom figure: (e) Tables 4-1 and 4-2 (Nicot et al., 2014;
IMicot et al., 201.2)—note, surface water, groundwater, and reuse values of 50%, 50%, and 5% in the tables
normalized to 48%, 48%, and 4%, respectively, for this chart (see reason for this above); (f) Appendix Table B-5
(U.S. EPA, 2015a: Nicot et al., 2012: Nicot et al., 2011)—note: see median value for Fort Worth Basin; (g) Table 7-2
(Nicot et al., 2014): (h) Table 8-6 (Nicot et al., 2012)—note, percentage going to disposal wells estimated by
subtracting reuse values from 100%,
10-7
-------
Chapter 10 - Synthesis
Finally, passby flows can be a factor reducing the frequency or severity of surface water impacts.
Passby flows are low stream flow thresholds below which withdrawals are not allowed. This
management practice has been shown to be protective of streams from over-withdrawals in the
Susquehanna River Basin in northern Pennsylvania fU.S. EPA. 2015el This is likely most important
for protecting aquatic life in smaller streams, but may also aid in protecting drinking water
supplies.
10.1.2 Chemical Mixing and Produced Water Handling
Like water acquisition, activities in the chemical mixing and produced water handling stages of the
hydraulic fracturing water cycle can impact drinking water in some instances. We combine our
discussion of the two stages here because activities in these stages both affect drinking water
resources primarily through spills. The chemical mixing stage encompasses management of fluids
on the well pad to create hydraulic fracturing fluid. Chemicals are mixed with a base fluid, typically
water, and then injected into the production well. After the pressure is released post-fracturing,
produced water flows from the well and needs to be collected and managed in the produced water
handling stage.
Chemical mixing and produced water handling activities can impact drinking water resources
through spills of chemicals used to make hydraulic fracturing fluid, hydraulic fracturing fluid itself,
or produced water reaching surface water or groundwater.1 There is some information on spill
frequencies—although limited—and spill severities are most often uncharacterized. Nevertheless,
we could identify factors affecting the frequency or severity of impacts from chemical mixing or
produced water spills. In the section below, we discuss these factors, with those affecting frequency
first, followed by those affecting severity. We discuss each of the factors individually, but spill
events in reality exhibit combinations of these factors. These factors can interact to increase or
decrease the frequency or severity of a spill beyond the effect of an individual factor.
10.1.2.1 Frequency
An impact on the quality of a drinking water resource from a spill first depends on a spill occurring.
Most spill frequency estimates are of spills in total, and not the subset reaching drinking water
resources. Spill estimates from three states (Colorado, North Dakota, and Pennsylvania) ranged
from 0.4 to 12.2 reported spills per 100 hydraulically fractured wells (Appendix C.4).2 The
estimates from Pennsylvania and Colorado included hydraulic fracturing chemicals, fluids, and
produced water; while the North Dakota estimate was based on spills of hydraulic fracturing
chemicals and fluids only.3 Spill rates can also be expressed on a per-active-well basis. This may be
1 In Chapter 5 and elsewhere in this assessment, the chemicals added to the base fluid (most often water] and proppant
(most often sand] are referred to as "additives" since this is the term used in FracFocus. Here, this chapter simply refers to
them as "chemicals." It does this to discuss chemicals in a unified manner in this combined section on chemical mixing and
produced water.
2 Since most wells are not reported hydraulically fractured in databases, these estimates used spudded, completed, or
installed wells as proxies for hydraulically fractured wells. (See Appendix Section C.4 for more detail.]
3 These estimates from Pennsylvania and Colorado also included spills of diesel fuel and drilling muds, which could not be
separated out from the total frequency estimate even though they were generally out-of-scope of this assessment (diesel
fuel was in scope if used in hydraulic fracturing fluid].
10-8
-------
Chapter 10 - Synthesis
more appropriate for produced water spills since they can occur years or even decades after
hydraulic fracturing. An analysis of North Dakota produced water spills found there were
approximately 5 to 7 spills of produced water per 100 active wells between 2010 and 2015
(Appendix E.5). We conclude from these data that spills do occur in both the chemical mixing and
produced water stages of the hydraulic fracturing water cycle, generally in the range of 1 to 10% of
hydraulically fractured or active wells.
Not all spills, however, reach and therefore impact a drinking water resource. In U.S. EPA f2015ml.
32 of the 457 (7%) spills characterized were reported to have reached surface water or
groundwater. The California Office of Emergency Services estimated 18% of produced water spills
reached waterways between January 2009 and December 2014 fCCST. 2015bl. It is unclear if this
estimate included groundwater, or was limited to surface water. If, however, roughly 5 to 20% of
spills reach surface water or groundwater (encompassing the U.S. EPA and California estimates
above), we would expect a spill to occur and reach a drinking water resource at approximately 0.05
to 2% of active or hydraulically fractured wells.1 This estimate of spills reaching drinking water
resources would be broadly consistent with estimates from the limited number of published studies
addressing this topic (e.g., Brantley etal.. 2014: Gross etal.. 20131.2 If a 0.05 to 2% frequency rate is
applied to the estimates of approximately 275,000 to 370,000 new wells hydraulically fractured
nationally between 2000 and part of 2013 and 2000 and part of 2014, respectively (Chapter 3), we
would expect roughly 140 to 7,400 spills to reach a drinking water resource during this almost 14-
to-15 year time-period. This would be approximately 10 to 500 spills per year reaching a drinking
water resource, dividing by the respective time periods. This large range reflects the high
uncertainty of these estimates and the lack of data on this topic.
Despite the data limitations and uncertainties surrounding estimates of spills, we can with more
certainty identify factors likely affecting the frequency of spills reaching drinking water resources.
These factors include spill characteristics, encompassing the volume of the chemical spilled; factors
related to the environmental fate and transport of the spill, such as properties of the chemical
spilled and characteristics of the site where the spill occurred; and finally, factors related to spill
prevention and response.
Everything else being equal, a larger volume spill will be more likely to reach a drinking water
resource than a smaller spill (U.S. EPA. 2015ml. On-site spills in the chemical mixing and produced
water handling stages are typically in the hundreds of gallons fU.S. EPA. 2015ml. Larger spills,
though less common, do occur. Well blowouts, pipeline leaks, and impoundment failures are
sources of some of the largest individual spill volumes. Well blowouts were responsible for the
1 Estimated by multiplying the 1 to 10% spill rate for active or hydraulically fractured wells by 5% to 20% for spills
reaching drinking water, and then reconverting to a percentage by multiplying by 100.
2 Brantley et al. ('2014') estimated approximately 0.4 to 0.8 spills per 100 hydraulically fractured wells reached surface
water in Pennsylvania between 2008 to September 2013. These were spills of 400 gal (1,514 L] or more, containing
hydraulic fracturing chemicals, fluids, or produced water. This might be an underestimate of spills reaching surface water
since spill volumes were limited to only 400 gal (1,514 L] or more. In estimate ofthe frequency of spills reaching
groundwater. Gross etal. ("20131 examined oil and produced water spills between July 2010 and July 2011 in Weld
County, Colorado. They counted 77 such spills reaching groundwater, approximately 0.4%> ofthe nearly 18,000 active
wells in the county.
10-9
-------
Chapter 10 - Synthesis
highest volume spills on average in 2015 in North Dakota. In Bradford County, Pennsylvania, a well
blowout resulted in a spill of approximately 10,000 gal (38,000 L) of produced water into a
tributary of Towanda Creek, a state designated trout fishery. The largest volume spill identified in
this assessment occurred in North Dakota, where approximately 2.9 million gal (11 million L) of
produced water spilled from a broken pipeline and impacted surface water and groundwater.
Though relatively rare compared to smaller volume spills, these types of spills are more likely to
reach—and therefore impact—a drinking water resource because they are of larger volumes.
By this same principle, produced water spills are more likely to impact drinking water resources
than chemical mixing spills. In an analysis of on-site spills, the median volume of produced water
spills was approximately twice as large as that in the chemical mixing stage (990 versus 420 gal, or
3,750 versus 1,590 L; U.S. EPA f2015mll Additionally, offsite, large pipeline spills of produced
water can occur. It is possible that spills of produced water are larger, in part, because they are less
likely to be stopped as quickly as spills in the chemical mixing stage. Spills in the chemical mixing
stage are likely to occur when people are on-site, and so the spills can be quickly addressed. In
contrast, spills of produced water may occur when no one is on-site or, in the case of pipelines, near
the off-site location of the spill. This may delay a response, allowing larger volumes to spill,
increasing the likelihood of the spill reaching a drinking water resource.
Properties of the chemicals spilled also affect the frequency of impacts. We identified or estimated
chemical and physical properties for almost half of the chemicals used in hydraulic fracturing fluids
between 2006 and 2013 (455 of the 1,084 chemicals). These were individual organic chemicals, not
inorganic chemicals, polymers, or mixtures. Volatility, solubility, and hydrophobicity/hydrophilicity
are three properties, among others, affecting whether a spill reaches a drinking water resource
(hydrophobic chemicals tend to repel or fail to mix with water, while hydrophilic chemicals tend to
mix with water). The vast majority of organic chemicals in hydraulic fracturing fluid do not readily
volatilize or evaporate, meaning these chemicals tend to remain in water if spilled. These chemicals
also vary widely in their solubility and hydrophobicity/hydrophilicity, defying a general
characterization. Nevertheless, of the 20 chemicals most frequently used according to our analysis
of FracFocus, most are highly soluble and hydrophilic, meaning they will be mobile if spilled
(Chapter 5). For example, methanol, isopropanol, and ethylene glycol are all likely to travel quickly
through the environment. Thus, these chemicals may more frequently reach drinking water
because of two unrelated, yet compounding factors: relatively high frequency of use in hydraulic
fracturing operations and relatively high mobility in the environment
Site characteristics are also an important factor determining whether a spill reaches a drinking
water resource (Figure 10-2). Site characteristics facilitating infiltration to groundwater are of
particular concern, since spills into groundwater are more likely to have severe impacts than those
into surface water (discussed in the severity section below). More permeable, sandier soils allow
greater infiltration of spilled fluids, whereas less permeable soils with more clay content can greatly
slow infiltration. More permeable rock also facilitates infiltration and movement of spills through
preferential flow paths—for example, in fractured or karst bedrock. Thus, sandier soils and more
permeable rock can increase the potential for spills to reach groundwater.
10-10
-------
Chapter 10 - Synthesis
Spilled Hydraulic
Fracturing Chemicals, Fluids
or Produced Water
Fluid Movement
Underground
Groundwater
Surface Water
Spill Characteristics
What chemicals were spilled?
How much was spilled?
Spill Response Activities
What actions were taken to remove the
spilled fluid from the environment?
Environmental Fate and Transport
How would the spilled fluid move
through the surface and
underground environment?
Figure 10-2. Fate and transport schematic for a spill of chemicals, hydraulic fracturing fluid, or
produced water.
Schematic shows the potential paths, transport processes, and factors governing potential impacts of spills to
drinking water resources.
There are spill prevention and response factors that reduce the frequency of impacts to drinking
water resources from spills. Spill containment systems include primary, secondary, and emergency
containment systems. Primary containment systems are the storage units, such as tanks or totes.
Secondary containment systems, such as liners and berms installed during site set-up, are intended
to contain spilled fluids until they can be cleaned up. Emergency containment systems, such as
berms, dikes, and booms, can be implemented temporarily in response to a spill. Remediation is the
action taken to clean up a spill and its affected environmental media. One of the most commonly
reported remediation activities is the removal of spilled fluid and/or affected media, typically soil
fU.S. EPA. 2015ml. Other remediation methods include the use of absorbent material, vacuum
trucks, flushing the affected area with water, and neutralizing the spilled material (U.S. EPA.
2015 ml. It was beyond the scope of this assessment to evaluate the implementation and efficacy of
spill prevention practices and spill response activities.
10.1.2.2 Severity
In addition to frequency, there are also factors affecting the severity of an impact on a drinking
water resource from a spill. For a given concentration, a larger volume spill will be more severe
than a smaller spill (see frequency section above for discussion of spill volumes). In addition to
10-11
-------
Chapter 10 - Synthesis
volume, the concentration and toxicity of the chemicals reaching a drinking water resource affect
severity, as well as site characteristics.
A spill with higher chemical concentrations will be more severe than a more dilute spill of equal
volume. In the chemical mixing stage, chemicals are stored in concentrated form on-site, prior to
diluting with abase fluid. Approximately 3,000 to 30,000 gal (11,000 to 114,000 L) of chemicals are
used per well on average, with up to twice that amount stored on site. If multiple wells are
fractured per site, tens to hundreds of thousands of gallons of chemicals are likely stored in
containers at a single site during the hydraulic fracturing of these wells. These storage containers
are a relatively frequent source of spills during the chemical mixing stage. Spills from these storage
containers, even if low in volume, may be severe if they reach a drinking water resource because
they often contain concentrated chemicals.
In the produced water handling stage, the severity of impacts from a spill also increases with higher
concentrations, especially if the spill reaches groundwater (see site characteristics below).
Produced water can vary substantially in chemical concentrations, including total dissolved solids
(TDS), metals, radioactive isotopes, and organic chemicals. Within the Marcellus Shale, for example,
produced water can range in TDS from less than 1,500 mg/L to over 300,000 mg/L (Rowan etal..
20111. By comparison, the average salinity concentration for seawater is 35,000 mg/L. The more
concentrated the produced water, the more likely impacts will be severe if a spill reaches a drinking
water resource. When a spilled fluid has greater concentrations of TDS than groundwater, the
higher-density fluid can move downward through the groundwater resource. Depending on the
flow rate and other properties of the groundwater, impacts from produced water spills can last for
years.
In addition to concentration, the toxicity of chemicals affects the severity of the impact if they enter
a drinking water resource. There were 37 chemicals listed in 10% or more of all FracFocus
disclosures between January 1, 2011 and February 28, 2013. Of these 37 chemicals, nine had
chronic oral reference values meeting the criteria used in this assessment.1 These nine chemicals
are associated with health effects including liver toxicity, kidney toxicity, developmental toxicity,
reproductive toxicity, and/or carcinogenesis. Chemicals used in hydraulic fracturing fluids and
detected in produced water will vary from site to site, so human health hazards are best evaluated
on a site-specific basis. Nevertheless, the multi-criteria decision analysis (MCDA) presented in
Chapter 9 highlighted certain chemicals that may have greater hazard potential. Propargyl alcohol,
2-butoxyethanol, and N,N-dimethylformamide are three such chemicals having relatively greater
hazard potential in the MCDA based on toxicity, frequency of use in hydraulic fracturing fluids, and
mobility in water.
Many of the chemicals in produced water are also known or suspected to cause cancer and/or non-
cancer health effects in humans. Associated health effects include liver toxicity, kidney toxicity,
neurotoxicity, reproductive and developmental toxicity, and carcinogenesis, based on the produced
1 The analysis of toxicity presented in Chapter 9 included chemicals regardless of accompanying concentration data in
FracFocus, and therefore listed 37 chemicals that were reported in 10% or more disclosures. Comparatively, Chapter 5
listed 35 chemicals that had valid concentration data from FracFocus and were reported in 10% of more disclosures.
10-12
-------
Chapter 10 - Synthesis
water chemicals having chronic oral reference values meeting the criteria used in this assessment.
Benzene, pyridine, and naphthalene are three of the chemicals highlighted in the MCDA as having
relatively greater hazard potential based on toxicity, measured concentrations in produced water,
and mobility in water.
We did not evaluate trends in chemical use by toxicity (e.g., the trends in the use of less toxic
chemicals). However, a more recent study of FracFocus data evaluated disclosures dating from
March 9, 2011, to April 13, 2015 fPavalu and Konschnik. 2016: Konschnik and Davalu. 20161. When
compared to the list of 1,084 chemicals used in hydraulic fracturing operations between 2005 and
2013 compiled for this assessment (Appendix H), an additional 263 chemicals were identified
(Chapter 5). Only one of these 263 chemicals was reported in more than 1% of disclosures. This
comparison of chemical lists does not address potential shifts in volumes of chemicals used, but it
does suggest that a shift to new types of chemicals-less toxic or otherwise-did not occur between
2013 and early 2015.
Finally, site characteristics also affect the severity of the impact Spills into groundwater are likely
to be more severe than spills into surface water, everything else being equal. This is not to say that
spills into surface water cannot be severe, especially in the immediate vicinity of the spill. For
instance, a tank overflowed on a well site in Kentucky spilling fluid into a nearby stream at
concentrations sufficient to kill fish in the area (Papoulias and Velasco, 2013). Chemicals can also
associate with stream sediments, forming a source of long-term contamination (e.g., radium). In
general, however, surface water dilutes a spilled chemical much more rapidly than groundwater.
Groundwater often moves slowly between areas of recharge and discharge. Groundwater
movement can be as slow as one foot per year or even one foot per decade (Alley et al.. 1999).
Because of this dynamic, chemicals from multiple spills can accumulate over time in groundwater.
Multiple chemical mixing and produced water spills, even if individually small, may impact a
groundwater resource in aggregate. Additionally, groundwater contamination may not be as readily
apparent as that in surface water because of the need to install monitoring wells to detect
contamination in groundwater. Lastly, groundwater can be difficult and expensive to remediate,
adding to the severity of impacts if spills reach groundwater (Alley etal., 1999).
10.1.3 Well Injection
Like the water acquisition, chemical mixing, and produced water handling stages, activities in the
well injection stage of the hydraulic fracturing water cycle can affect drinking water resources in
some instances. The well injection stage involves the injection of hydraulic fracturing fluids through
the production well and into the targeted rock formation at sufficient pressure to fracture the rock.
There are two fundamental pathways outlined in this assessment by which activities in the well
injection stage have the potential to affect drinking water resource quality. They are: (1) fluid
(meaning, liquid or gas) movement into a drinking water resource through defects or deficiencies
in the production well casing and/or cement; and (2) fluid movement into a drinking water
resource through the fracture network. The fluids potentially affecting drinking water resources
include hydraulic fracturing fluids, hydrocarbons (including methane gas), and naturally occurring
brines. The drinking water resources impacted directly in this stage are almost always groundwater
resources, rather than surface water.
10-13
-------
Chapter 10 - Synthesis
Though we could not in this assessment quantify an overall frequency of groundwater quality
impacts from the well injection stage, we can describe factors which make impacts more or less
frequent or more or less severe, as we did for other stages. We describe these factors below, first
with frequency and then severity. Within the frequency discussion, we address factors by each
pathway type.
10.1.3.1 Frequency
Pathway #1: Fluid movement into a drinking water resource through defects or deficiencies in the
production well casing and/or cement.
To reach and then fracture the production zone, an oil or gas well must first be drilled and
constructed down through the subsurface rock formations, often containing an overlying drinking
water resource. Since the well passes through the drinking water resource, this means defects or
deficiencies in the production well can lead to unintended movement of fluid into the drinking
water resource. This can occur regardless of the vertical separation between the drinking water
and the production zone.
The relatively brief hydraulic fracturing phase will likely impose the highest stresses to which the
well will be exposed during its entire life. If the well cannot withstand the stresses experienced
during hydraulic fracturing, the casing or cement can fail, resulting in the loss of mechanical
integrity and the unintended movement of fluids into the surrounding environment
A few studies have estimated rates of mechanical integrity failure of production wells resulting in
the loss of all barriers protecting the groundwater or in contamination of groundwater in areas
with hydraulic fracturing activity (Table 10-1). The estimates are all approximately equal to or less
than 1% of wells drilled or hydraulically fractured over varying time frames. For most of these
estimates, it is not possible to tell whether failures occurred during hydraulic fracturing or at some
other point in the well's life, with the exception of the EPA's Well File Review (U.S. EPA. 2015n). If
the failure rate from the Well File Review (0.5%) is applied to the estimates of 275,000 to 370,000
new wells hydraulically fractured nationally between 2000 and part of 2013 and 2000 and part of
2014, respectively (Chapter 3), we would expect roughly 1,370 to 1,850 mechanical integrity
failures during this time-period (almost 14 to 15 years). Dividing by each time period yields
approximately 100 to 125 mechanical integrity failures per year on average, resulting in the loss of
all barriers protecting the groundwater during hydraulic fracturing. These estimates also have a
high degree of uncertainty like the spills estimates. This not only stems from the lack of certainty
about failure rates, but also uncertainties surrounding the estimates of the number of wells
hydraulically fractured (Chapter 3). These are likely low estimates because they do not include
mechanical integrity failures occurring outside of the hydraulic fracturing process (e.g., during the
production phase), nor do they consider failures in re-fractured wells.
10-14
-------
Chapter 10 - Synthesis
Table 10-1. Literature estimates of mechanical integrity failure rates resulting in
contamination of groundwater or failure of all well barriers, potentially exposing the
groundwater.
Citation
Mechanical Integrity
Failure Rate (%)
Geographic
Scope
Key Findings & Description of Mechanical
Integrity Failure3,15
Fleckenstein et
al. (2015)
0.06
Colorado-
Wattenberg
Field
An overall catastrophic failure rate of 0.06% was found
for 16,828 wells studied (out of 17,948 total wells)
drilled in the Wattenberg Field between 1970 and 2013.
The timing of the failures was unknown, but most of the
failures occurred in the older wells. The Wattenberg
Formation is 4,400 ft (1,300m) below surface and
typically is hydraulically fractured. A catastrophic failure
was considered to have occurred when there was
contamination of drinking water aquifers (i.e., the
presence of thermogenic gas in a drinking water well)
and evidence of a well defect such as exposed
intermediate gas formations or casing leaks.
Considine et al.
(2012)
0.06
Pennsylvania
Two wells were cited between 2008 and 2011 by PA
DEP for causing methane migration into an aquifer. In
this same time period, 3,533 wells were drilled.
Brantlev et al.
(2014)
0.12-1.1
Pennsylvania
Based on positive determination letters (PDLs) for
violations that occurred between 2008 and 2012,
Brantley et al. estimated between 7 and 64 problematic
unconventional wells contaminated 85 properties.
Since PDLs are tied to drinking water wells and not gas
wells, Brantley et al. made assumptions about how
many unconventional gas wells were represented by
each PDL. This equates to problematic unconventional
gas wells compromising approximately 0.1 to 1% of the
6,061 wells spudded between 2008 and 2012.° Not all of
these PDLs may be due to mechanical integrity failures-
they could also be due to other causes, such as spills, or
methane migration from natural or other
anthropogenic sources.
Vidic et al.
(2013)
0.25
Pennsylvania
Of the 6,466 wells studied, 16 received notices
regarding contamination of groundwater with gas or
other fluids from the PA DEP associated with incidents
that occurred between 2008 and 2013.
10-15
-------
Chapter 10 - Synthesis
Citation
Mechanical Integrity
Failure Rate (%)
Geographic
Scope
Key Findings & Description of Mechanical
Integrity Failure3,15
U.S. EPA (2016c)
0.5
National
In an estimated 0.5% of the approximately 28,500
hydraulic fracturing jobs surveyed, a failure occurred
during hydraulic fracturing, such that there was no
additional barrier between the annular space with fluid
and the protected drinking water resource. While it
could not definitively be determined whether fluid
movement into the protected drinking water resource
occurred, in these cases, all of the protective barriers
intended to prevent such fluid migration had failed,
leaving the groundwater source vulnerable to
contamination.
a Note: While some information is available on the age of the wells studied, it is unclear whether the failure occurred during the
hydraulic fracturing event, with the exception of the U.S. EPA (2016c) study. In that study, the failures occurred during hydraulic
fracturing.
b While the Pennsylvania studies did not specifically identify whether the wells were involved in hydraulic fracturing operations,
a significant portion of Pennsylvania's recent oil and gas activity is in the Marcellus Shale; therefore, many of the wells in these
studies were most likely used for hydraulic fracturing.
cSpudding refers to starting the well drilling process by removing rock, dirt, and other sedimentary material with the drill bit
(U.S. EPA. 2013f).
Not all wells are equally likely to lose mechanical integrity; instead, there are factors that make
some wells more likely to experience a mechanical integrity failure than others. Well design and
construction are two such factors. First, a primary element of well design is the placement of at
least one additional layer of casing (besides the production casing) from the surface through the
lowest depth of the drinking water resource. This additional casing provides redundancy if the
production casing fails. In a study of 731 saltwater injection wells in the Williston Basin in North
Dakota, Michie and Koch (19911 found the risk of aquifer contamination from leaks into the
drinking water resource was 7 in 1,000,000 injection wells if a surface casing, in addition to the
production casing, was set deep enough to cover the drinking water resource. The risk increased to
6,000 per 1,000,000 wells (or 6 in 1,000) if this additional casing was not set deeper than the
bottom of the drinking water resource.
Second, fully cementing casing(s) through the entire drinking water resource affects the frequency
of impacts. Uncemented sections of surface casing increase the frequency of fluid leaks from the
well that can reach groundwater (Fleckenstein etal.. 2015: Watson and Bachu. 2009). The EPA's
Well File Review estimated that a portion of the protected groundwater resource identified by well
operators was uncemented in 3% of the wells surveyed fU.S. EPA. 2015nl With approximately
25,000 to 30,000 new wells hydraulically fractured a year (Chapter 3), this percentage means 750
to 900 of the wells used in hydraulic fracturing operations annually might lack this protection.
Adding re-fractured wells would increase the estimate of wells lacking this protection. Knowing the
depth of the groundwater resource at the point of drilling and then setting and cementing casings
below the lowest part of the drinking water resource can reduce the frequency or likelihood of an
impact
10-16
-------
Chapter 10 - Synthesis
Third, the well's casing, cement, and components need to be designed and constructed to withstand
the stresses applied to the well during hydraulic fracturing. In an example of inadequate well
construction, hydraulic fracturing of a gas well with insufficient and improperly placed cement in
Bainbridge Township, Ohio led to gas contamination of 26 domestic water supply wells and an
explosion in the basement of one of the nearby homes. This was due in part to a failure to cement
through the over-pressured gas formations and proceeding with the fracturing operation without
adequate cement fODNR. 20081. In another case, casings at an oil well near Killdeer, North Dakota,
ruptured in 2010 following a pressure spike during hydraulic fracturing, allowing fluids to escape
to the surface. Brine and tert-butyl alcohol were detected in two nearby water wells. Following an
analysis of potential sources, the only potential source consistent with the conditions observed in
the two impacted water wells was the ruptured well (U.S. EPA. 2015il
In addition to well design and construction, the degradation or corrosion of well components can
also increase the frequency of impacts to drinking water quality. Older wells exhibit more integrity
problems as cement and casings age. The EPA's Well File Review estimated at least 10% of the wells
represented in the national survey were greater than five years old at the time of hydraulic
fracturing. Hydraulic fracturing or re-fracturing older wells has the potential to increase the
frequency of casing or cement failures allowing unintended fluid migration into drinking water
resources.
Confirming well mechanical integrity can reduce the frequency of water quality impacts. Pressure
testing the casing used for hydraulic fracturing prior to the job can help detect problematic
casing—and provide an opportunity to make needed repairs if necessary. Monitoring the annular
space behind the casing used for hydraulic fracturing during the hydraulic fracturing job can detect
well component failure in real time and signal for an immediate shut down. Based on the EPA's Well
File Review study, casing pressure testing occurred at slightly less than 60% of the approximately
28,500 hydraulic fracturing jobs represented in that time frame (primarily 2009-2010) and annulus
monitoring took place during slightly more than 50% of these same jobs, implying these activities
did not always occur (U.S. EPA. 2016c). It is unclear whether the frequency of these practices have
changed since this time period.
Pathway #2: Fluid movement into a drinking water resource through the fracture network.
The other potential pathway for fluid movement into a drinking water resource is through the
fracture network. This could occur indirectly if the fracture network extends to a nearby well or its
fracture network, or to another permeable subsurface feature, such as natural fractures or faults,
which then allow the fluid to reach an underground drinking water resource. It could also occur
directly by the fracture network extending out of the production zone into a drinking water
resource, or hydraulic fracturing into a drinking water resource itself.1 Key factors affecting the
frequency of this pathway are the presence, distance, and condition of nearby wells; and the vertical
1 Hereafter, fractures extending out of the production zone are referred to as "out-of-zone" fractures, consistent with
Chapter 6.
10-17
-------
Chapter 10 - Synthesis
separation distance and the characteristics of the intervening rock between the production zone
and the drinking water resource.
Nearby wells (often called offset wells) can be a pathway for fluid movement, with hydraulic
fracturing fluid from one production well moving through the subsurface and entering another
nearby oil or gas well or its fracture network. These events are commonly referred to as "well
communication events" or "frac hits." The communication event might simply be registered as an
increase in pressure in the nearby well; yet there is also the possibility of damage to the nearby well
or its components, causing a surface spill or a subsurface release of fluids. The EPA's Well File
Review found 1% of the wells represented in the study experienced a frac hit, and the EPA spills
report identified 10 spills attributed to well communication events ("U.S. EPA. 2015m, n). It is
unknown whether any fluid reached a drinking water resource from these spills. Where active
nearby wells exist, operators of those wells can shut them in temporarily during the nearby
hydraulic fracturing to reduce the possibility of spills or damage to their wells, and therefore, the
potential for drinking water resource contamination.
The distance to the nearby well can affect the frequency of these communication events. In one
study, the likelihood of a frac hit was less than 10% in hydraulically fractured wells more than
4,000 ft (1,219 m) apart, while nearly 50% in wells less than 1,000 ft (300 m) apart (Aiani and
Kelkar. 20121. Distance was measured from the mid-point of each horizontal lateral. Thus, the
closer the nearby wells, the more likely a communication event
If nearby wells are in good condition and can withstand an increase in pressure, then an impact is
unlikely to occur. However, if the nearby well is not able to withstand the pressure of the fluid, well
components may fail and allow fluid to move into a drinking water resource. Because of this, nearby
older or abandoned wells are of particular concern. In older wells near a hydraulic fracturing
operation, plugs and cement may have degraded over time; in some cases, abandoned wells may
never have been plugged properly. This can be a significant issue in areas with legacy (i.e., historic)
oil and gas exploration. A Pennsylvania Department of Environmental Protection (PA DEP) report
cited three cases where migration of natural gas had been caused by well communication events via
old, abandoned wells fPA DEP. 2009cl. In Tioga County, Pennsylvania, following hydraulic
fracturing of a shale gas well, an abandoned well nearby produced a 30 ft (9 m) geyser of brine and
gas for more than a week (Dilmore etal.. 2015). Various studies estimate the number of abandoned
wells in the United States to be significant. For example, the Interstate Oil and Gas Compact
Commission flOGCC. 20081 estimates that approximately 1 million wells were drilled in the United
States prior to a formal regulatory system, and the status and location of many of these wells are
unknown. Hydraulic fracturing operators can reduce the possibility of impacts by identifying
nearby wells, and if necessary, plugging or otherwise addressing deficiencies in these wells.
If nearby wells serve as a pathway, fluid movement can bypass layers of intervening rock. In the
absence of this pathway, however, vertical distance and the intervening rock between the
production zone and the drinking water resource are factors affecting the possible movement of
fluid into a drinking water resource. The extension of fractures out of the oil and/or gas production
zone can—and does—occur. Examples have been reported in Greene County, Pennsylvania
fHammack etal.. 20141: at the Killdeer site in Dunn County, North Dakota fU.S. EPA. 2015il: and in
10-18
-------
Chapter 10 - Synthesis
other wells within the Bakken Shale ("Arkadakskiy and Rostron, 2013; Arkadakskiy and Rostron,
2012; Peterman etal., 2012"). In a study across several major shale formations, Davies etal. (2012)
found upward vertical fracture growth was often on the order of tens-to-hundreds of feet. One
percent of the fractures had a fracture height greater than 1,148 ft (350 m), and the maximum
fracture height among all of the data reported was 1,929 ft (588 m). This would suggest that
substantial vertical separation could preclude out-of-zone-fractures from directly reaching the
drinking water resource, although these measurements were only conducted in shale formations
and the extension of fractures is not the only way the drinking water resource could be
contaminated from out-of-zone fractures (see below). A modeling study also suggests fractures are
unlikely to extend from the production zone directly to a shallow drinking water resource in a deep
Marcellus-like environment (Kim and Moridis, 2015").
Not all fracturing occurs, however, with substantial vertical separation between the production
zone and the drinking water resource (Figure 10-3). The EPA's Well File Review found that 20% of
wells used for hydraulic fracturing had less than 2,000 ft (600 m) between the shallowest point of
fracturing and the base of the protected groundwater resource (U.S. EPA, 2015n). In coalbed
methane (CBM) plays, typically shallower than shale gas plays, these separation distances can be
smaller. For example, in the Raton Basin of southern Colorado and northern New Mexico,
approximately 10% of CBM wells have less than 675 ft (206 m) of separation between the
production zone and the depth of local water wells. In certain areas of the basin, this distance is less
than 100 ft (31 m) (Watts, 2006). Many of these areas are shallower in depth, and fracture growth
has been shown to be primarily horizontal, rather than vertical, at less than 2,000 ft (600 m) from
the surface (Fisher and Warpinski, 2012). Nevertheless, the possibility of an out-of-zone fracture
reaching a drinking water resource is more likely in a setting with less vertical separation than with
more.
Even if an out-of-zone fracture does not extend into a drinking water resource, it could connect to
other permeable subsurface features, such as natural fractures or faults, which could then connect
to a drinking water resource. Thus, properties of the intervening rock can also make this pathway
more or less frequent or likely. For instance, in the Pavillion gas field in Wyoming, there are no
laterally-continuous confining layers to prevent upward movement of fluids into the groundwater
(Digiulio and Jackson, 2016). While flow of subsurface fluids generally tends to be downward, local
areas of upward flow have been observed (Digiulio and lackson, 2016).
There are cases of hydraulic fracturing without vertical separation between the drinking water
resource and the production zone (Figure 10-3). The co-location of the oil or gas formation with the
drinking water resource is the factor affecting the frequency of an impact in these cases. Directly
fracturing into a drinking water resource causes an impact because it changes the quality of the
resource by introducing hydraulic fracturing fluids. The EPA's Well File Review found an estimated
0.4% of the wells represented in the study had perforations used for hydraulic fracturing shallower
than the base of the protected groundwater resource, as reported by well operators (U.S. EPA,
2015n). The EPA's Well File Review did not examine these instances by formation type. This
practice may be concentrated in locations in western states, especially in CBM plays. Examples
include the Raton Basin in Colorado (U.S. EPA, 2015k), the San Juan Basin of Colorado and New
10-19
-------
Chapter 10 - Synthesis
Mexico fU.S. EPA. 2004a). and the Powder River Basin of Montana and Wyoming fDahm etal.. 2011:
ALL Consulting. 2004: U.S. EPA. 2004a). This is a concern in the short term (should there be people
using these drinking water resources currently) and the long term (if drought or other conditions
necessitate the future use of these drinking water resources). For the most part in this chapter, we
focused on factors which can be managed, changed, or used to identify areas to target monitoring
efforts. In this situation, hydraulic fracturing directly into a drinking water resource would need to
cease if it was decided the resulting impacts to drinking water resource quality were unacceptable.
Targeted Rock Formation
Drinking Water Resource
No Vertical
x Separation
E " Distance
Drinking Water Resource
and Targeted Rock Formation
Targeted Rock Formation
15,000
10,000
5,000
(C)
, II
1 1 1
I I x m
^ J? ,/ V
-------
Chapter 10 - Synthesis
Lastly, the presence of gas, as opposed to liquids, in the subsurface may be a factor affecting the
frequency of impacts from fluid movement via defects or deficiencies in the well (pathway #1), or
through the fracture network (pathway #2). The low density of gas compared to liquids makes it
buoyant, which creates an upward drive toward the surface. Thus, gas found in the subsurface, such
as methane, can exploit pathways in a well (such as along a well lacking mechanical integrity), or in
the surrounding rock (such as induced or naturally occurring fractures). If a pathway exists and gas
is present, it can reach groundwater used for drinking. Consequently, gases could be more likely to
contaminate drinking water resources than liquids (Li etal.. 2016a).
10.1.3.2 Severity
The well injection chapter (Chapter 6) focused primarily on the potential for impacts to occur and
factors affecting frequency. By contrast, we have little-to-no information on factors affecting the
severity of impacts for this stage of the hydraulic fracturing water cycle. Severity would likely be
affected by the chemical composition of the fluid entering the drinking water resource; the volume
of the fluid; the duration in which that volume is delivered; and the concentration of the fluid and
its specific components, among other factors. Logically, the relatively simple pathway of a
mechanical integrity failure might result in the highest fluid volume delivered to a drinking water
resource over a short period of time—e.g., contamination of water wells in Bainbridge Township,
Ohio. By contrast, fluid movement through a fracture network, then through the intervening rock,
and finally into a drinking water resource may take a longer time and deliver a comparatively lower
volume. Even in this case, however, the impacts could still be severe if the fluid movement was to go
undetected and unaddressed.
10.1.4 Wastewater Disposal and Reuse
The last stage of the hydraulic fracturing water cycle is wastewater disposal and reuse. Produced
water from hydraulically fractured oil or gas production wells is managed predominantly through
disposal in underground Class II wells. Secondarily, it is disposed of via other practices, such as
discharge to surface waters or disposal in pits or evaporation ponds, or reused in other hydraulic
fracturing operations. Activities in the wastewater disposal and reuse stage of the hydraulic
fracturing water cycle can impact drinking water resources in some instances. Two such activities
are: the discharge of inadequately treated wastewater to surface water, and the storage or disposal
of wastewater in unlined pits or impoundments leading to contamination of surface water or
groundwater. In this section, we address factors increasing or decreasing the frequency or severity
of impacts from these activities. As in the water acquisition section, we combine our discussion of
frequency and severity here.
10.1.4.1 Frequency and Severity
Discharge of inadequately treated wastewater has impacted surface water. The quality of the
wastewater discharged is a factor affecting the frequency and severity of impacts. This factor is a
function of the chemical characteristics of the wastewater prior to treatment (i.e., the composition
and concentration of chemicals in the wastewater) and the efficacy of the treatment process. The
pre-2011 treatment of Marcellus wastewater in Pennsylvania illustrates this combination. In
10-21
-------
Chapter 10 - Synthesis
Pennsylvania before 2011, wastewater from shale gas operations was treated at centralized waste
treatment facilities (CWTs) and publicly owned treatment works (POTWs). The POTWs and some
CWTs at the time were not equipped to adequately treat high TDS wastewater. This resulted in
wastewater discharges containing elevated levels of TDS, including bromide and iodide, to surface
waters.
The elevated levels of TDS raised concerns about the formation of disinfection byproducts (DBPs)
after treatment at downstream drinking water facilities. Disinfection byproducts are formed when
organic material comes in contact with disinfectants (e.g., chlorine, chloramine, chlorine dioxide or
ozone). Many DBPs have long-term health effects including an increased risk of cancer, anemia,
liver and kidney effects, and central nervous system effects. Of particular concern are DBPs formed
in the presence of bromide or iodide, which are considered particularly toxic. Management of DBPs
places a burden on downstream drinking water utilities. Concerns regarding elevated TDS (in
particular high bromide) and the potential for formation of DBPs led the PA DEP to take steps in
2010 and 2011 to route Marcellus Shale wastewater away from POTWs and CWTs (that could not
treat for TDS) to alternate options such as disposal via injection wells, on-site reuse, or reuse after
limited treatment at CWTs. By 2014, only a small percentage (approximately less than 1%) of
Marcellus wastewater went to CWTs permitted to discharge to surface waters (Figure 10-1).
Additionally, the new EPA pretreatment standards prohibit oil and gas operators from sending
unconventional oil and gas wastewater directly to POTWs fU.S. EPA. 2016dl.
The combination of wastewater composition and inadequate treatment have also resulted in the
discharge of other constituents such as barium, strontium, and radium into surface waters in
Pennsylvania. Marcellus Shale wastewater contains radium, naturally occurring in the subsurface
formation. Radium has been found in stream sediments at discharge points for POTWs and CWT
facilities that have accepted Marcellus Shale wastewater. The ratio of radium isotopes (radium-228
to radium-226) in these sediments is consistent with ratios in Marcellus Shale wastewater (Warner
etal.. 2013a). Radium-226, with a half-life of approximately 1,600 years, causes long-term
contamination. The practice of management of wastewaters via POTWs and CWTs without TDS
removal has declined, yet it remains uncertain whether the discharge of radionuclides to surface
waters from the oil and gas industry in Pennsylvania has ceased entirely (PA DEP. 2015b).
The storage or disposal of wastewater in pits or impoundments can also lead to contamination of
surface water or groundwater resources. This can occur via surface spills or overflows. It can also
occur via infiltration into the soil and percolation to groundwater through the pit itself. Whether
the pit or impoundment is lined is an important factor affecting the frequency and severity of
impacts on groundwater due to subsurface leaching. Historically, unlined pits have been used to
dispose of wastewater via percolation (or evaporation). While this practice has been banned in
most states, it is allowed in certain locations or instances (e.g., storage of wastewater, but not
disposal) as of July 2016. Even where prohibited, unpermitted unlined disposal or storage pits exist
For example, approximately 1,000 unlined storage or disposal pits of oil and gas wastewater are
located in the Central Valley Region of California (California State Water Resources Control Board.
2016: Esser etal.. 2015). Of these, approximately 60% were still active as of July 2016, and roughly
20% of those pits lacked permits fCA Water Board. 20161.
10-22
-------
Chapter 10 - Synthesis
Unlined pits have been shown to cause contamination of drinking water resources. The presence of
BTEX (benzene, toluene, ethylbenzene, and xylenes) and other organics in groundwater are linked
to pits in California and New Mexico f California Regional Water Quality Control Board Central
Valley Region. 2015: Sumi. 2004: Eiceman. 19861. Groundwater impacts downgradient of an
unlined pit in Oklahoma included high salinity (3500-25,600 mg/L) and the presence of volatile
organic compounds (Kharaka et al.. 20021. Impacts can also occur in the case of disposal of
relatively low TDS wastewater fHealv etal.. 2011: Healv etal.. 20081. For example, a CBM
wastewater impoundment in the Powder River Basin of Wyoming resulted in high concentrations
of TDS, chloride, nitrate, and selenium in the groundwater (Healv etal.. 2011: Healv etal.. 2008).
Total dissolved solids exceeded 100,000 mg/L in one groundwater sample, despite the much lower
concentrations (2,300 mg/L) in the wastewater being discharged (Healv etal.. 2008). Most of the
solutes found in the groundwater did not originate with the CBM wastewater, but rather resulted
from dissolution of previously existing salts and minerals in the subsurface. Lining pits or using
closed-loop systems (i.e., tanks) can decrease the frequency of such impacts.
10.1.5 Summary
In the above section, we synthesized the information in this assessment by discussing factors
increasing or decreasing the frequency or severity of impacts from activities in the hydraulic
fracturing water cycle. We focused particularly on factors that could be managed, changed, or used
to identify locations for additional monitoring or alteration of practices. Based on the information
reviewed, we conclude the following combinations of activities and factors are more likely than
others to result in more frequent or more severe impacts:
• Water withdrawals for hydraulic fracturing in times or areas of low water availability,
particularly in areas with limited or declining groundwater resources;
• Spills during the management of hydraulic fracturing fluids and chemicals or produced
water that result in large volumes or high concentrations of chemicals reaching
groundwater resources;
• Injection of hydraulic fracturing fluids into wells with inadequate mechanical integrity,
allowing gases or liquids to move to groundwater resources;
• Injection of hydraulic fracturing fluids directly into groundwater resources;
• Discharge of inadequately treated hydraulic fracturing wastewater to surface water
resources; and
• Disposal or storage of hydraulic fracturing wastewater in unlined pits, resulting in
contamination of groundwater resources.
Conversely, the scientific literature and data provide evidence that certain factors can reduce the
frequency or severity of impacts. Based on the information reviewed in this assessment, we
conclude the following factors are likely to reduce the frequency or severity of impacts:
• Passby flows, or low-flow criteria, for surface water withdrawals, and the use of brackish
groundwater or reused wastewater as substitutes for fresh water withdrawals;
10-23
-------
Chapter 10 - Synthesis
• Implementation of spill prevention and response measures;
• Design and placement of well casing and cement able to withstand the stresses imposed by
hydraulic fracturing (including identifying the depth of the drinking water resource at the
point of drilling, and setting and cementing casings through the entire drinking water
resource);
• Confirming mechanical integrity of oil and gas wells prior to, during, and after hydraulic
fracturing, and correcting deficiencies if necessary;
• Identification of active or abandoned wells near hydraulic fracturing operations and, if
necessary, adjustment of the operations to minimize well-to-well communication and/or
plugging improperly abandoned wells;
• The use of treatment technologies to remove TDS, and other constituents, such as radium,
when present prior to discharge; and
• Storage of wastewater in lined pits or the use of closed-loop systems instead of pits.
The above factors are not the only factors that can reduce the frequency or severity of impacts, yet
are the ones most emphasized by the information reviewed for this assessment. It should be noted
that the above factors reduce, but do not completely eliminate, the possibility of an impact
occurring. In the case of hydraulic fracturing directly into a drinking water resource or disposal of
wastewater via unlined pits, we did not identify factors which could reduce the frequency or
severity of impacts, besides restricting the activity itself.
10.2 Uncertainties and Data Gaps
In this assessment, we identified impacts on drinking water resources in all stages of the hydraulic
fracturing water cycle and described the factors affecting the frequency or severity of impacts. The
major conclusions presented above (in Section 10.2.5) are the strongest conclusions based on data
and information synthesized for the assessment.
There were also many areas within the assessment for which strong conclusions could not be
reached. This was because of the lack of publicly available data and large uncertainties in available
sources of information. Below, we provide perspective on what data gaps and uncertainties
prevented us from drawing additional strong conclusions about the potential for impacts and/or
the factors affecting the frequency or severity of impacts.
We encountered uncertainties associated with, and gaps in, aggregated, publicly accessible
information about both activities in the hydraulic fracturing water cycle and groundwater data. In
general, comprehensive information on the location of activities in the hydraulic fracturing water
cycle is lacking, either because it is not collected, not publicly-available, or prohibitively difficult to
aggregate. Thus, we lacked complete information on the geographic locations of well sites (both
new and existing) where the chemical mixing, well injection, and produced water handling stages
take place; the depth(s) of zones that have been hydraulically fractured in these wells; where water
is being acquired (i.e., the source water) for hydraulic fracturing; and where hydraulic fracturing
wastewater is treated and/or disposed. FracFocus provided data on well locations, and water and
10-24
-------
Chapter 10 - Synthesis
other chemicals used at those locations. However, reporting to FracFocus at the time period studied
was not always required, making it difficult to determine the completeness or representativeness of
the information.
In addition, there are uncertainties about where groundwater resources are located. This includes
the thickness of the resource, from its top to its lowest depth, and its relation to the shallowest
depth where hydraulic fracturing occurred. If comprehensive data about the locations of both
drinking water resources and activities in the hydraulic fracturing water cycle were available, it
would have been possible to more completely identify areas in the United States where hydraulic
fracturing-related activities and drinking water resources overlap.
There are also uncertainties and data gaps related to chemicals used in hydraulic fracturing fluid
and those detected in produced water. Some chemicals and chemical mixtures remain undisclosed
because of confidential business information (CBI) claims. Well operators claimed at least one
chemical as CBI at more than 70% of disclosures reported to FracFocus between 2011 and early
2013. Data suggests this practice is increasing. Konschnik and Davalu (2016) reported that 92% of
FracFocus disclosures submitted between approximately March 2011 and April 2015 included at
least one chemical claimed as confidential. When chemicals are claimed as CBI, there is no public
means of accessing information on these chemicals. Furthermore, many of the chemicals and
chemical mixtures disclosed, or those detected in produced water, lack information on properties
affecting their movement, persistence, and toxicity in the environment should they be spilled.
Better information on these chemicals would allow for a more robust evaluation of potential human
health hazards posed, and thus a better understanding about the severity of impacts should the
chemicals reach drinking water resources.
In places where we know hydraulic fracturing water cycle activities have occurred, data to assess
impacts are often either not collected or are not publicly available in accessible forms. Specifically,
local water quality monitoring and well mechanical integrity integrity data are not consistently
collected or readily available. In particular, sufficient baseline data on local water quality are
needed to quantify any changes post-hydraulic fracturing. There are exceptions to this, for example,
the state of California recently implemented a plan to make water quality monitoring information
public (Text Box 10-1). In general, however, the limited amount of data collected before, during,
and after hydraulic fracturing activities and made public, reduces the ability to determine whether
hydraulic fracturing affected drinking water resources.
Text Box 10-1. Hydraulic Fracturing and Groundwater Quality Monitoring in California.
In July 2015, the California Water Resources Control Board adopted Senate Bill 4 (SB4), Model Criteria for
Groundwater Monitoring in Areas of Oil and Gas Well Stimulation. This resolution directed the establishment of
a "comprehensive regulatory groundwater monitoring and oversight program...in order to assess the
potential effects of well stimulation treatment activities on the state's groundwater resources" fCalifornia
State Water Resources Control Board. 2015). The adoption of SB4 concluded a multi-year process, which
incorporated stakeholder engagement, review by the public, and consultation with an expert scientific panel.
(Text Box 10-1 is continued on the following page.)
10-25
-------
Chapter 10 - Synthesis
Text Box 10-1 (continued). Hydraulic Fracturing and Groundwater Quality Monitoring in
California.
The recommendations of the expert panel informed the creation and implementation of SB4 with respect to
criteria "to be used for assessment, sampling, analytical testing, and reporting of water quality associated
with oil and gas well stimulation activities" (Esser et al.. 20151
The resolution requires two different scales of groundwater monitoring for different purposes. First, it
requires well-by-well (also called "area-specific") groundwater monitoring by well operators. This includes
groundwater monitoring conducted for all hydraulic fracturing projects initiated after July 2015. Each oil or
gas production well operator must submit a design and timeline for monitoring groundwater resources in
proximity to its proposed well. The State Water Resources Control Board approves the monitoring plan
before hydraulic fracturing can proceed. The groundwater monitoring plan must include:
• The installation of monitoring wells within 0.5 miles of the wellhead. At least one monitoring well must
be upgradient of the production well and two monitoring wells must be downgradient. Should the
production well penetrate more than one protected groundwater resource (as defined by the resolution),
monitoring wells must facilitate sampling of at least one that is shallow and one that is deep.
• A monitoring timeline that includes sampling prior to production well construction and hydraulic
fracturing, as well as semi-annual sampling after completion.
• A list of water quality parameters and constituents to be monitored, including TDS, specific metals, and
specific organic compounds.
The area-specific monitoring requirements also include submission of information by well operators about
geologic and human-made features in the subsurface that could serve as pathways for impacts to
groundwater, aspects of production well construction, and hydraulic fracturing fluid composition.
Second, a regional groundwater monitoring program will document trends in baseline water quality and
locate protected groundwater state-wide. In addition to monitoring for trends in groundwater quality related
to activities at well sites, it will also be designed to detect trends related to impacts from wastewater disposal
practices.
All data from the monitoring programs will be publicly accessible in a state-maintained database. The
database is intended to support public health, scientific, and academic needs, as well as future "investigation,
assessment, and research relevant to oil and gas development impacts on groundwater quality" (Esser et al..
20151
Together, the data and information collected and made publicly available as part of the area-specific and
regional groundwater monitoring in California will help fill data gaps identified in this section of the
assessment by locating groundwater resources, monitoring drinking water resources at spatial and temporal
scales relevant for detecting impacts from activities in the hydraulic fracturing water cycle, and distinguishing
impacts from hydraulic fracturing activities from the impacts of other potential sources.
In the cases where effects are suspected, it is often difficult to separate the potential effects of
hydraulic fracturing activities from effects of broader oil and gas industry activities and other
industries or causes. The use of long-lasting, mobile tracer chemicals added to hydraulic fracturing
fluids to monitor for impacts has been proposed (Kurose. 2014). but has received relatively little
attention in the scientific literature as of mid-2016. Instead, measured changes in water quality
parameters can be associated with, but not necessarily diagnostic of, impacts from hydraulic
fracturing activities. For instance, measurable changes in methane levels, TDS, ratios of geochemical
10-26
-------
Chapter 10 - Synthesis
constituents, and isotopic ratios might suggest impacts from hydraulic fracturing but could also be
from other sources, either natural or anthropogenic. To try to assign a cause, these measurements
often have to be followed with further collection of evidence supporting or refuting hydraulic
fracturing activities as the cause of the changes. (See Text Box 10-2 for discussion of causal
assessments.)
Text Box 10-2. Causal Assessment and Hydraulic Fracturing Water Cycle Activities.
A number of recent studies have conducted regional-scale assessments of trends in water quality in areas
with hydraulic fracturing activity, showing either no trend or trends linked temporally or spatially with
hydraulic fracturing (Burton et al.. 2016: Hildenbrand et al.. 2016: Hildenbrand et al.. 2015: Siegel etal.. 2015:
Darrah et al.. 2014: Fontenot et al.. 2013: Warner et al.. 2013bl Regional assessments can be important for
integrating information over broader scales, and for posing hypotheses about how hydraulic fracturing water
cycle activities may impact drinking water resources. Oftentimes, however, activities in the hydraulic
fracturing water cycle are merely one of several possible causes of an observed change in water quality or
quantity at a specific site. In this case, more thorough, site-specific investigations are often necessary. Causal
assessment (also called causal analysis) involves collecting multiple kinds of evidence to determine which of
several possible causes of contamination is most likely.
Causal assessment requires several steps. First, the spatial and temporal scope of the issue is defined,
including a description of all the possible causes of an observed impact, in this case the change in quality or
quantity of a drinking water resource. Once this is done, evidence is collected and assembled to support or
refute the potential causes. Evidence indicating how a potential cause and an observed effect are related in
time can help support or refute potential causes. Other kinds of evidence can also be useful in identifying a
cause, including: determining whether the composition and volume of a leaked, spilled, or treated and
discharged fluid are capable of causing observed impacts on water quality; and determining whether a
physical pathway between a well or well site exists by measuring the mechanical integrity of hydraulically
fractured wells and/or establishing the presence/absence of a contaminant plume.
Ideally, the evidence helps exclude possible causes of the reported contamination, narrowing down the list of
potential causes to a single cause. Causal assessments can take a long time to complete and can require a lot
of resources and expertise. In some situations, available data and resources are simply not sufficient to
definitively identify the cause. Nevertheless, causal assessments conducted in a consistent and transparent
way can help enable the identification of the likely cause(s) of contamination of drinking water resources.
The retrospective case studies conducted by the EPA under the Study Plan are examples of scientific
investigations using a multiple lines of evidence approach consistent with the principles of causal assessment
(U.S. EPA. 2015i. i, L 2014f. g). These case studies were cited throughout this report. For instance, as noted
previously, the Killdeer, North Dakota case study found that an inner string of casing burst during hydraulic
fracturing of an oil well, resulting in a release of hydraulic fracturing fluids and formation fluids that impacted
a groundwater resource (U.S. EPA. 2015i). Following an analysis of potential sources, the only potential
source consistent with the conditions observed was the ruptured well (U.S. EPA. 2015i).
Regardless of whether a single cause can be determined, actions can still be taken to mitigate one or more
potential causes of contamination. Information gained once the suspected activity has been halted or at least
reduced could elucidate whether hydraulic fracturing operations are more or less likely to have been the
source of the contamination.
Many members of the public are interested in understanding the national frequency of impacts to
drinking water resources from activities across the entire hydraulic fracturing water cycle. Because
10-27
-------
Chapter 10 - Synthesis
of the significant data gaps and uncertainties in the available data, it was not possible to estimate
the national frequency of impacts to drinking water resources from activities in the hydraulic
fracturing water cycle collectively. We were, however, able to estimate impact frequencies in some,
limited cases within the larger hydraulic fracturing water cycle (i.e., spills of hydraulic fracturing
fluids or produced water, and mechanical integrity failures). These more specific estimates had a
high degree of uncertainty, but were the best estimates that could be provided with the data and
literature currently available.
Finally, it should be recognized that this assessment is a snapshot in time. Our understanding of the
factors affecting the frequency or severity of impacts may change in the future as industry practices
evolve and new information becomes available.
10.3 Use of this Assessment
This assessment contributes to the understanding of the potential impacts to drinking water
resources by activities in the hydraulic fracturing water cycle and the factors influencing those
impacts. The scientific information presented can be used by federal, tribal, state, and local officials;
industry; and the public to better understand and address vulnerabilities of drinking water
resources to activities in the hydraulic fracturing water cycle.
The uncertainties and data gaps identified throughout this assessment could be used to identify
future data collection efforts. Data collection efforts could include, for example, surface water and
groundwater monitoring programs in areas with hydraulically fractured oil and gas production
wells; collection and the public dissemination of data on the condition of hydraulically fractured
wells; or targeted research programs to better characterize the environmental fate and transport
and human health hazards associated with chemicals in the hydraulic fracturing water cycle. Data
collected and analyzed through new data collection efforts may identify new factors increasing or
decreasing the frequency or severity of impacts.
In the near term, decision-makers could focus their attention on the combinations of activities and
factors that we conclude are more likely than others to result in more frequent or more severe
impacts (Section 10.2.5). By focusing attention on the above combinations, impacts to drinking
water resources from activities in the hydraulic fracturing water cycle can be prevented or reduced.
Overall, the practice of hydraulic fracturing is expanding and continues to change. Oil and gas
production associated with hydraulic fracturing was insignificant in 2000, but by 2015 it accounted
for an estimated 51% of U.S. oil production and 67% of U.S. gas production (EIA. 2016c. d). The
number of wells drilled and hydraulically fractured is likely to continue to increase in the coming
decades fEIA. 2014a], The work of evaluating potential impacts from combinations of activities and
factors in the hydraulic fracturing water cycle will need to keep pace with this industry and as new
scientific studies are produced. This assessment provides a foundation for those efforts, while
offering information to support the reduction of current vulnerabilities of drinking water resources.
10-28
-------
Chapter 11 - References
Chapter 11. References
11-1
-------
Chapter 11 - References
This page is intentionally left blank.
11-2
-------
Chapter 11 - References
Chapter 11. References
Hyperlinks to the reference citations throughout this document will take you to the ORD National
Center for Environmental Assessment HERO database (Health and Environmental Research Online)
at https:/ /hero.epa.gov/hero. HERO is a database of scientific literature used by the U.S. EPA in the
process of developing selected science assessments.
Abou-Saved. IS: Sorrell. MA: Foster. RA: Atwood. EL: Youngblood. DR. (2011). Haynesville Shale
development program: From vertical to horizontal. In 2011 SPE America's Unconventional
Gas Conference, 14-16 June, 2011, The Woodlands, TX: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/144425-MS
Abriola. LM: Pinder. GF. (1985a). A multiphase approach to the modeling of porous media
contamination by organic compounds: 1. Equation development Water Resour Res 21: 11-
18. http://dx.doi.org/10.1029/WR021 iOOlpOOOl 1
Abriola. LM: Pinder. GF. (1985b). A multiphase approach to the modeling of porous media
contamination by organic compounds: 2. Numerical simulation. Water Resour Res 21: 19-
26.
Adachi. 1: Siebrits. E: Peirce. A: Desroches. I. (2007). Computer simulation of hydraulic fractures.
International Journal of Rock Mechanics and Mining Sciences 44: 739-757.
http: / /dx. doi. or g/10.1016 /i. ii rmms. 2 006.11.006
Agbalagba. EO: Avwiri. GO: Chadumoren. YE. (2013). Radiological impact of oil and gas activities in
selected oil fields in production land area of Delta State, Nigeria. J Appl Sci Environ Manag
17: 279-288. http://dx.doi.org/10.4314%2Fiasem.vl7i2.ll
Aghakouchak. A: Cheng. L: Mazdivasni. 0: Farahmand. A. (2014). Global warming and changes in
risk of concurrent climate extremes: Insights from the 2014 California drought. Geophys
Res Lett 41: 8847-8852. http://dx.doi.org/10.1002/2014GL062308
Ajani. A: Kelkar. M. (2012). Interference study in shale plays. Presented at SPE Hydraulic Fracturing
Technology Conference, February 6-8, 2012, The Woodlands, TX.
Akob. DM: Cozzarelli. IM: Dunlap. PS: Rowan. EL: Lorah. MM. (2015). Organic and inorganic
composition and microbiology of produced waters from Pennsylvania shale gas wells. Appl
Geochem 60: 116-125. http://dx.doi.Org/10.1016/j.apgeochem.2015.04.011
Akob. DM: Mumford. AC: Orem. W: Engle. MA: Klinges. TG: Kent. DB: Cozzarelli. IM. (2016).
Wastewater disposal from unconventional oil and gas development degrades stream quality
at a West Virginia injection facility. Environ Sci Technol 50: 5517-5525.
http://dx.doi.org/10.1021/acs.est6b0Q428
Al-Ghazal. M: Al-Driweesh. S: Al-Shammari. F. (2013). First successful application of an
environment friendly fracturing fluid during on-the-fly proppant fracturing. Presented at
International Petroleum Technology Conference, March 26-28, 2013, Beijing, China.
http://www.earthdoc.org/publication/publicationdetails/?publication=69805
Al-Tubori. A: Johnston. S: Bover. C: Lambert. SW: Bustos. OA: Pashin. TC: Wrav. A. (2009). Coalbed
methane: Clean energy for the world. Oilfield Rev 21: 4-13.
Alanco. (2012). New subsidiary Alanco Energy Services, Inc. to provide produced water disposal
services to natural gas industry, http: //www.alanco.com/news 040912.asp
11-3
-------
Chapter 11 - References
Ali. M: Taoutaou. S: Shafqat. AU: Salehapour. A: Noor. S. (2009). The use of self-healing cement to
ensure long term zonal isolation for HPHT wells subject to hydraulic fracturing operations
in Pakistan. Presented at International Petroleum Technology Conference, December 7-9,
2009, Doha, Qatar.
Ali. SA: Clark. WT: Moore. WR: Dribus. TR. (2010). Diagenesis and reservoir quality. Oilfield Rev 22:
14-27.
AlKhowaildi. MA: AlGhazal. MA: Al-Driweesh. S: Abbad. E: Abdrabalnabi. H. (2016). Eco-friendly
hydraulic fracturing fluid: Field deployment and performance evaluation in Saudi Arabia's
tight gas reservoirs. In SPE Asia Pacific Hydraulic Fracturing Conference, 24-26 August,
2016, Beijing, China: Society of Petroleum Engineers, http: //dx.doi.org/10.2118/181861-
MS
ALL Consulting (ALL Consulting, LLC). (2004). Coal bed methane primer: New source of natural gas
and environmental implications. Tulsa, OK: U.S. Department of Energy, National Petroleum
Technology Center. http://bogc.dnrc.mtgov/PDF/Web%20Version.pdf
ALL Consulting (ALL Consulting, LLC). (2012). The modern practices of hydraulic fracturing: A
focus on Canadian resources. Tulsa, Oklahoma: ALL Consulting LLC.
Allard. S: Nottle. CE: Chan. A: Toll. C: von Gunten. U. (2013). Ozonation of iodide-containing waters:
Selective oxidation of iodide to iodate with simultaneous minimization of bromate and I-
THMs. Water Res 47: 1953-1960. http: //dx.d0i.0rg/l 0.1016/i.watres.2012.12.002
Allen. TE. (1994). Pregel blender prototype designed to reduce cost and environmental problems.
Presented at Permian Basin Oil and Gas Recovery Conference, March 16-18,1994, Midland,
Texas. https://www.onepetro.org/conference-paper/SPE-27708-MS
Alley. B: Beebe. A: Rodgers. 1: Castle. TW. (2011). Chemical and physical characterization of
produced waters from conventional and unconventional fossil fuel resources. Chemosphere
85: 74-82. http://dx.doi.Org/10.1016/i.chemosphere.2011.05.043
Alley. WM: Reillv. TE: Franke. PL. (1999). Sustainability of ground-water resources. (U.S. Geological
Survey Circular 1186). Denver, CO: U.S. Geological Survey.
http://pubs.usgs.gov/circ/circll86/
Allison. D: Folds. PS: Harless. DT: Howell. M: Vargus. GW: Stipetich. A. (2009). Optimizing openhole
completion techniques for horizontal foam-drilled wells. In 2009 SPE Eastern Regional
Meeting: Limitless Potential/Formidable Challenges, 23-25 September, 2009, Charleston,
WV: Society of Petroleum Engineers. http: //dx.doi.org/10.2118 /125642-MS
Alvarez-Cohen. L: Speitel. GE. (2001). Kinetics of aerobic cometabolism of chlorinated solvents.
Biodegradation 12: 105-126. http://dx.d0i.0rg/l0.1023/A:l012075322466
Alv. M: Clancev. B: Montgomery. IK: Bugti. MA: Ahmadzamri. AF: Al-Meraikhi. AA: Lawson. M.
(2015). Geochemical applications for identifying the source of hydrocarbons in well annuli.
Presented at International Petroleum Technology Conference, December 6-9, 2015, Doha,
Qatar.
AME (Acton Mickelson Environmental, Inc.). (2015). Pavillion, Wyoming area domestic water wells
draft final report and palatability study (Draft). (AME Project Number 90081). Cheyenne,
WY: Wyoming Department of Environmental Quality.
http://deq.wyoming.gov/media/attachments/Water%20Quality/Pavillion%20Investigatio
n/Draft%20Report/01 Pavillion%20WY%20Area%20Domestic%20Water%20Wells%20D
raft%20Final%20Reportpdf
11-4
-------
Chapter 11 - References
AME (Acton Mickelson Environmental, Inc.). (2016). Pavillion, Wyoming area domestic water wells
final report and palatability study. Cheyenne, WY: Wyoming Department of Environmental
Quality, http://deq.wyoming.gov/wqd/pavillion-investigation/resources/investigation-
final-report/
AMEC Environment & Infrastructure. Inc.. Hinckley Consulting. HDR Engineering. Inc.. (2014).
Hydrogeologic study of the Laramie County control area. Prepared for the Wyoming State
Engineers Office. Cheyenne, WY: Wyoming State Engineer's Office, http: //seo.wyo.gov/seo-
files/Final%20Draft%20Corrected%20Stamped.pdf?attredirects=0&d=l
Aminto. A: Olson. MS. (2012). Four-compartment partition model of hazardous components in
hydraulic fracturing fluid additives. Journal of Natural Gas Science & Engineering 7: 16-21.
http://dx.doi.org/DOI: 10.1016 /i.ingse.2012.03.006
Anderson. PB: Vanden Berg. MP: Carney. S: Morgan. C: Heuscher. S. (2012). Moderately saline
groundwater in the Uinta Basin, Utah. (Special Study 144). Salt Lake City, UT: Utah
Geological Survey.
Andrews. RD: Holland. A. (2015). Summary statement on Oklahoma seismicity April 21, 2015.
http://earthquakes.ok.gov/wp-
content/uploads/2015/04/QGS Summary Statement 2015 04 20.pdf
Anning. DW: Flvnn. ME. (2014). Dissolved-solids sources, loads, yields, and concentrations in
streams of the conterminous United States. (Scientific Investigations Report 2014-5012).
Reston, VA: U.S. Geological Survey, http://dx.doi.org/10.3133/sir20145012.
ANRC (Arkansas Natural Resources Commission). (2014). Non-riparian water use certification.
http://anrc.ark.org/divisions/water-resources-management/non-riparian-water-use-
certification-program/
API (American Petroleum Institute). (2000). Overview of exploration and production waste
volumes and waste management practices in the United States. Washington, D.C.: API
Publishing Services, http: //www.api.org/environment-health-and-safetv/environmental-
performance/~/media/Files/EHS/Environmental Performance/ICF-Waste-Survev-of-
EandP-Wastes-2000.ashx
API (American Petroleum Institute). (2009). Hydraulic fracturing operations - Well construction
and integrity guidelines [Standard] (1st ed.). Washington, D.C.: API Publishing Services.
http://www.shalegas.energy.gov/resources/HFl.pdf
API (American Petroleum Institute). (2010). Isolating potential flow zones during well construction
[Standard] (lsted.). (RP 65-2). Washington, D.C.: API Publishing Services.
http://www, techstreetcom/products/preview/1695866
API (American Petroleum Institute). (2011). Practices for mitigating surface impacts associated
with hydraulic fracturing. Washington, D.C.: API Publishing Services.
http://www.shalegas.energy.gov/resources/HF3 e7.pdf
API (American Petroleum Institute). (2015). Hydraulic fracturing well integrity and fracture
containment. Washington, D.C.: API Publishing Services.
Arkadakskiv. S..: Rostron. BT. (2012). Stable isotope geochemistry helps in reducing out-of-zone
hydraulic fracturing and unwanted brine production from the Bakken Reservoir.
http://isobrine.eom/wp-content/uploads/2013/04/Arkadakskiy Rostron 20121.pdf
Arkadakskiv. S: Rostron. B. (2013a). Tracking out-of-zone hydraulic fracturing in the Bakken with
naturally occuring tracers. Presented at GeoConvention 2013: Integration Canadian Society
of Petroleum Geologists, May 6-10, 2013, Calgary, AB, Canada.
11-5
-------
Chapter 11 - References
http://geoconvention.org/archives/2013abstracts/381 GC2013 Tracking Out-of-
zone Hydraulic Fracturing.pdf
Arkansas DEO (Arkansas Department of Environmental Quality). (2012). Complaints and
inspections database [Database], Little Rock, AR.
http://www.adeq.state.ar.us/home/pdssql/complaints inspections.aspx
Arthur. ID: Bohm. B: Cornue. D. (2009a). Environmental considerations of modern shale gas
development Presented at SPE Annual Technical Conference and Exhibition, October 4-7,
2009, New Orleans, LA. https://www.onepetro.org/conference-paper/SPE-122931-MS
Arthur. ID: Bohm. B: Coughlin. BT: Lavne. M. (2009b). Evaluating the environmental implications of
hydraulic fracturing in shale gas reservoirs. Presented at SPE Americas E&P Environmental
and Safety Conference, March 23-25, 2009, San Antonio, TX.
https://www.onepetro.org/conference-paper/SPE-121038-MS
Arthur. ID: Bohm. B: Lavne. M. (2008). Hydraulic fracturing considerations for natural gas wells of
the Marcellus Shale. Presented at The Ground Water Protection Council Annual Forum,
September 21-24, 2008, Cincinnati, OH.
Arthur. ID: Lavne. MA: Hochheiser. HW: Arthur. R. (2014). Spatial and statistical analysis of
hydraulic fracturing activities in U.S. shale plays and the effectiveness of the FracFocus
chemical disclosure system. In 2014 SPE Hydraulic Fracturing Technology Conference, 4-6
February, 2014, The Woodlands, TX: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/168640-MS
Astle. Ml: Weast. RC. (1984). Handbook of Chemistry and Physics (64th (1983-1984) ed.). Boca
Raton, FL: CRC Press.
Astrella. L: Wiemers. R. (1996). Technology: Closed loop drilling systems can eliminate reserve pit
costs. Oil and Gas Journal 94.
ATSDR (Agency for Toxic Substances and Disease Registry). (2011). Total petroleum hydrocarbons
(TPH). http://www.atsdr.cdc.gov/substances/toxsubstance.asp?toxid=75
ATSDR (Agency for Toxic Substances and Disease Registry). (2016). Minimal risk levels (MRLs).
March 2016. Atlanta, GA: Agency for Toxic Substances and Disease Registry (ATSDR).
Retrieved from http://www.atsdr.cdc.gov/mrls/index.asp
Avila. 1. (1976). Devonian shale as a source of gas. In Natural Gas from Unconventional Geologic
Sources (pp. 73-85). (FE22711). Washington, D.C.: U.S. Energy Research and Development
Administration.
Avala. S: Barber. T: Dessinges. MN: Frev. M: Horkowitz. I: Leugemors. E: Pessin. 1: Way. CH: Badrv. R:
Kholv. IE: Gait. A: Hiellesnet. M: Sock. D. (2006). Improving oilfield service efficiency.
Houston, TX: Schlumberger Limited.
http://www.slb.eom/~/media/Files/resources/oilfield review/ors06/aut06/improving oi
lfield service efficiencv.pdf
Avbar. U: Yu. W. ei: Eshkalak. MO: Sepehrnoori. K: Patzek. T. ad. (2015). Evaluation of production
losses from unconventional shale reservoirs. Journal of Natural Gas Science & Engineering
23: 509-516. http://dx.doi.org/l0.1016/i.ingse.2015.02.030
Bacher. D. (2013). Oil company fined $60,000 for illegally discharging fracking fluid.
https://www.indybay.org/newsitems/2013/ll/17/18746493.php7show comments=l
(accessed March 6, 2015).
11-6
-------
Chapter 11 - References
Bachu. S: Bennion. DB. (2009). Experimental assessment of brine and/or C02 leakage through well
cements at reservoir conditions. Int J Greenhouse Gas Control 3: 494-501.
http://dx.doi.Org/10.1016/i.iiggc.2008.ll.002
Bair. ES: Digel. RK. (1990). Subsurface transport of inorganic and organic solutes from experimental
road spreading of oilfield brine. Ground Water Monitoring and Remediation 10: 94-105.
Bair. ES: Freeman. DC: Senko. TM. (2010). Subsurface gas invasion Bainbridge Township, Geauga
County, Ohio. (Expert Panel Technical Report). Columbus, OH: Ohio Department of Natural
Resources.
https://oilandgas.ohiodnr.gov/portals/oilgas/pdf/bainbridge/DMRM%200%20Title%20P
age.%20Preface.%20Acknowledgements.pdf
Baker Hughes. (2014). Baker Hughes green chemicals program overview.
Baker Hughes. (2015). Well count. U.S. onshore well count [Database], Houston, TX: Baker Hughes,
Inc. http://phx.corporate-ir.net/phoenix.zhtml?c=79687&p=irol-wellcountus
Baker Hughes. f20161. North America rig count. http://phx.corporate-
ir.net/phoenix.zhtml?c=79687&p=irol-reportsother
Baldassare. F. (2011). The origin of some natural gases in Permian through Devonian Age systems
in the Appalachian Basin and the relationship to incidents of stray gas migration. Presented
at Technical Workshop for Hydraulic Fracturing Study: Chemical and Analytical Methods,
February 24-25, 2011, Arlington, VA. http://www2.epa.gov/hfstudy/origin-some-natural-
gases-permian-through-devonian-age-systems-appalachian-basin-and
Baldassare. FT: McCaffrey. MA: Harper. TA. (2014). A geochemical context for stray gas
investigations in the northern Appalachian Basin: Implications of analyses of natural gases
from Neogene-through Devonian-age strata. AAPG Bulletin 98: 341-372.
http://dx.doi. org/10.1306/06111312178
Barati. R: Liang. IT. (2014). A review of fracturing fluid systems used for hydraulic fracturing of oil
and gas wells. J Appl Polymer Sci Online pub. http: //dx.doi.org/10.1002 /app.40735
Barbot. E: Vidic. NS: Gregory. KB: Vidic. RD. (2013). Spatial and temporal correlation of water
quality parameters of produced waters from Devonian-age shale following hydraulic
fracturing. Environ Sci Technol 47: 2562-2569.
Barker. IF: Fritz. P. (1981). Carbon isotope fractionation during microbial methane oxidation.
Nature 293: 289-291. http://dx.doi.org/10.1038/293289a0
Barnes. T. (2010). 2 drillers fined for Pennsylvania gas well blowout, http://www.post-
gazette.com/news/environment/2 010/07/14/2-drillers-fined-for-Pennsylvania-gas-well-
blowout/stories/201007140241 (accessed March 3, 2015).
Barrett. ME. (2010). Evaluation of sand filter performance. (CRWR Online Report 10-7). Austin, TX:
Center for Research in Water Resources, University of Texas at Austin.
http://www.crwr.utexas.edu/reports/pdf/2010/rptl0-07.pdf
Barth-Naftilan. E: Alovsius. N: Saiers. IE. (2015). Spatial and temporal trends in freshwater
appropriation for natural gas development in Pennsylvania's Marcellus Shale play. Geophys
Res Lett 42: 6348-6356. http://dx.doi.org/10.1002/2015GL065240
Bartos. TT: Hallberg. LL. (2011). Generalized potentiometric surface, estimated depth to water, and
estimated saturated thickness of the high plains aquifer system, March-June 2009, Laramie
County, Wyoming, http://pubs.usgs.gov/sim/3180/
Beckwith. R. (2011). Proppants: Where in the world. J PetTech 63: 36-41.
11-7
-------
Chapter 11 - References
Bene. PG: Harden. B: Griffin. SW: Nicot. TP. (2007). Northern Trinity/Woodbine aquifer
groundwater availability model: Assessment of groundwater use in the Northern Trinity
aquifer due to urban growth and Barnett Shale development
(TWDB Contract Number: 0604830613). Austin, TX: R. W. Harden & Associates, Inc.
https://www.twdb.texas.gov/groundwater/models/gam/trnt n/TRNT N Barnett Shale Re
portpdf
Benko. KL: Drewes. IE. (2008). Produced water in the western United States: Geographical
distribution, occurrence, and composition. Environ Eng Sci 25: 239-246.
Benotti. Ml: Stanford. BP: Snyder. SA. (2010). Impact of drought on wastewater contaminants in an
urban water supply J Environ Qual 39: 1196-1200.
Bertoncello. A: Wallace. I: Honarpour. MM: Kabir. C: Blyton. CA. (2014). Imbibition and water
blockage in unconventional reservoirs: Well management implications during flowback and
early production. SPE Journal 17.
Betanzo. EA: Hagen. ER: Wilson. IT: Reckhow. KH: Hayes. L: Argue. DM.: Cangelosi. AA. (2016).
Water data to answer urgent water policy questions: Monitoring design, available data, and
filling data gaps for determining whether shale gas development activities contaminate
surface water or groundwater in the Susquehanna River Basin. Washington, D.C.: Northeast-
Midwest Institute, U.S. Geological Survey, http://www.nemw.org/wp-
content/uploads/2016/02/Susquehanna-River-Basin- Shale-Gas-Development- Studv-
Reportpdf
Bibbv. KT: Brantley. SL: Reible. DP: Linden. KG: Mouser. PI: Gregory. KB: Ellis. BR: Vidic. RD. (2013).
Suggested reporting parameters for investigations of wastewater from unconventional
shale gas extraction. Environ Sci Technol 47: 13220-13221.
http://dx.doi.org/10.1021/es404960z
Birdsell. DT: Rajaram. H: Dempsev. D: Viswanathan. HS. (2015a). Hydraulic fracturing fluid
migration in the subsurface: A review and expanded modeling results. Water Resour Res 51:
7159-7188. http://dx.doi.Org/10.1002/2015WR017810
Birdsell. DT: Raiaram. H: Lackey. G. (2015b). Imbibition of hydraulic fracturing fluids into partially
saturated shale. Water Resour Res 51: 6787-6796.
http://dx.doi.Org/10.1002/2015WR017621
Bishop. RE. (2013). Historical analysis of oil and gas well plugging in New York: Is the regulatory
system working? New Solutions: A Journal of Environmental and Occupational Health Policy
23: 103-116. http: //dx.doi.org/10.2190/NS.23.1 ,g
BT Services Company. (2009). BJ fracturing manual 2.0. service company confidential business
information document (Revision No. 1 ed.). Houston, TX.
Blauch. ME. (2010). Developing effective and environmentally suitable fracturing fluids using
hydraulic fracturing flowback waters. Presented at 2010 SPE unconventional gas
conference, 23-25 February, 2010, Pittsburgh, PA: Society of Petroleum Engineers.
http: //dx.doi.org/10.2118/131784-MS
Blauch. ME: Myers. RR: Moore. TR: Lipinski. BA. (2009). Marcellus shale post-frac flowback waters -
Where is all the salt coming from and what are the implications? In proceedings of the SPE
Eastern Regional Meeting, 23-25 September, 2009, Charleston, WV: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/125740-MS
BLM (Bureau of Land Management). (2013). Hydraulic fracturing white paper, Appendix E. Casper,
WY: Bureau of Land Management, Wyoming State Office.
11-8
-------
Chapter 11 - References
https://www.blm.gOv/style/medialib/blm/wy/information/NEPA/og/2015/02feb/verl.P
ar.62094.File.dat/Vl WRBBD-AppE.pdf
BLM (Bureau of Land Management). (2015). Abandoned land mines inventory [Database],
Washington, D.C.
http://www.blm.gov/wo/st/en/prog/more/Abandoned Mine Lands/abandoned mine sit
e.html
Blondes. MS: Gans. KD: Thordsen. 11: Reidv. ME: Thomas. B: Engle. MA: Kharaka. YK: Rowan. EL.
(2014). Data: U.S. Geological Survey National Produced Waters Geochemical Database v2.0
(Provisional) [Database]: U.S. Geological Survey.
http://energy.usgs.gov/EnvironmentalAspects/EnvironmentalAspectsofEnergyProduction
andUse/ProducedWaters.aspx#3822349-data
Boghici. R. (2009). Water quality in the Carrizo-Wilcox Aquifer, 1990-2006. (Report 372). Austin,
TX: Texas Water Development Board.
http://www.twdb.texas.gov/publications/reports/numbered reports/doc/R372Carrizo-
Wilcox.pdf
Boothrovd. IM: Almond. S: Passim. SM: Worrall. F: Davies. RT. (2016). Fugitive emissions of methane
from abandoned, decommissioned oil and gas wells. Sci Total Environ 547: 461-469.
http ://dx. doi. or g/10.1016/i. scitotenv. 2015.12.096
Boschee. P. (2012). Handling produced water from hydraulic fracturing. Oil and Gas Facilities 1: 23-
26.
Boschee. P. (2014). Produced and flowback water recycling and reuse: Economics, limitations, and
technology. Oil and Gas Facilities 3: 16-22.
Bosworth. S: El-Saved. HS: Ismail. G: Ohmer. H: Stracke. M: West. C: Retnanto. A. (1998). Key issues
in multilateral technology. Oilfield Rev 10: 14-28.
Bouchard. D: Cornaton. F: Hohener. P: Hunkeler. D. (2011). Analytical modelling of stable isotope
fractionation of volatile organic compounds in the unsaturated zone. J Contam Hydrol 119:
44-54. http://dx.doi.Org/10.1016/i.iconhyd.2010.09.006
Bowen. ZH: Oelsner. GP: Cade. BS: Gallegos. TT: Farag. AM: Mott. DN: Potter. CI: Cinotto. PI: Clark.
ML: Kappel. WM: Kresse. TM: Melcher. CP: Paschke. SS: Susong. DP: Varela. BA. (2015).
Assessment of surface water chloride and conductivity trends in areas of unconventional oil
and gas development-Why existing national data sets cannot tell us what we would like to
know. Water Resour Res 51: 704-715. http://dx.doi.org/10.1002/2014WR016382
Boyd. D: Al-Kubti. S: Khedr. OH: Khan. N: Al-Navadi. K: Degouv. D: Elkadi. A: Kindi. ZA. (2006).
Reliability of cement bond log interpretations compared to physical communication tests
between formations. Presented at Abu Dhabi International Petroleum Exhibition and
Conference, November 5-8, 2006, Abu Dhabi, UAE.
http://www.onepetro.org/mslib/app/Preview.do?paperNumber=SPE-101420-
MS&societyCode=SPE
Bover. CM: Glenn. SA: Clavpool. BR: Weida. SD: Adams. ID: Huck. PR: Stidham. IE. (2005).
Application of viscoelastic fracturing fluids in Appalachian Basin reservoirs (SPE 98068
ed.). Presented atthe SPE Eastern Regional Meeting, 14-16 September, 2005, Morgantown,
WV: Society of Petroleum Engineers. http: //dx.doi.org/10.2118 /98068-MS
Bovsen. PB: Bovsen. TA: Bovsen. IE. (2002). Creative Strategies for Produced Water Pisposal in the
Rocky Mountain Region. Presented at 9th Annual International Petroleum Environmental
11-9
-------
Chapter 11 - References
Conference, October 2002, Albuquerque, NM.
http://ipec.utulsa.edu/Conf2002/boysen 89.pdf
Bovsen. IE: Hariu. TA: Shaw. B: Fosdick. M: Grisanti. A: Sorensen. TA. (1999). The current status of
commercial deployment of the freeze thaw evaporation treatment of produced water.
Presented at the SPE/EPA Exploration and Production Environmental Conferece, 1-3 March,
1999, Austin, TX: Society of Petroleum Engineers. http: //dx.doi.org/10.2118/52700-MS
Brannon. HP: Daulton. DT: Hudson. HG: Tordan. AK. (2011). Progression toward implementation of
environmentally responsible fracturing processes (SPE-147534-MS ed.). Presented at SPE
Annual Technical Conference and Exhibition, 30 0ctober-2 November, 2011, Denver, CO:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/147534-MS
Brannon. HP: Daulton. DT: Post. MA: Hudson. HG: Tordan. AK. (2012). The quest to exclusive use of
environmentally responsible fracturing products and systems. Presented at the SPE
Hydraulic Fracturing Technology Conference, 6-8 February, 2012, The Woodlands, TX:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/152068-MS
Brannon. HP: Kendrick. DE: Luckev. E: Stipetich. A. (2009). Multistage fracturing of horizontal shale
gas wells using >90% foam provides improved production. In 2009 SPE Eastern Regional
Meeting: Limitless Potential/Formidable Challenges (SPE 124767 ed.), 23-25 September,
2009, Charleston, WV: Society of Petroleum Engineers, http: //dx.doi.org/10.2118/124767-
MS
Brannon. HP: Pearson. CM. (2007). Proppants and fracture conductivity. In Modern Fracturing:
Enhancing Natural Gas Production (1st ed.). Houston, TX: Energy Tribune Publishing, Inc.
Brantley. SL: Yoxtheimer. P: Arjmand. S: Grieve. P: Vidic. R: Pollak. I: Llewellyn. GT: Abad. I: Simon.
C. (2014). Water resource impacts during unconventional shale gas development: The
Pennsylvania experience. Int J Coal Geol 126: 140-156.
http://dx.doi.Org/10.1016/i.coal.2013.12.017
Breit. GN. (2002). USGS Produced waters database. Available online at
http://energy.usgs.gov/EnvironmentalAspects/EnvironmentalAspectsofEnergyProduction
andUse/ProducedWaters.aspx#3822349-data
Brinck. EL: Frost. CP. (2007). Petecting infiltration and impacts of introduced water using
strontium isotopes. Ground Water 45: 554-568. http://dx.doi.org/10.ll 11 /i. 1745-
6584.2007.00345.x
Brown. PG: Taffe. PR. (2001). Effects of nonionic surfactants on bacterial transport through porous
media. Environ Sci Technol 35: 3877-3883. http://dx.doi.org/10.1021/esQ10577w
Brown. HP: Grijalva. YE: Ravmer. LL. (1970). New developments in sonic wave train display and
analysis in cased holes. (SPWLA-1970-F). https://www.onepetro.org/conference-
paper/SPWLA-1970-F
Browne. PI: Lukocs. BP. (1999). The development of a rapid hydration on-the-fly crosslinked water
fracturing fluid. Presented at CSPG and Petroleum Society Joint Convention: Pigging Peeper,
Finding a Better Bottom Line, June 14 18,1999, Alberta, Canada.
http://www.tricanwellservice.com/sites/default/files/pdf/1356 SPE TechnicalPaper Frac
turing PetroleumSocietv99-59-FINAL.pdf
Brownlow. IW: Tames. SC: Yelderman. IC. (2016). Influence of hydraulic fracturing on overlying
aquifers in the presence of leaky abandoned wells. Ground Water.
http: / /dx. doi. or g/10.1111 /gwat. 12431
11-10
-------
Chapter 11 - References
Brufatto. C: Cochran. T: Conn. L: El-Zeghaty. SZA. A: Fraboulet. B: Griffin. T: Tames. S: Munk. T: Tustus.
F: Levine. TR: Montgomery. C: Murphy. D: Pfeiffer. T: Pornpoch. T: Rishmani. L. (2003). From
mud to cement - Building gas wells. Oilfield Rev 15: 62-76.
Burlingame. GA: Waer. MA. (2002). Committee report: Options for a taste and odor standard. J Am
Water Works Assoc 94: 80-87.
Burton. TG: Rifai. HS: Hildenbrand. ZL: Carlton. DP: Fontenot. BE: Schug. KA. (2016). Elucidating
hydraulic fracturing impacts on groundwater quality using a regional geospatial statistical
modeling approach. Sci Total Environ 545-546: 114-126.
http ://dx. doi. or g/10.1016/i. scitotenv. 2015.12.084
Byrnes. AP. (2011). Role of induced and natural imbibition in frac fluid transport and fate in gas
shales. Presented at Technical Workshops for Hydraulic Fracturing Study: Fate & Transport,
March 28-29, 2011, Arlington, VA.
https://www.epa.gov/sites/production/files/documents/roleofinducedandnaturalimbibiti
oninfracfluid.pdf
CA Water Board (California Water Board). (2016). Produced water pond status report July 30,
2016. Fresno, CA.
http://www.waterboards.ca.gov/water issues/programs/groundwater/sb4/docs/produce
d pond rpt 2016iul.pdf
CalEPA (California Environmental Protection Agency). (2016). Oil fields-food safety, responses to
13267 orders for chemicals used during petroleum production.
http://www.waterboards.ca.gov/centralvallev/water issues/oil fields/food safety/index.s
html
California Department of Water Resources. (2016). California data exchange center: Daily reservoir
storage summary. http://cdec.water.ca.gov/cgi-progs/reservoirs/RES
CARWOCB CVR (California Regional Water Quality Control Board Central Valley Region). (2015).
Water quality control plan for the Tulare Lake basin second edition. Sacramento, CA:
California Regional Water Quality Control Board.
http://www.waterboards.ca.gov/centralvallev/water issues/basin plans/tlbp.pdf
California State Water Resources Control Board. (2015). Model criteria for groundwater monitoring
in areas of oil and gas well stimulation. Sacramento, CA: California Water Boards.
http://www.waterboards.ca.gov/water issues/programs/groundwater/sb4/docs/model c
riteria final 070715.pdf
California State Water Resources Control Board. (2016). Water quality in areas of oil and gas
production Produced water ponds.
http://www.waterboards.ca.gov/water issues/programs/groundwater/sb4/oil field prod
uced/produced water ponds/index.shtml (accessed September 26, 2016).
Calvert. DG: Smith. DK. (1994). Issues and techniques of plugging and abandonment of oil and gas
wells. Presented at SPE Annual Technical Conference and Exhibition, September 25-28,
1994, New Orleans, LA. http://dx.doi.org/10.2118/28349-MS
Camobreco. VI: Richards. BK: Steenhuis. TS: Peverlv. TH: McBride. MB. (1996). Movement of heavy
metals through undisturbed and homogenized soil columns. Soil Sci 161: 740-750.
http: //dx.doi.org/10.1097/00010694-199611000-00003
Caniglia. 1. (2014). Youngstown contractor sentenced to 28 months for dumping fracking waste,
(accessed March 12, 2015).
11-11
-------
Chapter 11 - References
Carter. KE: Hammack. RW: Hakala. TA. (2013). Hydraulic fracturing and organic compounds - Uses,
disposal and challenges. PresEnted at the SPE Eastern Regional Meeting, 20-22 August,
2013, Pittsburgh, PA: Society of Petroleum Engineers, http: //dx.doi.org/10.2118/165692-
MS
Castle. SL: Thomas. BF: Reager. IT: Rodell. M: Swenson. SC: Famiglietti. IS. (2014). Groundwater
depletion during drought threatens future water security of the Colorado River Basin.
Geophys Res Lett 41: 5904-5911. http://dx.doi.org/10.1002/2014GL061055
Catlett. RD: Spencer. ID: Lolon. E: Bucior. D. (2013). Evaluation of two horizontal wells in the Eagle
Ford using oil-based chemical tracer technology to optimize stimulation design. In SPE
Hydraulic Fracturing Technology Conference, 4-6 February, 2013, The Woodlands, TX:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/163846-MS
CCST (California Council on Science and Technology). (2014). Advanced well stimulation
technologies in California: An independent review of scientific and technical information.
Sacramento, CA. http://ccstus/publications/2014/2014wstpdf
CCST (California Council on Science and Technology). (2015a). An independent scientific
assessment of well stimulation in California, Volume 2: Potential environmental impacts of
hydraulic fracturing and acid stimulations. Sacramento, CA.
https://ccstus/publications/2015/2015SB4-v2.pdf
CCST (California Council on Science and Technology). (2015b). An independent scientific
assessment of well stimulation in California, Volume 1: Well stimulation technologies and
their past, present, and potential future use in California. Sacramento, CA.
http: / /www. ccstus /publications /2 015 /2 015 SB4-vl .pdf
Ceres (Coalition for Environmentally Responsible Economies). (2014). Hydraulic fracturing & water
stress: Water demand by the numbers. Boston, MA.
Chambers. DB: Kozar. MP: Messinger. T: Mulder. ML: Pelak. AT: White. IS. (2014). Water quality of
groundwater and stream base flow in the Marcellus Shale gas field of the Monongahela
River Basin, West Virginia, 2011-2012. Reston, VA: U.S. Geological Survey.
http://dx.doi.org/10.3133/sir20145233
Chapman. EC: Capo. RC: Stewart. BW: Kirbv. CS: Hammack. RW: Schroeder. KT: Edenborn. HM.
(2012). Geochemical and strontium isotope characterization of produced waters from
Marcellus Shale natural gas extraction. Environ Sci Technol 46: 3545-3553.
Chaudhuri. S: Ale. S. (2013). Characterization of groundwater resources in the Trinity and
Woodbine aquifers in Texas. Sci Total Environ 452: 333-348.
http://dx.doi.Org/10.1016/i.scitotenv.2013.02.081
Chaudhuri. S: Ale. S. (2014a). Long term (1960-2010) trends in groundwater contamination and
salinization in the Ogallala aquifer in Texas. J Hydrol 513: 376-390.
http://dx.doi.Org/10.1016/i.ihydrol.2014.03.033
Chaudhuri. S: Ale. S. (2014b). Temporal evolution of depth-stratified groundwater salinity in
municipal wells in the major aquifers in Texas, USA. Sci Total Environ 472: 370-380.
http: / /dx. doi. or g/10.1016/i. scitotenv. 2013.10.120
Chebotarev. II. (1955). Metamorphism of natural waters in the crust of weathering. Geochim Cosmo
Act 8: 22-48. http://dx.doi.org/10.1016/0016-7037r55190015-6
ChemicalBook (ChemicalBook Inc.). (2010). Sorbitan trioleate. Available online at
http://www.chemicalbook.com/chemicalproductproperty en cb4677178.htm (accessed
April 6, 2015).
11-12
-------
Chapter 11 - References
Chester Engineers. (2012). The Pittsburgh Water and Sewer Authority 40 year plan. (PWSA Project
No. R-Dl.10015-11). Pittsburgh, PA: The Pittsburgh Water and Sewer Authority.
http://apps.pittsburghpa.gov/pwsa/PWSA 40-vear Plan.pdf
Christian. KM: Lautz. LK: Hoke. GD: Siegel. PI: Lu. Z: Kessler. T. (2016). Methane occurrence is
associated with sodium-rich valley waters in domestic wells overlying the Marcellus Shale
in New York State. Water Resour Res 52: 206-226.
http://dx.d0i.0rg/l 0.1002/2015WR017805
Cipolla. C: Weng. X: Mack. M: Ganguly. U: Gu. H: Kresse. 0: Cohen. C. (2011). Integrating
microseismic mapping and complex fracture modeling to characterize hydraulic fracture
complexity. Presented at SPE Hydraulic Fracturing Technology Conference, January 24-26,
2011, The Woodlands, TX.
http://www.onepetro.org/mslib/app/Preview.do?paperNumber=SPE-140185-
MS&societyCode=SPE
Clark. CE: Horner. RM: Harto. CB. (2013). Life cycle water consumption for shale gas and
conventional natural gas. Environ Sci Technol 47: 11829-11836.
http://dx.doi.org/10.1021/es4013855
Clark. CE: Veil. TA. (2009). Produced water volumes and management practices in the United States.
(ANL/EVS/R-09/1). Argonne, IL: Argonne National Laboratory.
http://www.ipd.anl.gOv/anlpubs/2009/07/64622.pdf
Cluff. M: Hartsock. A: Macrae. 1: Carter. K: Mouser. PI. (2014). Temporal changes in microbial
ecology and geochemistry in produced water from hydraulically fractured Marcellus Shale
gas wells. Environ Sci Technol 48: 6508-6517. http://dx.doi.org/10.1021/es501173p
COGCC (Colorado Oil and Gas Conservation Commission). (2004). Colorado Oil and Gas
Conservation Commission Order No. IV-276.
https://cogcc.state.co.us/orders/orders/lv/276.html
COGCC (Colorado Oil and Gas Conservation Commission). (2012). COGIS (Colorado Oil and Gas
Information System ) [Database], Denver, CO. http://cogcc.state.co.us/#/home
COGCC (Colorado Oil and Gas Conservation Commission). (2013). COGCC 2013 flood response.
Denver, CO.
COGCC (Colorado Oil and Gas Conservation Commission). (2014). Risk-based inspections:
Strategies to address environmental risk associated with oil and gas operations.
(COGCC-2014-PROJECT #7948). Denver, CO.
COGCC (Colorado Oil and Gas Conservation Commission). (2015). COGIS (Colorado Oil and Gas
Information System ) - All production reports to date. Denver, CO. http: //cogcc.state.co.us/
COGCC (Colorado Oil and Gas Conservation Commission). (2016). COGIS - Facility inquiry
[Database], Denver, CO. http://cogcc.state.co.us/cogis/FacilitySearch.asp
Colaneri. K. (2014). Range Resources to pay $4M for violations at western Pa. impoundments.
https://stateimpactnpr.org/pennsylvania/2014/09/18/range-resources-to-pay-4m-for-
violations-at-western-pa-impoundments/ (accessed September 9, 2016).
Colborn. T: Kwiatkowski. C: Schultz. K: Bachran. M. (2011). Natural gas operations from a public
health perspective. HumEcol Risk Assess 17: 1039-1056.
http://dx.doi.Org/10.1080/10807039.2011.605662
11-13
-------
Chapter 11 - References
Coleman. DP: Liu. CL: Riley. KM. (1988). Microbial methane in the shallow Paleozoic sediments and
glacial deposits of Illinois, U.S.A. Chem Geol 71: 23-40. http: //dx.doi.org/10.1016/0009-
2541 (~88"I90103-9
Colorado Division of Water Resources et al. (Colorado Division of Water Resources; Colorado Water
Conservation Board; Colorado Oil and Gas Conservation Commission). (2014). Water
sources and demand for the hydraulic fracturing of oil and gas wells in Colorado from 2010
through 2015 [Fact Sheet], http://cewc.colostate.edu/2012/02 /water-sources-and-
demand-for-the-hydraulic-fracturing-of-oil-and-gas-wells-in-colorado-from-2010-through-
2015/
Considine. T: Watson. R: Considine. N: and Martin. 1. (2012). Environmental impacts during
Marcellus shale gas drilling: Causes, impacts, and remedies. (Report 2012-1). Buffalo, NY:
Shale Resources and Society Institute.
http://cce.cornell.edu/EnergyClimateChange/NaturalGasDev/Documents/UBSRSI-
Environmental%20Impact%20Report%2 02012.pdf
Coolev. H: Gleick. PH: Wolff. G. (2006). Desalination, with a grain of salt: A California perspective.
Oakland, CA: Pacific Institute for Studies in Development, Environment, and Security.
http://www.pacinstorg/wp-content/uploads/2013/02/desalination report3.pdf
Corseuil. HX: Kaipper. BI: Fernandes. M. (2004). Cosolvency effect in subsurface systems
contaminated with petroleum hydrocarbons and ethanol. Water Res 38: 1449-1456.
http://dx.doi.Org/10.1016/j.watres.2003.12.015
Corseuil. HX: Monier. AL: Fernandes. M: Schneider. MR: Nunes. CC: do Rosario. M: Alvarez. PI.
(2011). BTEX plume dynamics following an ethanol blend release: Geochemical footprint
and thermodynamic constraints on natural attenuation. Environ Sci Technol 45: 3422-3429.
http://dx.doi.org/10.1021/esl04055q
Coulter. GR: Gross. BC: Benton. EG: Thomson. CL. (2006). Barnett shale hybrid fracs - One operator's
design, application, and results. Presented at SPE Annual Technical Conference and
Exhibition, 24-27 September, 2006, San Antonio, TX: Society of Petroleum Engineers.
http://dx.doi. org/10.2118/102063-MS
Council of Canadian Academies. (2014). Environmental impacts of shale gas extraction in Canada.
Ottawa, Ontario.
http://www.scienceadvice.ca/uploads/eng/assessments%20and%20publications%20and
%20news%20releases/Shale%20gas/ShaleGas fullreportEN.pdf
Countess. S: Boardman. G: Hammack. R: Hakala. A: Sharma. S: Parks. 1. (2014). Evaluating
leachability of residual solids from hydraulic fracturing in the Marcellus shale. In Shale
Energy Engineering 2014: Technical Ehallenges, Environmental Issues, and Public Policy.
Reston, VA: American Society of Civil Engineers.
http://dx.doi.org/10.1061/9780784413654.012
Craig. MS: Wendte. SS: Buchwalter. TL. (2012). Barnett shale horizontal restimulations: A case study
of 13 wells. In SPE America's Unconventional Resources Conference, June 5-7, 2012,
Pittsburgh, PA.
Cramer. DP. (2008). Stimulating unconventional reservoirs: Lessons learned, successful practices,
areas for improvement. In SPE Unconventional Reservoirs Conference, February 10-12,
2008, Keystone, CO.
Crescent (Crescent Consulting, LLC). (2011). East Mamm creek project drilling and cementing
study. Oklahoma City, OK.
11-14
-------
Chapter 11 - References
Crook. R. (2008). Cementing: Cementing horizontal wells. Halliburton.
Curtice. RT: Salas. WDT: Paterniti. ML. (2009). To gel or notto gel? In 2009 SPE Annual Technical
Conference and Exhibition, 4-7 October, 2009, New Orleans, LA: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/124125-MS
Cusick. M. (2013). EPA fines western PA treatment plants for Marcellus wastewater violations.
http://stateimpact.npr.org/pennsylvania/2013/Q5/24/epa-fines-western-pa-treatment-
plants-for-marcellus-wastewater-violations/ (accessed March 6, 2015).
D'Amico. D: Davis. T. (2015). Monitoring hydraulic fracturing through the use of time-lapse,
multicomponent seismic data, Pouce Coupe Field, Alberta. (SEG-2015-5934718). Tulsa, OK:
Society of Exploration Geophysicists.
Dahi Taleghani. A: Ahmadi. M: Olson. IE. (2013). Secondary fractures and their potential impacts on
hydraulic fractures efficiency. In Bunger; J McLennan; R Jeffrey (Eds.), Effective and
Sustainable Hydraulic Fracturing. Croatia: InTech. http: //dx.doi.org/10.5772 /56360
Dahi Taleghani. A: Olson. IE. (2011). Numerical modeling of multi-stranded hydraulic fracture
propagation: Accounting for the interaction between induced and natural fractures. In 2009
SPE Annual Technical Conference and Exhibition. Richardson, TX: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/124884-PA
Dahm. K: Chapman. M. (2014). Produced water treatment primer: Case studies of treatment
applications. (S&T Research Project #1617). Denver CO: U.S. Department of the Interior.
http://www.usbr.gov/research/proiects/download productcfm?id=1214.
Dahm. KG: Guerra. KL: Xu. P: Drewes. IE. (2011). Composite geochemical database for coalbed
methane produced water quality in the Rocky Mountain region. Environ Sci Technol 45:
7655-7663. http://dx.doi.org/10.1021/es201021n
Pake. LP. (1978). Fundamentals of reservoir engineering. Boston, MA: Elsevier.
http://www.ing.unp.edu.ar/asignaturas/reservorios/Fundamentals%20of%20Reservoir%
20Engineering%20%28L.P.%20Dake%29.pdf
Dane shy. AA. (2009). Factors controlling the vertical growth of hydraulic fractures. (SPE-118789-
MSl. http://dx.d0i.0rg/l 0.2118/118789-MS
Darrah. TH: Vengosh. A: Tackson. RB: Warner. NR: Poreda. RT. (2014). Noble gases identify the
mechanisms of fugitive gas contamination in drinking-water wells overlying the Marcellus
and Barnett Shales. Proc Natl Acad Sci USA 111: 14076-14081.
http://dx.doi.org/10.1073/pnas.13221Q7111
Daulton. D: Post. M: McMahon. 1: Kuc. B: Ake. C: Hughes. B: Hill. D. (2012). Global chemical
evaluation process review to qualify regulatory and environmental characteristics for
oilfield chemical products. Presented at SPE Annual Technical Conference and Exhibition,
October 8-10, 2012, San Antonio, TX.
Davies. RT: Almond. S: Ward. RS: Tackson. RB: Adams. C: Worrall. F: Herringshaw. LG: Gluvas. TG:
Whitehead. MA. (2014). Oil and gas wells and their integrity: Implications for shale and
unconventional resource exploitation. Marine and Petroleum Geology 56: 239-254.
http://dx.doi.Org/10.1016/i.marpetgeo.2014.03.001
Davies. RT: Foulger. GR: Mathias. S: Moss. 1: Hustoft. S: Newport. L. (2013). Reply: Davies etal.
(2012), "Hydraulic fractures: How far can they go?" Marine and Petroleum Geology 43: 519-
521. http://dx.doi.0rg/lO.lOl6/i.marpetgeo.2Oi3.O2.OOl
11-15
-------
Chapter 11 - References
Davies. RT: Mathias. SA: Moss. T: Hustoft. S: Newport. L. (2012). Hydraulic fractures: How far can
they go? Marine and Petroleum Geology 37: 1-6.
http://dx.doi.Org/10.1016/i.marpetgeo.2012.04.001
Davis. TP: Struchtemever. CG: Elshahed. MS. (2012). Bacterial communities associated with
production facilities of two newly drilled thermogenic natural gas wells in the Barnett Shale
(Texas, USA). Microb Ecol 64: 942-954. http://dx.doi.org/l 0.1007/s00248-012-0073-3
Davalu. A: Konschnik. K. (2016). FracFocus Chemical Disclosure Registry 1.0 and 2.0: Data
conversion and cleaning methods paper. Cambridge, MA: Harvard Law School.
De Andrade. 1: Sangesland. S: Todorovic. 1: Vralstad. T. (2015). Cement sheath integrity during
thermal cycling: A novel approach for experimental tests of cement systems. SPE Bergen
One Day Seminar, April 22, 2015, Bergen, Norway.
de Pater. HI. (2015). Hydraulic fracture containment: New insights into mapped geometry. In 2015
SPE Hydraulic Fracturing Technology Conference, 3-5 February, The Woodlands, TX:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/173359-MS
Deans. C. (2007). The single-well chemical tracer test: A method for measuring reservoir fluid
saturation in situ. In ED Holstein (Ed.), Petroleum Engineering Handbook: Reservoir
Engineering and Petrophysics (pp. 615). Richardson, TX: Society of Petroleum Engineers.
Dehghanpour. H: Lan. 0: Saeed. Y: Fei. H: Oi. Z. (2013). Spontaneous imbibition of brine and oil in
gas shales: Effect of water adsorption and resulting microfractures. Energy Fuels 27: 3039-
3049. http://dx.doi.org/10.1021/ef4002814
Dehghanpour. H: Zubair. HA: Chhabra. A: Ullah. A. (2012). Liquid intake of organic shales. Energy
Fuels 26: 5750-5758. http://dx.d0i.0rg/l 0.1021 /ef3009794
DeSimone. LA: McMahon. PB: Rosen. MR. (2014). The quality of our nations waters: Water quality
in principal aquifers of the United States, 1991-2010. (Circular 1360). Reston, VA: U.S.
Geological Survey, http://dx.doi.org/10.3133/cirl360
Diehl. TH: Harris. MA. (2014). Withdrawal and consumption of water by thermoelectric power
plants in the United States, 2010. (Scientific Investigations Report 20145184). Reston, VA:
U.S. Geological Survey, http://dx.doi.org/10.3133/sir20145184
Digiulio. DC: Tackson. RB. (2016). Impact to underground sources of drinking water and domestic
wells from production well stimulation and completion practices in the Pavillion, Wyoming,
Field. Environ Sci Technol 50: 4524-4536. http://dx.doi.org/10.1021/acs.est.5b04970
Dilmore. RM: Sams. II: Glosser. D: Carter. KM: Bain. DT. (2015). Spatial and temporal characteristics
of historical oil and gas wells in Pennsylvania: Implications for new shale gas resources.
Environ Sci Technol 49: 12015-12023. http://dx.doi.org/10.1021/acs.est.5b00820
DOE (U.S. Department of Energy). (2002). Handbook on best management practices and mitigation
strategies for coalbed methane in the Montana portion of the Powder River Basin. Tulsa,
OK: U.S. Department of Energy, National Energy Technology Laboratory.
http://bogc.dnrc.mt.gov/PDF/BMPHandbookFinal.pdf
DOE (U.S. Department of Energy). (2003). Handbook on coalbed methane produced water:
Management and beneficial use alternatives. Tulsa, OK: ALL Consulting, http: //www.all-
llc.com/publicdownloads/CBM BU Screen.pdf
DOE (U.S. Department of Energy). (2006). A guide to practical management of produced water from
onshore oil and gas operations in the United States. Washington, D.C.: U.S. Department of
Energy, National Petroleum Technology Office.
11-16
-------
Chapter 11 - References
http://fracfocus.0rg/sites/default/files/publications/a guide to practical management of
produced water from onshore oil and gas operations in the united states.pdf
DOE (U.S. Department of Energy). (2011a). Acomparative study ofthe Mississippian Barnett Shale,
Fort Worth Basin, and Devonian Marcellus Shale, Appalachian Basin. (DOE/NETL-
2011/1478). http://www.netl.doe.gov/technologies/oil-gas/publications/brochures/DQE-
NETL-2011-1478%20Marcellus-Barnett.pdf
DOE (U.S. Department of Energy). (2011b). Shale gas: Applying technology to solve Americas
energy challenges, http: //www.netl.doe.gov/file%201ibrary/research/oil-
gas/Shale Gas March 2011.pdf
DOE (U.S. Department of Energy). (2014a). Secretary of Energy Advisory Board: Task force report
on FracFocus 2.0. Washington, D.C.
http://energy.gOv/sites/prod/files/2014/04/fl4/20140328 SEAB TF FracFocus2 Report
Final.pdf
DOE (U.S. Department of Energy). (2014b). Water management strategies for improved coalbed
methane production in the Black Warrior Basin, http: //www.netl.doe.gov/research/oil-
and- gas /proj ect-summaries /natural- gas-resources /de- fe 0000888
DOE (U.S. Department of Energy). (2015). Chapter 7: Advancing systems and technologies to
produce cleaner fuels, technology assessments, unconventional oil and gas. In Quadrennial
Technology Review 2015. Washington, D.C.
http://www.energy.gOv/sites/prod/files/2016/03/f30/0TR2015-7G-Unconventional-Oil-
and-Gas.pdf
Domenico. PA: Schwartz. FW. (1997). Physical and Chemical Hydrology (2nd ed.). Hoboken, NJ:
Wiley.
Drewes. 1: Cath. T: Debroux. 1: Veil. 1. (2009). An integrated framework for treatment and
management of produced water - Technical assessment of produced water treatment
technologies (lsted.). (RPSEA Project 07122-12). Golden, CO: Colorado School ofMines.
http://aqwatec.mines.edu/research/proiects/Tech Assessment PW Treatment Tech.pdf
Drillinglnfo. Inc.. (2014a). DI Desktop June 2014 download [Database], Austin, TX: Drillinglnfo.
Retrieved from http: //info.drillinginfo.com/di-desktop-downloads/
Drillinglnfo. Inc.. (2014b). Drillinglnfo Inc. DI Desktop raw data feed [Database],
Drohan. PI: Brittingham. M. (2012). Topographic and soil constraints to shale-gas development in
the northcentral Appalachians. Soil Sci Soc Am J 76: 1696-1706.
http: / /dx. doi. or g/10.213 6 /sssai 2 012.0 0 87
Drollette. BP: Hoelzer. K: Warner. NR: Darrah. TH: Karatum. 0: O'Connor. MP: Nelson. RK:
Fernandez. LA: Reddv. CM: Vengosh. A: Tackson. RB: Eisner. M: Plata. PL. (2015). Elevated
levels of diesel range organic compounds in groundwater near Marcellus gas operations are
derived from surface activities. Proc Natl Acad Sci USA 112: 13184-13189.
http://dx.doi.org/10.1073/pnas.1511474112
Pugstad. A. (2007). Well to well tracer tests. In EP Holstein (Ed.), Petroleum Engineering
Handbook: Reservoir Engineering and Petrophysics (pp. 651). Richardson, TX: Society of
Petroleum Engineers.
Puraisamv. RT: Beni. AH: Henni. A. (2013). State of the art treatment of produced water. In W
Elshorbagy; RK Chowdhury (Eds.), Water Treatment (pp. 199-222). Rijeka, Croatia: InTech.
http://dx.doi.org/10.5772/53478
11-17
-------
Chapter 11 - References
Dusseault. MB: Gray. MN: Nawrocki. PA. (2000). Why oilwells leak: Cement behavior and long-term
consequences. Presented at SPE International Oil and Gas Conference and Exhibition in
China, November 7-10, 2000, Beijing, China, https://www.onepetro.org/conference-
paper/SPE-64733-MS
Dutta. R: Lee. C-H: Odumabo. S: Ye. P: Walker. SC: Karpvn. ZT: Avala. LF. (2014). Experimental
investigation of fracturing-fluid migration caused by spontaneous imbibition in fractured
low-permeability sands. SPE Reserv Eval Engin 17: 74-81.
Easton. 1. (2014). Optimizing fracking wastewater management. Pollution Engineering, January 13.
Eberhard. M. (2011). Fracture design and stimulation - Monitoring. Presented atTechnical
Workshops for the Hydraulic Fracturing Study: Well Construction & Operations, March 10-
11, 2011, Arlington, VA.
Economides. Ml: Hill. A. d: Ehlig-Economides. C: Zhu. D. (2013). Petroleum production systems (2nd
ed.). Englewood Cliffs, NJ: Prentice Hall.
Economides. Ml: Martin. T. (2007). Modern fracturing: Enhancing natural gas production. Houston,
TX: ET Publishing.
Economides. Ml: Mikhailov. DN: Nikolaevskiv. VN. (2007). On the problem of fluid leakoff during
hydraulic fracturing. Transport in Porous Media 67: 487-499.
http://dx.doi. org/10.1007/sl 1242-006-9038-7
Edwards. DA: Andriot. MP: Amoruso. MA: Tummev. AC: Bevan. CI: Tveit. A: Hayes. LA: Youngren.
SH: Nakles. DV. (1997). Development of fraction specific reference doses (RFDs) and
reference concentrations (RFCs) for total petroleum hydrocarbons (TPH). Amherst, MA:
Amherst Scientific Publishers.
EERC (Energy and Environmental Research Center, University of North Dakota). (2010). Bakken
water opportunities assessment- Phase 1. (2010-EERC-04-03). Grand Forks, ND: Energy
and Environmental Research Center.
http://www.undeerc.org/bakken/pdfs/FracWaterPhaseIreport.pdf
EERC (Energy and Environmental Research Center, University of North Dakota). (2011). Bakken
water opportunities assessment- Phase 2: Evaluation of brackish groundwater treatment
for use in hydraulic fracturing of the Bakken Play, North Dakota. (2011-EERC-12-05). Grand
Forks, ND: Energy and Environmental Research Center.
http://www.undeerc.org/Water/pdf/BakkenWater0ppPhase2.pdf
EERC (Energy and Environmental Research Center, University of North Dakota). (2013).
BakkenSmart: Water [Fact Sheet], Grand Forks, ND: Energy and Environmental Research
Center. http://www.undeerc.org/bakken/pdfs/NDIC-NDPC-Water-Fact-Sheet.pdf
EIA (U.S. Energy Information Administration). (2011a). Review of emerging resources: U.S. shale
gas and shale oil plays. Washington, D.C: U.S. Department of Energy, U.S. Energy Information
Administration, https://www.eia.gov/analysis/studies/usshalegas/
EIA (U.S. Energy Information Administration). (2011b). Shale gas and oil plays, lower 48 States
[Map], https://www.eia.gov/special/shaleplays/
EIA (U.S. Energy Information Administration). (2012). Today in energy: Geology and technology
drive estimates of technically recoverable resources. Washington, D.C.
http://www.eia.gov/todayinenergy/detail.cfm?id=7190
EIA (U.S. Energy Information Administration). (2013). Technically recoverable shale oil and shale
gas resources: An assessment of 137 shale formations in 41 countries outside the United
11-18
-------
Chapter 11 - References
States (pp. 730). Washington, D.C.: U.S. Department of Energy, U.S. Energy Information
Administration, http://www.eia.gov/analysis/studies/worldshalegas/
EIA (U.S. Energy Information Administration). (2014a). Annual energy outlook 2014 with
projections to 2040. (DOE/EIA-0383(2014)). Washington, D.C.
http://www.eia.gOv/forecasts/aeo/pdf/0383r2014Ipdf
EIA (U.S. Energy Information Administration). (2014b). Natural gas. U.S. Energy Information
Administration: Independent statistics and analysis, http: //www.eia.gov/naturalgas/
EIA (U.S. Energy Information Administration). (2015). Lower 48 states shale plays.
http://www.eia.gov/oil gas/rpd/shale gas.pdf
EIA (U.S. Energy Information Administration). (2016a). Annual energy outlook 2016 early release:
Annotated summary of two cases. (DOE/EIA-0383ER(2016)). Washington, D.C.
http://www.eia.gov/forecasts/aeo/er/
EIA (U.S. Energy Information Administration). (2016b). Crude oil and natural gas drilling activity
measurements (Number of rigs). Washington, D.C.
EIA (U.S. Energy Information Administration). (2016c). Hydraulic fracturing accounts for about half
of current U.S. crude oil production.
http://www, eia. gov/todavinenergy/detail. cfm?id=2 5 372
EIA (U.S. Energy Information Administration). (2016d). Hydraulically fractured wells provide two-
thirds of U.S. natural gas production.
http://www, eia. gov/todavinenergy/detail. cfm?id=2 6112
EIA (U.S. Energy Information Administration). (2016e). Open data API. U.S. crude oil first purchase
price, Annual. http://www.eia.gov/opendata/qb.cfm?sdid=PET.FOOOOOO 3.A
EIA (U.S. Energy Information Administration). (2016f). Open data API. U.S. natural gas wellhead
price, Annual. http://www.eia.gov/opendata/qb.cfm?sdid=NG.N9190US3.A
EIA (U.S. Energy Information Administration). (2016g). Shale in the United States.
http: //www.eia.gov/energy in brief/article/shale in the united states.cfm
E ice man. GA. (1986). Hazardous organic wastes from natural gas production, processing and
distribution: Environmental fates. (WRRI report, no. 227). New Mexico: Water Resources
Research Institute, http://wrri.nmsu.edu/publish/techrpt/abstracts/abs227.html
Eisner. L: Fischer. T: Le Calvez. TH. (2006). Detection of repeated hydraulic fracturing (out-of-zone
growth) by microseismic monitoring. The Leading Edge (Tulsa) 25: 548-554.
http://dx.doi.Org/10.1190/l.2202655
El Paso Water Utilities. (2016). Personal communication: Phone conversation and email exchange
between Scott Reinert, El Paso Water Utilities, to Nathan Wiser, U.S. EPA.
Elbel. 1: Britt. L. (2000). Fracture treatment design. In MJ Economides; KG Nolte (Eds.), Reservoir
Stimulation (3rd ed.). New York, NY: John Wiley & Sons.
Elliott. EG: Ettinger. AS: Leaderer. BP: Bracken. MB: Deziel. NC. (2016). A systematic evaluation of
chemicals in hydraulic-fracturing fluids and wastewater for reproductive and
developmental toxicity. J Expo Sci Environ Epidemiol.
http://dx.doi.org/10.1038/ies.2015.81
Ellis. RT. (1997). Severe hypernatremia from sea water ingestion during near-drowning in a
hurricane. West J Med 167: 430-433.
11-19
-------
Chapter 11 - References
Eltschlager. KK: Hawkins. TW: Ehler. WC: Baldassare. F. (2001). Technical measures for the
investigation and mitigation of fugitive methane hazards in areas of coal mining.
Washington, D.C.: U.S. Department of the Interior.
http://www.osmre.gov/resources/librarv/ghm/methane.pdf
Ely. TW. (1989). Chapter 7: Fracturing fluids and additives. In JL Gidley; SA Holditch; DE Nierode;
RW Veatch Jr (Eds.), Recent Advances in Hydraulic Fracturing (pp. 131-146). Richardson,
TX: Society of Petroleum Engineers.
Enform. (2013). Interim industry recommended practice 24: Fracture stimulation: Interwellbore
communication 3/27/2013 (lsted.). (IRP 24). Calgary, Alberta: Enform Canada.
Engelder. T. (2012). Capillary tension and imbibition sequester frack fluid in Marcellus gas shale
[Letter], Proc Natl Acad Sci USA 109: E3625; author reply E3626.
http://dx.doi.org/10.1073/pnas.1216133110
Engelder. T: Cathles. LM: Brvndzia. LT. (2014). The fate of residual treatment water in gas shale.
J Unconven Oil Gas Resour 7: 33-48. http://dx.doi.Org/10.1016/i.iuogr.2014.03.002
Engle. MA: Bern. CR: Healv. RW: Sams. II: Zupancic. TW: Schroeder. KT. (2011). Tracking solutes and
water from subsurface drip irrigation application of coalbed methaneproduced waters,
Powder River Basin, Wyoming. Environmental Geosciences 18: 169-187.
Engle. MA: Reyes. FR: Varonka. MS: Orem. WH: Ma. L: Ianno. AT: Schell. TM: Xu. P: Carroll. KC.
(2016). Geochemistry of formation waters from the Wolfcamp and "Cline" Shales: Insights
into brine origin, reservoir connectivity, and fluid flow in the Permian Basin, USA. Chem
Geol 425: 76-92. http://dx.doi.Org/10.1016/i.chemgeo.2016.01.025
Entrekin. S: Evans-White. M: Tohnson. B: Hagenbuch. E. (2011). Rapid expansion of natural gas
development poses a threatto surface waters. Front Ecol Environ 9: 503-511.
http://dx.doi.org/10.1890/110053
Entrekin. SA: Malonev. K0: Kapo. KE: Walters. AW: Evans-White. MA: Klemow. KM. (2015). Stream
vulnerability to widespread and emergent stressors: A focus on unconventional oil and gas.
PLoS ONE 10: e0137416. http://dx.doi.org/10.1371 /iournal.pone.0137416
Environment Canada. (2004). Threats to water availability in Canada, http://www, ec.gc.ca/inre-
nwri/default asp?lang=En&n=0 CD 66675 -1
ERCB (Energy Resource Conservation Board). (2012). Midway Energy Ltd. hydraulic fracturing
incident: Interwellbore communication January 13, 2012. (ERCB Investigation Report, Red
Deer Field Centre). Calgary, Alberta: Energy Resources Conservation Board.
ESN Rocky Mountain (Environmental Services Network Rocky Mountain). (2003). Produced gas
and water testing of CBM gas wells in the Raton Basin. Golden, CO.
http://hermes.cde.state.co.us/drupal/islandora/obiect/co%3A7346
ESRI (Environmental Systems Research Institute Inc.). (2010). US states shapefile. Redlands,
California.
http://www.arcgis.com/home/item.html?id=la6cae723afl4f9cae228bl33aebc620
Esser. BK: Beller. HR: Carroll. SA: Cherry. TA: Gillespie. I: Tackson. RB: Tordan. PD: Madrid. V: Morris.
TP: Parker. BL.: Stringfellow. W. T.: Varadharajan. C: Vengosh. A. (2015). Recommendations
on model criteria for groundwater sampling, testing, and monitoring of oil and gas
development in California. (LLNLTR669645). Livermore, CA: Lawrence Livermore National
Laboratory.
http://www.waterboards.ca.gov/water issues/programs/groundwater/sb4/docs/llnl reco
mmendations reportpdf
11-20
-------
Chapter 11 - References
EWI (Energy Water Initiative). (2015). U.S. onshore unconventional exploration and production
water management case studies. Houston, TX: CH2M Hill Engineers, Inc.
https://www.owrb.ok.gOv/2060/PWWG/Resources/EWI Case Studies Reportpdf
Fakhru'l-Razi. A: Pendashteh. A: Abdullah. LC: Biak. PR: Madaeni. SS: Abidin. ZZ. (2009). Review of
technologies for oil and gas produced water treatment [Review], J Hazard Mater 170: 530-
551.
Famiglietti. IS: Lo. M: Ho. SL: Bethune. 1: Anderson. KT: Sved. TH: Swenson. SC: de Linage. CR: Rodell.
M. (2011). Satellites measure recent rates of groundwater depletion in California's Central
Valley. Geophys Res Lett 38: L03403. http://dx.doi.org/1 0.1029/2010GL046442
Feng. Y: Podnos. E: Gray. KE. (2015). Well integrity analysis: 3D numerical modeling of cement
interface debonding. (ARMA 16-246). Alexandria, TX: American Rock Mechanics
Association.
Ferrar. KT: Michanowicz. PR: Christen. CL: Mulcahv. N: Malone. SL: Sharma. RK. (2013). Assessment
of effluent contaminants from three facilities discharging Marcellus Shale wastewater to
surface waters in Pennsylvania. Environ Sci Technol 47: 3472-3481.
Fertl. WH: Chilingar. GV. (1988). Total organic carbon content determined from well logs. SPE
Formation Evaluation 3: 407-419. http://dx.doi.org/10.2118/15612-PA
Feth. THF. (1965). Preliminary map of the conterminous United States showing depth to and quality
of shallowest ground water containing more than 1,000 parts per million dissolved solids.
(Hydrologic Investigations Atlas HA-199). Washington, D.C.: U.S. Geological Survey.
https://pubs.er.usgs.gov/publication/hal99
Fink. IK. (2003). Oil field chemicals. Boston, MA: Gulf Professional Publishing.
Finkel. M: Hays. 1: Law. A. (2013). The shale gas boom and the need for rational policy. Am J Public
Health 103: 1161-1163. http://dx.doi.org/10.2105/ATPH.2013.301285
Fisher. IB: Sublette. KL. (2005). Environmental releases from exploration and production
operations in Oklahoma: Type, volume, causes, and prevention. Environmental Geosciences
12: 89-99. http://dx.doi.org/10.1306/eg.l 1160404039
Fisher. K. (2012). Green frac fluid chemistry optimizes well productivity, environmental
performance [Magazine], The American Oil and Gas Reporter, March 2012, 4.
Fisher. M: Warpinski. N. (2012). Hydraulic fracture height growth: Real data. SPE Prod Oper 27: 8-
19. http: //dx.doi.org/10.2118/145949-PA
Fisher. RS. (1998). Geologic and geochemical controls on naturally occurring radioactive materials
(NORM) in produced water from oil, gas, and geothermal operations. Environmental
Geosciences 5: 139-150. http://eg.geoscienceworld.Org/content/5/3/139
Fitzgerald. DP: McGhee. BF: McGuire. TA. (1985). Guidelines for 90% accuracy in zone-isolation
decisions. J PetTech 37: 2013-2022. http://dx.doi.org/10.2118/12141-PA
Fiaer. E: Holt. RM: Horsrud. P: Raaen. AM: Risnes. R. (2008). Petroleum Related Rock Mechanics
(2nd edition ed.). Amsterdam, The Netherlands: Elsevier.
Fleckenstein. WW: Eustes. AW: Stone. CH: Howell. PK. (2015). An assessment of risk of migration of
hydrocarbons or fracturing fluids to fresh water aquifers: Wattenberg Field, CO. Presented
at SPE Kuwait Oil and Gas Show and Conference, 11-14 October, 2015, Mishref, Kuwait:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/1754Ql-MS
Flewelling. SA: Sharma. M. (2014). Constraints on upward migration of hydraulic fracturing fluid
and brine. Ground Water 52: 9-19. http://dx.doi.org/10.1111 /gwat. 12095
11-21
-------
Chapter 11 - References
Flewelling. SA: Sharma. M. (2015). Comment on "Hydraulic fracturing in faulted sedimentary
basins: Numerical simulation of potential contamination of shallow aquifers over long time
scales" by Claire Gassiatetal. Water Resour Res 51: 1872-1876.
http://dx.doi.Org/10.1002/2014WR015904
Flewelling. SA: Tvmchak. MP: Warpinski. N. (2013). Hydraulic fracture height limits and fault
interactions in tight oil and gas formations. Geophys Res Lett 40: 3602-3606.
http: / /dx. doi. or g/10.10 0 2 /grl. 50707
Flournov. RM: Feaster. TH. (1963). Field observations on the use of the cement bond log and its
application to the evaluation of cementing problems. Presented at the Fall Meeting of the
Society of Petroleum Engineers of AIME, 6-9 October, 1963, New Orleans, LA: Society of
Petroleum Engineers. http://dx.doi.org/10.2118/632-MS
Fontenot. BE: Hunt. LR: Hildenbrand. ZL: Carlton. DP. Tr: Oka. H: Walton. TL: Hopkins. D: Osorio. A:
Biorndal. B: Hu. OH: Schug. KA. (2013). An evaluation of water quality in private drinking
water wells near natural gas extraction sites in the Barnett Shale formation. Environ Sci
Technol 47: 10032-10040. http://dx.doi.org/l 0.1021 /es4011724
Fountain. TC: Tacobi. RD. (2000). Detection of buried faults and fractures using soil gas analysis.
Environmental and Engineering Geoscience 6: 201-208.
http://dx.doi.Org/10.2113/gseegeosci.6.3.201
FracFocus. (2016). FracFocus. Available online at www.fracfocus.org
Fredd. CN: Olsen. TN: Brenize. G: Ouintero. BW: Bui. T: Glenn. S: Bonev. CL. (2004). Polymer-free
fracturing fluid exhibits improved cleanup for unconventional natural gas well applications.
Presented at SPE Eastern Regional Meeting, 15-17 September, 2004, Charleston, WV:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/91433-MS
Freeze. RA: Cherry. TA. (1979a). Groundwater. Englewood Cliffs, NJ: Prentice-Hall.
Freeze. RA: Cherry. TA. (1979b). Groundwater Contamination. In Groundwater. Upper Saddle River,
NJ: Prentice Hall.
Gadhamshettv. V: Shrestha. N: Chilkoor. G: Bathi. TR. (2015). Emerging environmental impacts of
unconventional oil development in the Bakken Formation in the Williston Basin of western
North Dakota. In DL Drogos (Ed.), Hydraulic Fracturing: Environmental Issues (pp. 151-
180). Washington, D.C.: American Chemical Society, http: / /dx. doi. o r g/10.10 21 /bk- 2 015 -
1216.ch007
Gallegos. TT: Varela. BA. (2015). Trends in hydraulic fracturing distributions and treatment fluids,
additives, proppants, and water volumes applied to wells drilled in the United States from
1947 through 2010: Data analysis and comparison to the literature. (U.S. Geological Survey
Scientific Investigations Report 20145131). Reston, VA: U.S. Geological Survey.
http://dx.doi.org/10.3133/sir20145131
Gallegos. TT: Varela. BA: Haines. SS: Engle. MA. (2015). Hydraulic fracturing water use variability in
the United States and potential environmental implications. Water Resour Res 51: 5839-
5845. http://dx.doi.org/10.1002/2015WR017278
Gassiat. C: Gleeson. T: Lefebvre. R: Mckenzie. I. (2013). Hydraulic fracturing in faulted sedimentary
basins: Numerical simulation of potential contamination of shallow aquifers over long time
scales. Water Resour Res 49: 8310-8327. http://dx.doi.org/10.1002/2013WRQ14287
Geiver. L. (2013). Frac water treatment yields positive results for Houston Co.
http://www.thebakken.eom/articles/20/frac-water-treatment-yields-positive-results-for-
houston-co
11-22
-------
Chapter 11 - References
Georgakakos. A: Fleming. P: Dettinger. M: Peters-Lidard. C: Richmond. TC: Reckhow. K: White. K:
Yates. D. (2014). Water resources. In JM Melillo; TC Richmond; GW Yohe (Eds.), Climate
Change Impacts in the United States (pp. 69-112). Washington, D.C.: U.S. Global Change
Research Program, http://www.globalchange.gov/ncadac
George. PG: Mace. RE: Petrossian. R. (2011). Aquifers of Texas. (Report 380). Austin, TX: Texas
Water Development Board.
http://www.twdb.texas.gov/publications/reports/numbered reports/doc/R3 80 Aquiferso
fTexas.pdf
Gidlev. TL: Holditch. SA: Nierode. DE: Veatch Tr.. RW. (1989). Recent advances in hydraulic
fracturing. Richardson, TX: Society of Petroleum Engineers.
Gilmore. K: Hupp. R: Glathar. 1. (2013). Transport of hydraulic fracturing water and wastes in the
Susquehanna River Basin, Pennsylvania. J Environ Eng 140: B4013002.
http://dx.d0i.0rg/l 0.1061 /fASCElEE.l 943-7870.0000810
Gleick. P. (2008). Why dontwe get our drinking water from the ocean by taking the salt out of
seawater? http://www.scientificamerican.com/article/why-dont-we-get-our-drinking-
water-from-the-ocean/
GNB (Government of New Brunswick). (2013). Responsible environmental managementof oil and
natural gas activities in New Brunswick - Rules for industry. New Brunswick, Canada.
http://www2.gnb.ca/content/dam/gnb/Corporate/pdf/ShaleGas/en/RulesforIndustry.pdf
GNB (Government of New Brunswick). (2015). FAQs hydraulic fracturing (fraccing). New
Brunswick, Canada.
http://www2.gnb.ca/content/dam/gnb/Corporate/pdf/ShaleGas/en/FA0 HvdraulicFractu
ring.pdf
Goldstein. BP: Brooks. BW: Cohen. SD: Gates. AE: Honevcutt. ME: Morris. IB: Orme-Zavaleta. I:
Penning. TM: Snawder. 1. (2014). The role of toxicological science in meeting the challenges
and opportunities of hydraulic fracturing. Toxicol Sci 139: 271-283.
Gonneea. ME: Morris. PI: Dulaiova. H: Charette. MA. (2008). New perspectives on radium behavior
within a subterranean estuary. Mar Chem 109: 250-267.
http://dx.doi.0rg/lO.lOl6/j.marchem.2OO7.i2.OO2
Goodwin. KT: Crook. RT. (1992). Cement sheath stress failure. SPE Drilling & Completion 7: 291-296.
http://dx.doi.org/10.2118/2Q453-PA
Goodwin. S: Carlson. K: Knox. K: Douglas. C: Rein. L. (2014). Water intensity assessment of shale gas
resources in the Wattenberg field in northeastern Colorado. Environ Sci Technol 48: 5991-
5995. http://dx.doi.org/10.1021/es404675h
Gorodv. AW. (2012). Factors affecting the variability of stray gas concentration and composition in
groundwater. Environmental Geosciences 19: 17-31.
http://dx.doi.org/10.1306/eg.12081111Q13
Government of Canada and U.S. EPA (Government of Canada and U.S. Environmental Protection
Agency). (1995). The great lakes: An environmental atlas and resource book [EPA Report]
(3rd ed.). Chicago, IL: U.S. Environmental Protection Agency.
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1004ICU.txt
Gradient. (2013). National human health risk evaluation for hydraulic fracturing fluid additives.
httD://www.energv.senate.gov/public/index.cfm/files/serve?File id=53a41a78-c06c-
4695-a7be-84225aa7230f
11-23
-------
Chapter 11 - References
Greeman. DT: Rose. AW: Washington. TW: Dobos. RR: Ciolkosz. ET. (1999). Geochemistry of radium in
soils of the Eastern United States. Appl Geochem 14: 365-385.
http://dx.doi. org/10.1016/S0883-2927r98100059-6
Greenhunter (Greenhunter Resources). (2014). Oilfield water management solutions.
Gregory. KB: Vidic. RD: Dzombak. DA. (2011). Water management challenges associated with the
production of shale gas by hydraulic fracturing. Elements 7: 181-186.
Grinberg. A. (2014). In the pits: Oil and gas wastewater disposal into open unlined pits and the
threat to California's water and air.
http://www.cleanwateraction.org/files/publications/ca/CA%20Qil%20and%20Gas%20In
%20the%20Pits%20Facesheetpdf
Grinberg. A. (2016). Still in the pits: Update on oil and gas wastewater disposal in California.
Washington, D.C.: Clean Water Action/Clean Water Fund.
Gross. SA: Avens. HI: Banducci. AM: Sahmel. 1: Panko. TM: Tvermoes. BE. (2013). Analysis of BTEX
groundwater concentrations from surface spills associated with hydraulic fracturing
operations. J Air Waste Manag Assoc 63: 424-432.
http://dx.d0i.0rg/l 0.1080/10962247.2012.759166
Gu. H: Siebrits. E. (2008). Effect of formation modulus contrast on hydraulic fracture height
containment. SPE Prod Oper 23: 170-176. http://dx.doi.org/10.2118/103822-PA
Gu. M: Mohanty. KK. (2014). Effect of foam quality on effectiveness of hydraulic fracturing in shales.
International Journal of Rock Mechanics and Mining Sciences 70: 273-285.
http: / /dx. doi. or g/10.1016 /i. ii rmms. 2 014.05.013
Guerra. K: Dahm. K: Dundorf. S. (2011). Oil and gas produced water management and beneficial use
in the western United States. (Science and Technology Program Report No. 157). Denver,
CO: U.S. Department of the Interior, Bureau of Reclamation.
Gupta. DVS: Hlidek. BT. (2009). Frac fluid recycling and water conservation: A case history. In 2009
Hydraulic Fracturing Technology Conference, 19-21 January, 2009, The Woodlands, TX:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/119478-MS
Gupta. DVS: Valko. P. (2007). Fracturing fluids and formation damage. In M Economides; T Martin
(Eds.), Modern Fracturing: Enhancing Natural Gas Production (pp. 227-279). Houston, TX:
Energy Tribune Publishing Inc.
Guralnick. MS. (2016). Jurisdiction. In Fracking: Law and Policy. New York, NY: Wolters Kluwer.
Gurdak. II: McMahon. PB: Dennehv. K: Oi. SL. (2009). Water quality in the high plains aquifer,
Colorado, Kansas, Nebraska, New Mexico, Oklahoma, South Dakota, Texas, and Wyoming,
1999 - 2004. (GSC 1337). Reston, VA: U.S. Geological Survey.
http: //pubs.usgs. gov/circ/13 3 7/
Gustafson. IB: Griffith-Tell. I: Orem. D. (1997). TPH Working Group series, volume 3: Selection of
representative TPH fractions based on fate and transport considerations. Amherst, MA:
Association for Environmental Health and Sciences.
GWPC (Groundwater Protection Council). (2009). State oil and natural gas regulations designed to
protect water resources. Morgantown, WV: U.S. Department of Energy, National Energy
Technology Laboratory.
http://www.gwpc.org/sites/default/files/state oil and gas regulations designed to protec
t water resources O.pdf
11-24
-------
Chapter 11 - References
GWPC (Groundwater Protection Council). (2012). FracFocus well records: January 1, 2011, through
February 27, 2012. Available online at
https://fracfocusdata.org/DisclosureSearch/Search.aspx
GWPC (Groundwater Protection Council). (2014). State oil and natural gas regulations designed to
protect water resources. Morgantown, WV: U.S. Department of Energy, National Energy
Technology Laboratory.
http://www.gwpc.org/sites/default/files/files/0il%20and%20Gas%20Regulation%20Rep
ort%20Hyperlinked%20Version%20Final-rfs.pdf
GWPC and ALL Consulting (Ground Water Protection Council and ALL Consulting). (2009). Modern
shale gas development in the United States: A primer. (DE-FG26-04NT15455). Washington,
DC.: U.S. Department of Energy, Office of Fossil Energy and National Energy Technology
Laboratory.
http://www.gwpc.org/sites/default/files/Shale%20Gas%20Primer%202009.pdf
GWPC and IOGCC (Ground Water Protection Council and Interstate Oil and Gas Compact
Commission). (2014). Fracturing fluid management http://fracfocus.org/hydraulic-
fracturing-how-it-works/drilling-risks-safeguards (accessed December 12, 2014).
Habicht. S: Hanson. L: Faeth. P. (2015). The potential environmental impact from fracking in the
Delaware River basin. (CNA Research Memorandum, Control Number IRM-2015-U-
011300). Arlington, VA: CNA Corporation.
Halliburton. (1988). Primer on Hydraulic Fracturing. Provided to EPA on March 2, 2011. Available
at Docket ID: EPA-HQ-ORD-2010-0674-1634. (HESI-3031).
http://www, re gulations. gov /#! docume ntD etail :D=EP A-HO-ORD- 2010-0674-1634
Halliburton. (2008). Hydraulic fracturing of coalseams. In Coalbed Methane: Principles and
Practices. Houston, TX.
Halliburton. (2013). Hydraulic fracturing 101.
Halliburton. (2014). Hydraulic fracturing 101.
Haluszczak. LP: Rose. AW: Kump. LR. (2013). Geochemical evaluation of flowback brine from
Marcellus gas wells in Pennsylvania, USA. Appl Geochem 28: 55-61.
http://dx.doi.Org/10.1016/i.apgeochem.2012.10.002
Hamieh. BM: Beckman. JR. (2006). Seawater desalination using Dewvaporation technique:
Theoretical development and design evolution. Desalination 195: 1-13.
http://dx.doi.Org/10.1016/i.desal.2005.09.034
Hammack. R: Harbert. W: Sharma. S: Stewart. B: Capo. R: Wall. A: Wells. A: Diehl. R: Blaushild. D:
Sams. 1: Veloski. G. (2014). An evaluation of fracture growth and gas/fluid migration as
horizontal Marcellus Shale gas wells are hydraulically fractured in Greene County,
Pennsylvania. (NETL-TRS-3-2014). Pittsburgh, PA: U.S. Department of Energy, National
Energy Technology Laboratory.
http://www.netl.doe.gov/File%20Library/Research/onsite%20research/publications/NE
TL-TRS-3-2014 Greene-Countv-Site 20140915 1 l.pdf
Hammer. R: VanBriesen. 1. (2012). In fracking's wake: New rules are needed to protect our health
and environment from contaminated wastewater. New York, NY: Natural Resources
Defense Council (NRDC). http://www.nrdc.org/energv/files/fracking-wastewater-
fullreportpdf
11-25
-------
Chapter 11 - References
Hammond. PA. (2016). The relationship between methane migration and shale-gas well operations
near Dimock, Pennsylvania, USA. Hydrogeo J 24: 503-519.
http://dx.doi. org/10.1007/sl 0040-015-1332-4
Hansen. E: Mulvanev. D: Betcher. M. (2013). Water resource reporting and water footprint from
Marcellus Shale development in West Virginia and Pennsylvania. Durango, CO: Earthworks
Oil & Gas Accountability Project
http://www.downstreamstrategies.com/documents/reports publication/marcellus wv pa.
pdf
Harkness. IS: Dwver. GS: Warner. NR: Parker. KM: Mitch. WA: Vengosh. A. (2015). Iodide, bromide,
and ammonium in hydraulic fracturing and oil and gas wastewaters: Environmental
implications. Environ Sci Technol 49: 1955-1963. http://dx.doi.org/10.1021/es504654n
Harms. WM: Yeager. R. (1987). Diesel-based gel concentrate reduces stimulation costs. Oil and Gas
Journal 85: 37-39.
Harrison. SS. (1983). Evaluating system for ground-water contamination hazards due to gas-well
drilling on the glaciated Appalachian Plateau. Ground Water 21: 689-700.
http://dx.d0i.0rg/l 0.1111 /i.l 745-6584.1983.tb01940.x
Harrison. SS. (1985). Contamination of aquifers by overpressurizing the annulus of oil and gas
wells. Ground Water 23: 317-324.
Harto. CB: Smith. KP: Kamboi. S: Ouinn. 11. (2014). Radiological dose and risk assessment of landfill
disposal of technologically enhanced naturally occurring radioactive materials (TEN0RM)
in North Dakota. (ANL/EVS-14/13). US DOE Argonne National Laboratory.
http://www.ndhealth.gov/ehs/tenorm/ArgonneStudy/ANL-
NDDH%20TENQRM%20Landfill%20Studv%20fANL%20EVS-
14 13")%20Final%20Report.pdf
Hayes. T. (2009). Sampling and analysis of water streams associated with the development of
Marcellus Shale gas. Des Plaines, IL: Marcellus Shale Coalition.
http://energyindepth.org/wp-content/uploads/marcellus/2012/ll/MSCommission-
Reportpdf
Hayes. T. (2011). Characterization of Marcellus and Barnett shale flowback waters and technology
development for water reuse. Presented at Hydraulic Fracturing Study: March 2011
Technical Workshop on Water Resource Management, March 29-30, 2011, Arlington, VA.
http://www2.epa.gov/hfstudy/characterization-marcellus-and-barnett-shale-flowback-
waters-and-technologv-development
Hayes. T: Severin. B. (2012a). Characterization of flowback water from the the Marcellus and the
Barnett Shale regions. Barnett and Appalachian Shale water management and reuse
technologies. (08122-05.09; Contract 08122-05).
http://www.rpsea.org/media/files/proiect/2146b3a0/08122-05-RT-
Characterization Flowback Waters Marcellus Barnett Shale Regions-03-20-12.pdf
Hayes. T: Severin. BF. (2012b). Evaluation of the aqua pure mechanical vapor recompression
system in the treatment of shale gas flowback water - Barnett and Appalachian Shale water
management and reuse technologies. (08122-05.11).
http://barnettshalewater.org/documents/08122-05.ll-EvaluationofMVR-3-12-2012.pdf
Hayes. TP: Halldorson. B: Horner. P: Ewing. 1: Werline. TR: Severin. BF. (2014). Mechanical vapor
recompression for the treatment of shale-gas flowback water. Oil and Gas Facilities 3: 54-
62.
11-26
-------
Chapter 11 - References
Havmond. D. (1991). The Austin Chalk - An overview. HGS Bulletin 33: 27-30, 32, 34.
He. Z. (2011). Flow of gas and water in hydraulically fractured shale gas reservoirs. EPA HF
Workshop, March 28-29, 2011, Arlington, VA.
Healv. RW: Bartos. TT: Rice. CA: Mckinley. MP: Smith. BP. (2011). Groundwater chemistry near an
impoundment for produced water, Powder River Basin, Wyoming, USA. J Hydrol 403: 37-48.
http://dx.doi.Org/10.1016/j.jhydrol.2011.03.042
Healv. RW: Rice. CA: Bartos. TT: Mckinley. MP. (2008). Infiltration from an impoundment for coal-
bed natural gas, Powder River Basin, Wyoming: Evolution of water and sediment chemistry.
Water Resour Res 44. http://dx.doi.org/10.1029/2007wr006396
Heermann. SE: Powers. SE. (1998). Modeling the partitioning of BTEX in water-reformulated
gasoline systems containing ethanol. J Contam Hydrol 34: 315-341.
http://dx.doi. org/10.1016/S01 69-7722r98100099-0
Hildenbrand. ZL: Carlton. DP: Fontenot. BE: Meik. TM: Walton. TL: Taylor. TT: Thacker. IB: Korlie. S:
Shelor. CP: Henderson. D: Kadio. AF: Roelke. CE: Hudak. PF: Burton. T: Rifai. HS: Schug. KA.
(2015). A comprehensive analysis of groundwater quality in the Barnett Shale region.
Environ Sci Technol 49: 8254-8262. http://dx.doi.org/10.1021/acs.est.5b01526
Hildenbrand. ZL: Carlton. DP: Fontenot. BE: Meik. TM: Walton. TL: Thacker. IB: Korlie. S: Shelor. CP:
Kadjo. AF: Clark. A: Usenko. S: Hamilton. IS: Mach. PM: Verbeck. GF: Hudak. P: Schug. KA.
(2016). Temporal variation in groundwater quality in the Permian Basin of Texas, a region
of increasing unconventional oil and gas development Sci Total Environ 562: 906-913.
http://dx.doi.Org/10.1016/i.scitotenv.2016.04.144
Hladik. ML: Focazio. Ml: Engle. M. (2014). Discharges of produced waters from oil and gas
extraction via wastewater treatment plants are sources of disinfection by-products to
receiving streams. Sci Total Environ 466: 1085-1093.
http ://dx. doi. or g/10.1016/i. scitotenv. 2013.08.008
Hoelzer. K: Sumner. AT: Karatum. 0: Nelson. RK: Drollette. BP: O'Connor. MP: D'Ambro. EL:
Getzinger. GT: Ferguson. PL: Reddv. CM: Eisner. M: Plata. PL. (2016). Indications of
transformation products from hydraulic fracturing additives in shale-gas wastewater.
Environ Sci Technol 50: 8036-8048. http://dx.doi.org/10.1021/acs.est.6b00430
Holditch. SA. (2007). Chapter 8: Hydraulic fracturing. In JP Clegg (Ed.), Petroleum Engineering
Handbook (pp. IV-323 - IV-366). Richardson, TX: Society of Petroleum Engineers.
Holtsclaw. 1: Loveless. P: Saini. R: Fleming. I. (20111. SPE 146832: Environmentally-focused
crosslinked gel system results in high retained proppant-pack conductivity. Presented at
Society of Petroleum Engineers Annual Conference, November 2, 2011, Penver, CO.
https://www.onepetro.org/conference-paper/SPE-146832-MS
Horn. A: Hu. I: Patton. M. (2013). QA/QC of water blending enhances crosslinked gel completions.
http://content.stockpr.com/hydrozonix/files/downloads/1013HEP-hydrozonix.pdf
Horner. RM: Harto. CB: Murphy. PI: Clark. CE: Shupvrt. TP. (2014). Water use and management in
the Bakken shale oil play. (POE Award No.: FWP 49462). Pittsburgh, PA: Pepartment of
Energy (POE), National Energy Technology Laboratory (NETL).
http://www.ipd.anl.gOv/anlpubs/2014/05/104645.pdf
Horsey. CA. (1981). Pepositional environments of the Pennsylvanian Pottsville Formation in the
Black Warrior Basin of Alabama. Journal of Sedimentary Research 51: 799-806.
http: //dx.doi.org/10.1306/212F7PB5-2B24-11P7-8648000102C1865P
11-27
-------
Chapter 11 - References
Horton. S. (2012). Disposal of hydrofracking waste fluid by injection into subsurface aquifers
triggers earthquake swarm in central Arkansas with potential for damaging earthquake.
Seismological Research Letters 83: 250-260. http://dx.doi.Org/10.1785/gssii.83.2.250
Pennsylvania General Assembly (2011). House bill No. 1950, sessions of 2011 - Amending title 58
(oil and gas) of the Pennsylvania consolidated statues, and making a related repeal.
http://www.legis.state.pa.us/cfdocs/billinfo/billinfo.cfm?syear=2011&body=H&type=B&B
N=1950
House of Representatives (2009). Rep. 111-316. Department of the Interior, Environment, and
Related Agencies Appropriation Act, 2010: Conference report 111-316 (to accompany H.R.
2996). http://www.gpo.gov/fdsys/pkg/CRPT-lllhrpt316/pdf/CRPT-lllhrpt316.pdf
House of Representatives (U.S. House of Representatives). (2011). Chemicals used in hydraulic
fracturing. Washington, D.C.: U.S. House of Representatives, Committee on Energy and
Commerce, Minority Staff.
http://www.conservation.ca.gov/dog/general information/Documents/Hydraulic%20Frac
turing%20Report%204%2018%2011 .pdf
Houston. N: Blauch. M: Weaver. D: Miller. PS: O'Hara. D. (2009). Fracture-stimulation in the
Marcellus shale: Lessons learned in fluid selection and execution. In 2009 SPE Eastern
Regional Meeting: Limitless Potential/Formidable Challenges, 23-25 September, 2009,
Charleston, WV: Society of Petroleum Engineers. http: //dx.doi.org/10.2118/125987-MS
Hristozov. PR: Zabeo. A: Foran. C: Isigonis. P: Critto. A: Marcomini. A: Linkov. I. (2014). A weight of
evidence approach for hazard screening of engineered nanomaterials. Nanotoxicology 8:
72-87. http://dx.doi.org/10.3109/17435390.2012.750695
Hua. GH: Reckhow. PA: Kim. I. (2006). Effect of bromide and iodide ions on the formation and
speciation of disinfection byproducts during chlorination. Environ Sci Technol 40: 3050-
3056. http://dx.doi.org/10.1021/es0519278
Huang. IB: Keisler. 1: Linkov. I. (2011). Multi-criteria decision analysis in environmental sciences:
Ten years of applications and trends [Review], Sci Total Environ 409: 3578-3594.
http ://dx. doi. or g/10.1016/i. scitotenv. 2011.06.022
Hulme. P. (2005). CBM co-produced water management, disposal, treatment and use. Laramie, WY:
University of Wyoming, William P. Ruckelshaus Institute of Environment and Natural
Resources, http://www, uwyo.edu/haub// files/ docs/ruckelshaus/pubs/2005-cbm-
water-management-hulme.pdf
Hunter. TA: Moser. PH. (1990). Ground water availability in Jefferson County, Alabama: Geological
survey special map 224. Tuscaloosa, AL: Geological Survey of Alabama.
http://www.ogb.state.al.us/documents/pubs/onlinepubs/Special Maps/SM224 .PPF
Hvne. NT. (2012). Nontechnical Guide to Petroleum Geology, Exploration, Prilling and Production
(3rd ed.). Tulsa, OK: PennWell Corporation.
IAEA (International Atomic Energy Agency). (2014). The Environmental Behaviour of Radium
[Revised edition], Vienna, Austria.
http://www-pub.iaea.org/MTCP/Publications/PPF/trs476 web.pdf
Igunnu. ET: Chen. GZ. (2014). Produced water treatment technologies. International Journal of Low-
Carbon Technologies 9: 157-177. http://dx.doi.org/10.1093/iilct/cts049
IHS (Global Insight). (2013). Americas new energy future: The unconventional oil and gas
revolution and the US economy. Pouglas County, CO.
11-28
-------
Chapter 11 - References
http://www.energyxxi.org/sites/default/files/pdf/Americas New Energy Future Phase3.p
df
IHS Inc. (Information Handling Services, Inc.). (2014). U.S. Petroleum Information/Dwights LLC
database, PIDM 2.5 - Data management system, edited. Englewood, CO: IHS Inc.
Ingraffea. AR: Wells. MT: Santoro. RL: Shonkoff. SB. (2014). Assessment and risk analysis of casing
and cement impairment in oil and gas wells in Pennsylvania, 2000-2012. Proc Natl Acad Sci
USA 111: 10955-10960. http://dx.doi.org/10.1073/pnas.1323422111
IOGCC (Interstate Oil and Gas Compact Commission). (2002). States experience with hydraulic
fracturing: A survey of the Interstate Oil and Gas Compact Commission.
http://groundwork.iogcc.ok.gOv/sites/default/files/09IOG5571 StatesExperience%20w-
HvdFrac.pdf
IOGCC (Interstate Oil and Gas Compact Commission). (2008). Protecting our country's resources:
The states' case, orphaned well plugging initiative. Oklahoma City, OK.
http://iogcc.publishpath.com/Websites/iogcc/pdfs/2008-Protecting-Our-Country's-
Resources-The-States'-Case.pdf
IOGCC (Interstate Oil and Gas Compact Commission). (2015). Groundwork: Hydraulic fracturing -
State progress. httpV/groundwork.iogcc.ok.gov/topics-index/hydraulic-fracturing/state-
progress
IPCC (Intergovernmental Panel on Climate Change). (2007). Climate change 2007: Impacts,
adaptation and vulnerability. Cambridge, UK: Cambridge University Press.
https://www.ipcc.ch/publications and data/publications ipcc fourth assessment report
wg2 report impacts adaptation and vulnerabilitv.htm
Irwin. C. (2013). Hydraulic fracturing: A way to go greener?
http://breakingenergy.eom/2013/04/23/hydraulic-fracturing-a-way-to-go-greener/
Tackson. G: Flores. C: Abolo. N: Lawal. H. (2013a). A novel approach to modeling and forecasting frac
hits in shale gas wells. Presented at EAGE Annual Conference & Exhibition incorporating
SPE Europec, June 10-13, 2013, London, UK. https://www.onepetro.org/conference-
paper/SPE-164898-MS
Tackson. RB: Carpenter. SR: Dahm. CN: Mcknight. DM: Naiman. RT: Postel. SL: Running. SW. (2001).
Water in a changing world. Ecol Appl 11: 1027-1045. http://dx.doi.org/10.189Q/1051-
0761 ("20011011 [1027:WTACW|2.0.CQ:2
Tackson. RB: Lowrv. ER: Pickle. A: Kang. M: Digiulio. D: Zhao. K. (2015). The depths of hydraulic
fracturing and accompanying water use across the United States. Environ Sci Technol 49:
8969-8976. http://dx.doi.org/10.1021/acs.est.5b01228
Tackson. RB: Vengosh. A: Carey. TW: Davies. RT: Darrah. TH: O'Sullivan. F: Petron. G. (2013b). The
Environmental Costs and Benefits of Fracking. Annu Rev Environ Resour 39: 327-362.
http://dx.doi.org/10.1146/annurev-environ-031113-144051
Tackson. RB: Vengosh. A: Darrah. TH: Warner. NR: Down. A: Poreda. RT: Osborn. SG: Zhao. K: Karr.
TP. (2013c). Increased stray gas abundance in a subset of drinking water wells near
Marcellus shale gas extraction. Proc Natl Acad Sci USA 110: 11250-11255.
http://dx.doi.org/10.1073/pnas.1221635110
Tackson. RE: Dussealt. MB. (2014). Gas release mechanisms from energy wellbores. Presented at
48th US Rock Mechanics/Geomechanics Symposium, June 1-4, 2014, Minneapolis,
Minnesota. https://www.onepetro.org/conference-paper/ARMA-2014-7753
11-29
-------
Chapter 11 - References
Tackson. RE: Gorodv. AW: Mayer. B: Roy. TW: Ryan. MC: Van Stempvoort. DR. (2013d). Groundwater
protection and unconventional gas extraction: The critical need for field-based
hydrogeological research. Ground Water 51: 488-510.
http: //dx. doi. or g/10.1111 /gwat. 12074
lacob. R. (2011). Incident action plan, Franchuk 44-20 SWH incident. Piano, Texas: Denbury
Onshore, LLC.
Tiang. M: Hendrickson. CT: Vanbriesen. TM. (2014). Life cycle water consumption and wastewater
generation impacts of a marcellus shale gas well. Environ Sci Technol 48: 1911-1920.
http://dx.doi.org/10.1021/es4047654
Tones. TR: Britt. LK. (2009). Design and placement In Design and Appraisal of Hydraulic Fractures.
Richardson, TX: Society of Petroleum Engineers.
Tudson. R: Richard. A: Dix. DT: Houck. K: Martin. M: Kavlock. R: Dellarco. V: Henry. T: Holderman. T:
Savre. P: Tan. S: Carpenter. T: Smith. E. (2009). The toxicity data landscape for
environmental chemicals [Review], Environ Health Perspect 117: 685-695.
http://dx.doi.org/10.1289/ehp.0800168
Tudson. RS: Martin. MT: Egeghv. P: Gangwal. S: Reif. DM: Kothiva. P: Wolf. M: Cathev. T: Transue. T:
Smith. D: Vail. 1: Frame. A: Mosher. S: Cohen Hubal. EA: Richard. AM. (2012). Aggregating
data for computational toxicology applications: The U.S. Environmental Protection Agency
(EPA) Aggregated Computational Toxicology Resource (ACToR) System [Review],
International Journal of Molecular Sciences 13: 1805-1831.
http://dx.doi.org/10.3390/iimsl3021805
Kahrilas. GA: Blotevogel. I: Stewart. PS: Borch. T. (2015). Biocides in hydraulic fracturing fluids: A
critical review of their usage, mobility, degradation, and toxicity. Environ Sci Technol 49:
16-32. http://dx.doi.org/10.1021/es5Q3724k
Kang. M: Baik. E: Miller. AR: Bandilla. KW: Celia. MA. (2015). Effective permeabilities of abandoned
oil and gas wells: Analysis of data from Pennsylvania. Environ Sci Technol 49: 4757-4764.
http://dx.doi.org/10.1021/acs.est5b0Q132
Kang. M: Tackson. RB. (2016). Salinity of deep groundwater in California: Water quantity, quality,
and protection. Proc Natl Acad Sci USA 113: 7768-7773.
http://dx.doi.Org/10.1073/pnas.1600400113
Kappel. WM. (2013). Dissolved methane in groundwater, Upper Delaware River Basin,
Pennsylvania and New York (pp. 1-6). (2013-1167). U. S. Geological Survey.
http://pubs.usgs.gov/of/2013/1167/pdf/ofr2013-1167.pdf
Kappel. WM: Nvstrom. EA. (2012). Dissolved methane in New York groundwater, 1999-2011.
(Open-File Report 20121162). Washington, D.C.: U.S. Geological Survey.
http: //pubs.usgs. gov/of/2 012/1162/
Kappel. WM: Williams. TH: Szabo. Z. (2013). Water resources and shale gas/oil production in the
Appalachian basin critical issues and evolving developments. (Open-File Report 20131137).
Troy, NY: U.S. Geological Survey, http: //pubs.usgs.gov/of/2013/1137/pdf/ofr2013-
1137.pdf
Kargbo. DM: Wilhelm. RG: Campbell. DT. (2010). Natural gas plays in the Marcellus Shale: Challenges
and potential opportunities. Environ Sci Technol 44: 5679-5684.
http://dx.doi.org/10.1021/es903811p
Kassotis. CD: Iwanowicz. LR: Akob. DM: Cozzarelli. IM: Mumford. AC: Orem. WH: Nagel. SC. (2016).
Endocrine disrupting activities of surface water associated with a West Virginia oil and gas
11-30
-------
Chapter 11 - References
industry wastewater disposal site. Sci Total Environ 557-558: 901-910.
http ://dx. doi. or g/10.1016/i. scitotenv. 2016.03.113
Kassotis. CD: Tillitt. DE: Wade Davis. 1: Hormann. AM: Nagel. SC. (2014). Estrogen and androgen
receptor activities of hydraulic fracturing chemicals and surface and ground water in a
drilling-dense region. Endocrinology 155: 897-907. http://dx.doi.org/10.1210/en.2Q13-
1697
Kaushal. SS: Groffman. PM: Likens. GE: Belt. KT: Stack. WP: Kelly. VR: Band. LE: Fisher. GT. (2005).
Increased salinization of fresh water in the northeastern United States. Proc Natl Acad Sci
USA 102: 13517-13520. http://dx.doi.org/10.1073 /pnas.0506414102
KDHE (Kansas Department of Health and Environment). (2014). Kansas seismic action plan.
http://www.kansas.com/news/state/articlel491Q84.ece/binary/PDF%20-
%20State%20of%20Kansas%20Seismic%20Action%20Plan.pdf
Kell. S. (2011). State oil and gas agency groundwater investigations and their role in advancing
regulatory reforms, a two-state review: Ohio and Texas. Ground Water Protection Council
(GWPC).
http://fracfocus.org/sites/default/files/publications/state oil gas agency groundwater in
vestigations optimized.pdf
Kelly. WR. (2008). Long-term trends in chloride concentrations in shallow aquifers near Chicago.
Ground Water 46: 772-781. http: //dx.doi.Org/10.llll /i.1745-6584.2008.00466.x
Kennedy /Tenks Consultants. (2002). Evaluation of technical and economic feasibility of treating
oilfield produced water to create a new water resource.
http://www.gwpc.org/sites/default/files/event-sessions/Roger Funston PWC2002 O.pdf
Khan. NA: Engle. M: Dungan. B: Holguin. F0: Xu. P: Carroll. KC. (2016). Volatile-organic molecular
characterization of shale-oil produced water from the Permian Basin. Chemosphere 148:
126-136. http://dx.doi.Org/10.1016/i.chemosphere.2015.12.116
Khanna. A: Kotousov. A. (2016). Controlling the height of multiple hydraulic fractures in layered
media. SPE Journal 21: 256-263. http://dx.doi.org/10.2118/176017-PA
Kharaka. YK: Kakouros. E: Abbott. MM. (2002). Environmental impacts of petroleum production: 1-
The fate of inorganic and organic chemicals in produced water from the Osage-Skiatook
Petroleum Environmental Research B site, Osage County, OK. 9th International Petroleum
Environmental Conference, October 22-25, 2002, Albuquerque, NM.
http://toxics.usgs.gov/pubs/ipec 9/environmental impactsl.html
Kharaka. YK: Kakouros. E: Thordsen. IT: Ambats. G: Abbott. MM. (2007). Fate and groundwater
impacts of produced water releases at OSPER B site, Osage County, Oklahoma. Appl
Geochem 22: 2164-2176. http://dx.doi.Org/10.1016/j.apgeochem.2007.04.005
Kim. GH: Wang. TY. (2014). Interpretation of hydraulic fracturing pressure in tight gas formations.
Journal of Energy Resources Technology 136: 032903.
http://dx.doi.Org/10.1115/l.4026460
Kim. 1: Moridis. GT. (2013). Development of the T+M coupled flowgeomechanical simulator to
describe fracture propagation and coupled flowthermalgeomechanical processes in
tight/shale gas systems. Computers and Geosciences 60: 184-198.
http://dx.doi.Org/10.1016/i.cageo.2013.04.023
Kim. 1: Moridis. GT. (2015). Numerical analysis of fracture propagation during hydraulic fracturing
operations in shale gas systems. International Journal of Rock Mechanics and Mining
Sciences 76: 127-137.
11-31
-------
Chapter 11 - References
Kim. T: Moridis. GT: Martinez. ER. (2016). Investigation of possible wellbore cement failures during
hydraulic fracturing operations. Journal of Petroleum Science and Engineering 139: 254-
263. http://dx.doi.Org/10.1016/i.petrol.2016.01.035
Kim. 1: Um. ES: Moridis. GT. (2014). Fracture propagation, fluid flow, and geomechanics of water-
based hydraulic fracturing in shale gas systems and electromagnetic geophysical
monitoring of fluid migration. SPE Hydraulic Fracturing Technology Conference, February
4-6, 2014, The Woodlands, TX. http://dx.doi.org/l 0.2118/168578-MS
King. G: King. D. (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 Prod Oper 28: 323-344.
http://dx.doi.org/10.2118/166142-PA
King. GE. (2010). Thirty years of gas shale fracturing: what have we learned? Presented at SPE
Annual Technical Conference and Exhibition, 19-22 September, 2010, Florence, Italy:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/133456-MS
King. GE. (2012). Hydraulic fracturing 101: Whatevery representative, environmentalist, regulator,
reporter, investor, university researcher, neighbor and engineer should know about
estimating frac risk and improving frac performance in unconventional gas and oil wells.
SPE Hydraulic Fracturing Technology Conference, February 6-8, 2012, The Woodlands, TX.
http://fracfocus.org/sites/default/files/publications/hydraulic fracturing 101.pdf
King. GE: Durham. D. (2015). Environmental aspects of hydraulic fracturing: What are the facts? In
DL Drogos (Ed.), Hydraulic fracturing: Environmental issues (pp. 144). Washington, D.C.:
American Chemical Society, http://dx.doi.org/10.1021/bk-2015-1216.ch001
King. GE: Valencia. RL. (2016). Well integrity for fracturing and re-fracturing: What is needed and
why? SPE Hydraulic Fracturing Technology Conference, February 9-11, 2016, The
Woodlands, TX. https://www.onepetro.org/conference-paper/SPE-179120-MS
King. P: Soraghan. M. (2016). Spills dropped 8% in 2015 as new drilling slowed [Magazine],
EnergyWire, July 21, 2016, Article 1060040567.
Kingsbury. TA: Delzer. GC: Hopple. TA. (2008). Anthropogenic organic compounds in source water of
nine community water systems that withdraw from streams, 2002-2005. (Scientific
Investigations Report 2008-5208). Reston, VA: U.S. Geological Survey.
http: / /pubs, usgs. gov/sir /2 008/5208/
Kirksev. 1. (2013). Optimizing wellbore integrity in well construction. Presented at North American
Wellbore Integrity Workshop, October 16-17, 2013, Denver, CO.
http://ptrc.ca/+pub/document/Kirksey%20-
%20Qptimizing%20Wellbore%2 01ntegritv.pdf
Klein. M: Kenealev. G: Makowecki. B. (2012). Comparison of hydraulic fracture fluids in multi-stage
fracture stimulated horizontal wells in the Pembina Cardium formation. In 2012 SPE
Hydrocarbon Economics and Evaluation Symposium, 24-25 September, 2012, Calgary,
Alberta, Canada: Society of Petroleum Engineers. http: //dx.doi.org/10.2118 /162916-MS
Knappe. D: Fireline. TL. (2012). Fracking 101: Shale gas extraction using horizontal drilling and
hydraulic fracturing. Presented atNCAWWA-WEA Annual Conference, November 14, 2012,
Raleigh, N C.
Konikow. LF. (2013). Groundwater depletion in the United States (1900-2008). U.S. Geological
Survey Scientific Investigations Report 2013-5079. Reston, VA: U.S. Geological Survey.
http://pubs.usgs.gOv/sir/2013/5079
11-32
-------
Chapter 11 - References
Konikow. LF: Kendv. E. (2005). Groundwater depletion: A global problem. Hydrogeo J 13: 317-320.
http://dx.doi.Org/10.1007/sl0040-004-0411-8
Konschnik. K: Davalu. A. (2016). Hydraulic fracturing chemicals reporting: Analysis of available
data and recommendations for policymakers. Energy Policy 88: 504-514.
http://dx.doi.Org/10.1016/i.enpol.2015.ll.002
Korfmacher. KS: Tones. WA: Malone. SL: Vinci. LF. (2013). Public health and high volume hydraulic
fracturing. New Solutions: A Journal of Environmental and Occupational Health Policy 23:
13-31. http://dx.doi.org/10.2190 /NS.23.1 .c
Krasner. SW. (2009). The formation and control of emerging disinfection by-products of health
concern [Review], Philos Transact A Math Phys Eng Sci 367: 4077-4095.
https://dx.doi.Org/10.1098/rsta.2009.0108
Kresse. TM: Warner. NR: Hays. PD: Down. A: Vengosh. A: Tackson. RB. (2012). Shallow groundwater
quality and geochemistry in the Fayetteville Shale gas-production area, north-central
Arkansas, 2011. (2012-5273). Denver, CO: U.S. Geological Survey.
http://pubs.usgs.gOv/sir/2012/5273/sir2012-5273.pdf
Kundert. D: Mullen. M. (2009). Proper evaluation of shale gas reservoirs leads to a more effective
hydraulic-fracture stimulation. In SPE Rocky Mountain Petroleum Technology Conference,
14-16 April, 2009, Denver, CO: Society of Petroleum Engineers.
http: //dx.doi.org/10.2118 /123586-MS
Kurose. S. (2014). Requiring the use of tracers in hydraulic fracturing fluid to trace alleged
contamination. Sustainable Development Law & Policy 14: 43-54.
Kuwavama. Y: Olmstead. S: Krupnick. A. (2015a). Water quality and quantity impacts of hydraulic
fracturing. Current Sustainable/Renewable Energy Reports 2: 17-24.
http: //dx.doi.org/10.1007/s40518-014-0023-4
Kuwavama. Y: Roeshot. S: Krupnick. A: Richardson. N: Mares. 1. (2015b). Pits versus tanks: Risks
and mitigation options for on-site storage of wastewater from shale gas and tight oil
development (RFF DP 15-53). Washington, D.C.: Resources for the Future.
http://www.rff.org/research/publications/pits-versus-tanks-risks-and-mitigation-options-
site-storage-wastewater-shale
LA GWRC (Louisiana Ground Water Resources Commission). (2012). Managing Louisiana's
groundwater resources: An interim report to the Louisiana Legislature. Baton Rouge, LA:
Louisiana Department of Natural Resources.
http://dnr.louisiana.gov/index.cfm?md=pagebuilder&tmp=home&pid=907
Lacazette. A: Geiser. P. (2013). Comment on Davies etal., 2012 Hydraulic fractures: How far can
they go? Marine and Petroleum Geology 43: 516-518.
http://dx.doi.Org/10.1016/i.marpetgeo.2012.12.008
Landis. MS: Kamal. AS: Kovalcik. KD: Croghan. C: Norris. GA: Bergdale. A. (2016). The impact of
commercially treated oil and gas produced water discharges on bromide concentrations
and modeled brominated trihalomethane disinfection byproducts at two downstream
municipal drinking water plants in the upper Allegheny River, Pennsylvania, USA. Sci Total
Environ 542: 505-520. http://dx.doi.org/10.1016/i.scitotenv.2015.10.074
Landry. G: Weltv. RD: Thomas. M: Vaughan. ML: Tatum. D. (2015). Bridging the gap: An integrated
approach to solving sustained casing pressure in the Cana Woodford Shale. SPE Well
Integrity Symposium, June 2-3, 2015, Galveston, TX.
https://www.onepetro.org/conference-paper/SPE-174525-MS
11-33
-------
Chapter 11 - References
Lange. T: Sauter. M: Heitfeld. M: Schetelig. K: Brosig. K: Tahnke. W: Kissinger. A: Helmig. R: Ebigbo. A:
Class. H. (2013). Hydraulic fracturing in unconventional gas reservoirs: Risks in the
geological system part 1. Environ Earth Sci 70: 3839-3853.
http://dx.doi.org/10.1007/sl266S-013-2803-3
Langmuir. D: Riese. AC. (1985). The thermodynamic properties of radium. Geochim Cosmo Act 49:
1593-1601.
Lauer. NE: Harkness. IS: Vengosh. A. (2016). Brine spills associated with unconventional oil
development in North Dakota. Environ SciTechnol 50: 5389-5397.
http://dx.doi.org/10.1021/acs.est5b06349
Laurenzi. IT: Tersev. GR. (2013). Life cycle greenhouse gas emissions and freshwater consumption of
Marcellus shale gas. Environ Sci Technol 47: 4896-4903.
http://dx.doi.org/10.1021/es305162w
Lawal. H: Abolo. NU: Tackson. G: Sahai. V: Flores. C. (2014). A quantitative approach to analyze
fracture area loss in shale gas reservoirs. In SPE Latin America and Caribbean Petroleum
Engineering Conference, May 21-23, 2014, Maracaibo, Venezuela.
https://www.onepetro.org/conference-paper/SPE-169406-MS
LDEO (Louisiana Department of Environmental Quality). (2008). Ground water use advisory:
Commissioner of conservation recommends wise water use planning in the Haynesville
Shale. http://dnr.louisiana.gov/index.cfm?md=newsroom&tmp=detail&aid=509
LDEO (Louisiana Department of Environmental Quality). (2013). Personal communication, M.
Molieri, January 3, 2013.
Le Roux. 1: Gallard. H: Croue. TP. (2012). Formation of NDMA and halogenated DBPs by
chloramination of tertiary amines: The influence of bromide ion. Environ Sci Technol 46:
1581-1589. http://dx.doi.org/10.1021/es203785s
Leadscope Inc. (2012). Leadscope [Computer Program], Columbus, OH. http://www.leadscope.com
L'Eau LLC. (2008). Dewvaporation desalination 5,000-gallon-per-day pilot plant. (Desalination and
Water Purification Research and Development Program Report No. 120). Denver, CO:
Bureau of Reclamation, U.S. Department of the Interior.
http://www.usbr.gov/research/AWT/reportpdfs/reportl20.pdf
LeBas. R: Lord. P: Luna. D: Shahan. T. (2013). Development and use of high-TDS recycled produced
water for crosslinked-gel-based hydraulic fracturing. In 2013 SPE Hydraulic Fracturing
Technology Conference, 4-6 February, 2013, The Woodlands, TX: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/163824-MS
Lecampion. B: Teffrev. R: Detournav. E. (2005). Resolving the geometry of hydraulic fractures from
tilt measurements. Pure Appl Geophys 162: 2433-2452. http: //dx.doi.org/10.1007/s00024-
005-2786-4
Lee. PS: Herman. TP: Elsworth. D: Kim. HT: Lee. HS. (2011). A critical evaluation of unconventional
gas recovery from the marcellus shale, northeastern United States. KSCE Journal of Civil
Engineering 15: 679-687. http://dx.doi.org/10.1007/sl2205-011-00Q8-4
Lee. HP: Olson. I. onE: Holder. 1. on: Gale. TFW: Myers. RP. (2015). The interaction of propagating
opening mode fractures with preexisting discontinuities in shale. J Geophys Res Solid Earth
120: 169-181. http://dx.doi.org/10.1002/2014TB011358
Lee. K: Neff. I. (2011). Produced water: Environmental risks and advances in mitigation
technologies. New York, NY: Springer, http: //dx.doi.org/10.1007/978-l-4614-0046-2
11-34
-------
Chapter 11 - References
Lefebvre. 0: Moletta. R. (2006). Treatment of organic pollution in industrial saline wastewater: A
literature review [Review], Water Res 40: 3671-3682.
http://dx.doi.Org/10.1016/i.watres.2006.08.027
Lefebvre. R: Gleeson. T. om: Mckenzie. TM: Gassiat. C. f2015I Reply to comment by Flewelling and
Sharma on "Hydraulic fracturing in faulted sedimentary basins: Numerical simulation of
potential contamination of shallow aquifers over long time scales". Water Resour Res 51:
1877-1882. http://dx.doi.org/10.1002/2014WR016698
LePage. 1: De Wolf. C: Bemelaar. 1: Nasr-El-din. HA. (2013). An environmentally friendly stimulation
fluid for high-temperature applications. SPE Journal 16: 104-110.
http://dx.d0i.0rg/l 0.2118/121709-PA
Lester. Y: Ferrer. I: Thurman. EM: Sitterlev. KA: Korak. TA: Aiken. G: Linden. KG. (2015).
Characterization of hydraulic fracturing flowback water in Colorado: Implications for water
treatment. Sci Total Environ 512-513: 637-644.
http ://dx. doi. or g/10.1016/i. scitotenv. 2015.01.043
Levick. L: Fonseca. 1: Goodrich. D: Hernandez. M: Semmens. DT: Leidv. R: Scianni. M: Guertin. DP:
Stromberg. 1: Tluczek. M: Kepner. WG. (2008). The ecological and hydrological significance
of ephemeral and intermittent streams in the arid and semi-arid American Southwest [EPA
Report], (EPA/600/R-08/134 and ARS/233046). Washington, D.C.: U.S. Environmental
Protection Agency, Office of Research and Development and U.S. Department of
Agriculture/Agricultural Research Service. Southwest Watershed Research Center.
https://www.epa.gov/sites/production/files/2015-
03/documents/ephemeral streams report final 508-kepner.pdf
Levis. E. (2011). Texas company pays $93,710 settlement for polluting Clearfield County Creek.
https://fishandboat.com/newsreleases/news2011.htm
Levorsen. AI. (1965). Geology of Petroleum (2nd ed.). San Francisco, CA: W.H. Freeman and
Company.
Li. H: Carlson. KH. (2014). Distribution and origin of groundwater methane in the Wattenberg oil
and gas field of northern Colorado. Environ Sci Technol 48: 1484-1491.
http://dx.doi.org/10.1021/es404668b
Li. H: Son. TH: Carlson. KH. (2016a). Concurrence of aqueous and gas phase contamination of
groundwater in the Wattenberg oil and gas field of northern Colorado. Water Res 88: 458-
466. http://dx.doi.Org/10.1016/i.watres.2015.10.031
Li. H: Zou. Y: Valko. PP: Ehlig-Economides. C. (2016b). Hydraulic fracture height predictions in
laminated shale formations using finite element discrete element method. In SPE Hydraulic
Fracturing Technology Conference, 9-11 February, The Woodlands, Texas: Society of
Petroleum Engineers. http://dx.doi.org/10.2118/179129-MS
Li. L: Xia. Y: Huang. B. o: Zhang. L: Li. M: Li. A. (2016c). The behaviour of fracture growth in
sedimentary rocks: A numerical study based on hydraulic fracturing processes. Energies 9.
http://dx.doi.org/10.3390/en9030169
Linaric. M: Markic. M: Sipos. L. (2013). High salinity wastewater treatment Water Sci Technol 68:
1400-1405. http://dx.d0i.0rg/l0.2166/wst.2013.376
Linkov. I: Welle. P: Lonev. D: Tkachuk. A: Canis. L: Kim. IB: Bridges. T. (2011). Use of multicriteria
decision analysis to support weight of evidence evaluation. Risk Anal 31: 1211-1225.
http://dx.doi.Org/10.llll/i.1539-6924.2011.01585.x
11-35
-------
Chapter 11 - References
Lira-Barragan. LF: Ponce-Ortega. TM: Serna-Gonzalez. M: El-Halwagi. MM. (2016). Optimal reuse of
flowback wastewater in hydraulic fracturing including seasonal and environmental
constraints. AIChE J 62: 1634-1645. http://dx.doi.org/10.10Q2/aic.15167
Liu. 1: Zhang. F: Gardner. RP: Hou. G: Zhang. 0: Li. H. (2015). A method to evaluate hydraulic fracture
using proppant detection. Appl Radiat Isot 105: 139-143.
http://dx.doi.Org/10.1016/i.apradiso.2015.08.003
Llewellyn. GT. (2014). Evidence and mechanisms for Appalachian Basin brine migration to shallow
aquifers in NE Pennsylvania, USA. Hydrogeo J 22: 1055-1066.
http://dx.doi. org/10.1007/sl 0040-014-1125-1
Llewellyn. GT: Dorman. F: Westland. TL: Yoxtheimer. D: Grieve. P: Sowers. T: Humston-Fulmer. E:
Brantley. SL. (2015). Evaluating a groundwater supply contamination incident attributed to
Marcellus Shale gas development Proc Natl Acad Sci USA 112: 6325-6330.
http://dx.doi.org/10.1073/pnas.1420279112
Loveless. D: Holtsclaw. 1: Saini. R: Harris. PC: Fleming. 1. (2011). Fracturing fluid comprised of
components sourced solely from the food industry provides superior proppant transport In
2011 SPE Annual Technical Conference and Exhibition. Richardson, TX: Society of
Petroleum Engineers. http://dx.doi.org/10.2118/147206-MS
Lowe. T: Potts. M: Wood. D. (2013). A case history of comprehensive hydraulic fracturing
monitoring in the Cana Woodford. In 2013 SPE Annual Technical Conference and Exhibition,
30 September-2 October, 2013, New Orleans, LA: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/166295-MS
Luh. I: Marinas. BT. (2012). Bromide ion effect on N-nitrosodimethylamine formation by
monochloramine. Environ Sci Technol 46: 5085-5092.
http://dx.doi.org/10.1021/es300Q77x
Lutz. BP: Lewis. AN: Doyle. MW. (2013). Generation, transport, and disposal of wastewater
associated with Marcellus Shale gas development. Water Resour Res 49: 647-656.
Ma. G: Geza. M: Xu. P. (2014). Review of flowback and produced water management, treatment, and
beneficial use for major shale gas development basins. Shale Energy Engineering
Conference 2014, Pittsburgh, Pennsylvania, United States.
MacDonald. RT: Frantz. TH: Schlotterbeck. ST: Adams. B: Sikorski. D. (2003). An update of recent
production responses obtained from Devonian shale and Berea wells stimulated with
nitrogen foam (with proppant) vs. nitrogen-only, Pike Co., KY. Presented at SPE Eastern
Regional Meeting, 6-10 September, 2003, Pittsburgh, PA: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/84834-MS
Maguire-Bovle. ST: Barron. AR. (2014). Organic compounds in produced waters from shale gas
wells. Environ Sci Process Impacts 16: 2237-2248. http://dx.doi.org/10.1039/c4em00376d
Malone. M: Ely. IW. (2007). Execution of hydraulic fracturing treatments. In M Economides; T
Martin (Eds.), Modern Fracturing: Enhancing Natural Gas Production (pp. 323-360).
Houston, TX: ET Publishing.
Malone. S: Kelso. M: Auch. T: Edelstein. K: Ferrar. K: Talbert. K. (2015). Data inconsistencies from
states with unconventional oil and gas activity. J Environ Sci Health Part A Tox Hazard Subst
Environ Eng 50: 501-510. http://dx.doi.org/10.1080/10934529.2015.992678
Malonev. KO: Yoxtheimer. DA. (2012). Production and disposal of waste materials from gas and oil
extraction from the Marcellus shale play in Pennsylvania. Environmental Practice 14: 278-
287. https://pubs.er.usgs.gOv/publication/70043410
11-36
-------
Chapter 11 - References
Mantell. ME. (2011). Produced water reuse and recycling challenges and opportunities across major
shale plays. Presented at Hydraulic Fracturing Study: March 2011, technical workshop #4
on Water Resources Management, March 29-30, 2011, Arlington, VA.
http://www2.epa.gov/sites/production/files/documents/09 Mantell - Reuse 508.pdf
Mantell. ME. (2013). Recycling and reuse of produced water to reduce freshwater use in hydraulic
fracturing operations. In Summary of the Technical Workshop on Water Acquisition
Modeling: Assessing Impacts through Modeling and Other Means (pp. A20-A27).
Washington, D.C.: U.S. Environmental Protection Agency.
http://www2.epa.gov/hfstudv/summarv-technical-workshop-water-acquisition-modeling-
assessing-impacts-through-modeling-and
Manthos. D. (2014). SkyTruth releases map of drilling-related impoundments across PA.
http://skytruth.org/2014/10/PA-drilling-impoundments-2005-2Q13 / (accessed
September 9, 2016).
Marshak. S. (2004). Essentials of Geology (1st ed.). New York, NY: W.W. Norton and Company.
Martin. T: Valko. P. (2007). Hydraulic fracture design for production enhancement. In M
Economides; T Martin (Eds.), Modern Fracturing: Enhancing Natural Gas Production.
Houston, TX: ET Publishing.
Matamoros. V: Muieriego. R: Bavona. TM. (2007). Trihalomethane occurrence in chlorinated
reclaimed water at full-scale wastewater treatment plants in NE Spain. Water Res 41: 3337-
3344. http://dx.doi.Org/10.1016/i.watres.2007.04.021
Material Safety Data Sheets, (a) Encana/Halliburton Energy Services, Inc.: Duncan, Oklahoma.
Provided by Halliburton Energy Services during an onsite visit by the EPA on May 10, 2010;
(b) Encana Oil and Gas (USA), Inc.: Denver, Colorado. Provided to US EPA Region 8.
Matthews. TC: Li. S: Swann. CT: Ericksen. RL. (2006). Incubation with moist top soils enhances
solubilization of radium and other components from oilfield scale and sludge:
Environmental concerns from Mississippi. Environmental Geosciences 13: 43-53.
Maupin. MA: Kenny. IF: Hutson. SS: Lovelace. IK: Barber. NL: Linsev. KS. (2014). Estimated use of
water in the United States in 2010. (USGS Circular 1405). Reston, VA: U.S. Geological Survey.
http://dx.doi.org/10.3133/cirl405
Mauter. M: Palmer. V. (2014). Expert elicitation of trends in marcellus oil and gas wastewater
management. J Environ Eng 140: B4014004. http://dx.doi.org/10.1061/(ASCE)EE. 1943-
7870.0000811
Maxwell. SL: Culligan. BK: Warren. RA: McAlister. DR. (2016). Rapid method for the determination
of 226Ra in hydraulic fracturing wastewater samples. Journal of Radioanal Chem 309:
1333-1340. http://dx.doi.org/10.1007/sl 0967-016-4745-1
McCartney. H. (2011). Hydraulic fracturing in coal bed methane development, Raton Basin,
Southern Colorado, USA. In Proceedings of the Technical Workshops for the Hydraulic
Fracturing Study: Well Construction & Operations (pp. 69-70). (EPA 600/R-11/046).
Washington, D.C.: U.S. Environmental Protection Agency.
https://www.epa.gov/sites/production/files/2015-
10/documents/hydraulicfracturingstudywellconstructionandoperation.pdf
McDaniel. 1: Watters. L: Shadravan. A. (2014). Cement sheath durability: Increasing cement sheath
integrity to reduce gas migration in the Marcellus Shale Play. In SPE Hydraulic Fracturing
Technology Conference proceedings, 4-6 February, 2014, The Woodlands, TX: Society of
Petroleum Engineers. http://dx.doi.org/10.2118/168650-MS
11-37
-------
Chapter 11 - References
McDaniel. RR: McCarthy. SM: Smith. M. (2010). Methods and compositions for determination of
fracture geometry in subterranean formations. (U.S. Patent No. 7,726,397 B2). Washington,
D.C.: U.S. Patent and Trademark Office. https: //www.google.com/patents/US7726397
McElreath. D. (2011). Comparison of hydraulic fracturing fluids composition with produced
formation water following fracturing: Implications for fate and transport Presented at
Hydraulic Fracturing Study: March 2011 Technical Workshop on Fate and Transport, March
28-29, 2011, Arlington, VA.
http://www2.epa.gov/sites/production/files/documents/comparisonofhffluidscompositio
nwithproducedformationwater.pdf
McGuire. Ml: Karanfil. T: Krasner. SW: Reckhow. DA: Roberson. TA: Summers. RS: Westerhoff. P: Xie.
Y. (2014). Notyour granddad's disinfection by-product problems and solutions. JAWWA
106: 54-73. http://dx.doi.org/10.5942/iawwa.2014.106.0128
Mcintosh. TC: Walter. LM. (2005). Volumetrically significant recharge of Pleistocene glacial
meltwaters into epicratonic basins: Constraints imposed by solute mass balances. Chem
Geol 222: 292-309. http://dx.doi.Org/10.1016/i.chemgeo.2005.07.010
Mcintosh. TC: Walter. LM: Martini. AM. (2002). Pleistocene recharge to midcontinent basins: Effects
on salinity structure and microbial gas generation. Geochim Cosmo Act 66: 1681-1700.
http://dx.doi. org/10.1016/S0016-7037C01100885-7
McKay. SF: King. AT. (2006). Potential ecological effects of water extraction in small, unregulated
streams. River Research and Applications 22: 1023-1037.
http: / /dx. doi. or g/10.10 0 2 /rra. 958
Mckenzie. LM: Guo. R: Witter. RZ: Savitz. DA: Newman. L: Adgate. TL. (2014). Birth outcomes and
maternal residential proximity to natural gas development in rural Colorado. Environ
Health Perspect 122: 412-417. http://dx.doi.org/10.1289/ehp.1306722
McMahon. PB: Dennehv. KF: Bruce. BW: Gurdak. IT: Oi. SL. (2007). Water-quality assessment of the
high plains aquifer, 1999-2004. (Professional Paper 1749). Reston, VA: U.S. Geological
Survey, http: / /pubs.usgs. gov/pp /1749 /
MCOR (Marcellus Center for Outreach and Research). (2012). Extent and thickness of Marcellus
Shale. University Park, PA: Pennsylvania State University.
http://www.marcellus.psu.edu/images/Marcellus thickness.gif
McPhillips. LE: Creamer. AE: Rahm. BG: Walter. MT. (2014). Assessing dissolved methane patterns
in central New York groundwater. 1: 57-73. http: //dx.doi.Org/10.1016/j.ejrh.2014.06.002
Meixner. T: Manning. AH: Stonestrom. DA: Allen. DM: Aiami. H: Blasch. KW: Brookfield. AE: Castro.
CL: Clark. IF: Gochis. DT: Flint. AL: Neff. KL: Niraula. R: Rodell. M: Scanlon. BR: Singha. K:
Walvoord. MA. (2016). Implications of projected climate change for groundwater recharge
in the western United States. J Hydrol 534: 124-138.
http: / /dx. doi. or g/10.1016 /i .ihydrol. 2 015.12.027
Melillo. TM: Richmond. TC: Yohe. GW. (2014). Ch. 20: Southwest. Climate change impacts in the
United States. In JM Melillo; TC Richmond; GW Yohe (Eds.), Climate Change Impacts in the
United States: The Third National Climate Assessment (pp. 462-486). Washington, D.C.: U.S.
Global Change Research Program. http://dx.doi.org/10.7930/TQ8G8HMN
Methanol Institute. (2013). Methanol safe handling manual. Alexandria, VA.
Michie. TW: Koch. CA. (1991). Evaluation of injection-well risk management in the Williston Basin. J
Pet Tech 43: 737-741. http: //dx.doi.org/10.2118/20693-PA
11-38
-------
Chapter 11 - References
Minnich. K. (2011). A water chemistry perspective on flowback reuse with several case studies. In
Proceedings of the Technical Workshops for the Hydraulic Fracturing Study: Water
Resources Management
https://www.epa.gov/sites/production/files/documents/10 Minnich - Chemistry 508.pdf
Miskimins. TL. (2008). Design and life cycle considerations for unconventional reservoir wells. In
2008 SPE Unconventional Reservoirs Conference, 10-12 February, 2008, Keystone, CO:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/114170-MS
Mitchell. BT. (1970) Viscosity of foam. (Doctoral Dissertation). The University of Oklahoma, Norman,
OK.
Mitchell. AL: Small. M: Casman. EA. (2013a). Surface water withdrawals for Marcellus Shale gas
development: Performance of alternative regulatory approaches in the Upper Ohio River
Basin. Environ Sci Technol 47: 12669-12678. http://dx.doi.org/10.1021/es403537z
Mitchell. I: Pabon. P: Collier. ZA: Egeghv. PP: Cohen-Hubal. E: Linkov. I: Vallero. DA. (2013b). A
decision analytic approach to exposure-based chemical prioritization. PLoS ONE 8: e70911.
http://dx.doi.org/10.1371/iournal.pone.0070911
Molofskv. LI: Connor. TA: Farhat. SK: Wvlie. AS: Wagner. T. (2011). Methane in Pennsylvania water
wells unrelated to Marcellus shale fracturing. Oil and Gas Journal 109: 54-59.
Molofskv. LT: Connor. TA: McHugh. TE: Richardson. SD: Woroszylo. C: Alvarez. PI. (In Press)
Environmental factors associated with natural methane occurrence in the Appalachian
Basin. Ground Water, http: //dx.d0i.0rg/l0. Ill 1 /gwat. 12401
Molofskv. LT: Connor. TA: Wvlie. AS: Wagner. T: Farhat. SK. (2013). Evaluation of methane sources in
groundwater in northeastern Pennsylvania. Ground Water 51: 333-349.
http: / /dx. doi. or g/10.1111 /gwat. 12056
Montague. TA: Pinder. GF. (2015). Potential of hydraulically induced fractures to communicate with
existing wellbores. Water Resour Res 51: 8303-8315.
http: / /dx. doi.org/10.1002/2014WR016771
Montana GWIC (Ground Water Information Center). (2009). Coal bed methane Powder River Basin:
Montana Bureau of Mines and Geology Ground Water Information Center. Available online
at http://mbmggwic.mtech.edu/
Montgomery. C. (2013). Fracturing fluid components. In A Bunder; J McLennon; R Jeffrey (Eds.),
Effective and Sustainable Hydraulic Fracturing. Croatia: InTech.
http://dx.doi.org/10.5772/56422
Montgomery. CT: Smith. MB. (2010). Hydraulic fracturing - History of an enduring technology. J Pet
Tech 62: 26-32.
Morillon. A: Vidalie. IF: Svahnudi. U: Suripno. S: Hadinoto. EK. (2002). Drilling and waste
management; SPE 73931. Presented at the SPE International Conference on Health, Safety,
and Environment in Oil and Gas Exploration and Production, March 20-22, 2002, Kuala
Lumpur, Malaysia.
MTA. Inc.. (Moore Twining Associates, Inc.). (2014). Lab report with cover letter dated 06/03/2014.
Fresno, CA: Central Valley Regional Water Quality Control Board (CVRWQCB).
Muehlenbachs. L: Spiller. E: Timmins. C. (2012). Shale gas development and property values:
Differences across drinking water sources. (NBER Working Paper No. 18390). Cambridge,
MA: National Bureau of Economic Research, http: //www.nber.org/papers/wl8390
11-39
-------
Chapter 11 - References
Mukheriee. H: Poe jr.. B: Heidt. T: Watson. T: Barree. R. (2000). Effect of pressure depletion on
fracture-geometry evolution and production performance. SPE ProdFacil 15: 144-150.
http://dx.doi.org/10.2118/65Q64-PA
Murali Mohan. A: Hartsock. A: Hammack. RW: Vidic. RD: Gregory. KB. (2013). Microbial
communities in flowback water impoundments from hydraulic fracturing for recovery of
shale gas. FEMS Microbiol Ecol. http://dx.doi.org/10.llll/1574-6941.12183
Murdoch. PS: Baron. IS: Miller. TL. (2000). Potential effects of climate chance on surface-water
quality in North America. J Am Water Resour Assoc 36: 347-366.
Murray. KE. (2013). State-scale perspective on water use and production associated with oil and
gas operations, Oklahoma, U.S. Environ Sci Technol 47: 4918-4925.
http://dx.doi.org/10.1021/es400Q593
Mvshakin. E: Siriwardane. H: Hulcher. C: Lindner. E: Sams. N: King. S: Mckov. M. (2015). Numerical
simulations of vertical growth of hydraulic fractures and brine migration in geological
formations above the Marcellus shale. Journal of Natural Gas Science & Engineering 27: 531-
544. http://dx.doi.Org/10.1016/i.ingse.2015.08.030
Narasimhan. S: McClearv. M: Fluckiger. SD: Gray. 1: Cherian. BY: Shaikh. H: Olaove. 0. (2015). Using
the wrong method to estimate stress from depletion causes significant errors in predicting
wellbore integrity and fracture geometry. In SPE Eastern Regional Meeting, 13-15 October,
2015, Morgantown, WV: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/177310-MS
National Drought Mitigation Center (2015). U.S. drought monitor. Available online at
http://droughtmonitor.unl.edu/Home.aspx (accessed February 27, 2015).
Nelson. AW: Eitrheim. ES: Knight. AW: May. D: Mehrhoff. MA: Shannon. R: Litman. R: Burnett. WC:
Forbes. TZ: Schultz. MK. (2015). Understanding the radioactive ingrowth and decay of
naturally occurring radioactive materials in the environment: An analysis of produced fluids
from the Marcellus Shale. Environ Health Perspect 123: 689-696.
http://dx.doi.org/10.1289/ehp.1408855
Nelson. AW: May. D: Knight. AW: Eitrheim. ES: Mehrhoff. M: Shannon. R: Litman. R: Schultz. MK.
(2014). Matrix complications in the determination of radium levels in hydraulic fracturing
flowback water from Marcellus Shale. Environ Sci Technol Lett 1: 204-208.
http://dx.doi.org/10.1021/ez500Q379
NETL (National Energy Technology Laboratory). (2013). Modern shale gas development in the
United States: An update. Pittsburgh, PA: U.S. Department of Energy, National Energy
Technology Laboratory. http: //www.netl.doe.gov/File%20Library/Research/Oil-
Gas /shale- gas-primer-update- 2 013 .pdf
NETL (National Energy Technology Laboratory). (2014). Evaporation [Fact Sheet], Pittsburgh, PA:
US Department of Energy, National Energy Technology Laboratory.
http://www.netl.doe.gov/research/coal/crosscutting/pwmis/tech-desc/evap
Newell. R. (2011). Shale gas and the outlook for U.S. natural gas markets and global gas resources.
Presented at US EIA presentation at OECD Meetings, June 21, 2011, Paris, France.
http://photos.state.gov/libraries/usoecd/19452/pdfs/DrNewell-EIA-Administrator-Shale-
Gas-Presentation-lune212011.pdf
Nicot. TP: Hebel. AK: Ritter. SM: Walden. S: Baier. R: Galuskv. P: Beach. 1: Kyle. R: Svmank. L: Breton.
C (2011). Current and projected water use in the Texas mining and oil and gas industry -
Final Report. (TWDB Contract No. 0904830939).
11-40
-------
Chapter 11 - References
http://www.twdb.texas.gov/publications/reports/contracted reports/doc/0904830939 M
iningWaterUse.pdf
Nicot. TP: Reedy. RC: Costlev. RA: Huang. Y. (2012). Oil &gas water use in Texas: Update to the 2011
mining water use report Austin, TX: Bureau of Economic Geology, University of Texas at
Austin.
http://www.twdb.texas.gov/publications/reports/contracted reports/doc/0904830939 2
012Update MiningWaterUse.pdf
Nicot. TP: Scanlon. BR. (2012). Water use for shale-gas production in Texas. U.S. Environ Sci Technol
46: 3580-3586. http://dx.doi.org/10.1021/es204602t
Nicot. TP: Scanlon. BR: Reedy. RC: Costlev. RA. (2014). Source and fate of hydraulic fracturing water
in the Barnett Shale: A historical perspective. Environ Sci Technol 48: 2464-2471.
http://dx.doi.org/10.1021/es404Q50r
NIOSH (National Institute for Occupational Safety and Health). (2000). International chemical
safety card (ICSC) #0291 [Fact Sheet], Atlanta, GA: Center for Disease Control and
Prevention, http://www.cdc.gov/niosh/ipcsneng/neng0291.html
NM EMNRD OCD (New Mexico Energy, Minerals and Natural Resource Department, Oil
Conservation Division). (2013). Attachment A to June 6, 2013 Order and statementof
reasons [Order No. R-13506-D], New Mexico Commission of Public Records.
https://www.nmoga.org/wp-content/uploads/2011/06/New-Pit-Rule.pdf
Nolen-Hoeksema. R. (2013). Elements of hydraulic fracturing [Editorial], Oilfield Rev 25: 2.
North Dakota Department of Health. (2011). Incidents reported to the Oil & Gas Division from
11/1/2012-11/11/2013.
http://www.ndhealth.gov/ehs/foia/spills/ChartWebPageOG 20121101 20131111.pdf
North Dakota Department of Health. (2015). Oil field environmental incident summary, incident
20150107160242.
http://www.ndhealth.gov/EHS/FOIA/Spills/Summary Reports/20150107160242 Summa
rv Reportpdf
North Dakota State Water Commission. (2010). Water appropriation requirements, current water
use, & water availability for energy industries in North Dakota: A 2010 summary. Bismarck,
ND.
http://www.swc.nd.gov/info edu/reports and publications/pdfs/wr investigations/wr49
reportpdf
North Dakota State Water Commission. (2014). Facts about North Dakota fracking and water use.
Bismarck, ND. http://www.swc.nd.gov/pdfs/fracking water use.pdf
Nowak. N: Bradish. 1. (2010). High density polyethylene (HDPE) lined produced water evaporation
ponds. Presented at 17th International Petroleum and Biofuels Environmental Conference,
August 31-September 2, 2010, San Antonio, TX.
http://ipec.utulsa.edu/Conf2010/Powerpoint%20presentations%20and%20papers%20re
ceived/Nowak 83 received9-8-10.pdf
Nowamooz. A: Lemieux. TM: Molson. 1: Therrien. R. (2015). Numerical investigation of methane and
formation fluid leakage along the casing of a decommissioned shale gas well. Water Resour
Res 51: 4592-4622. http://dx.doi.org/10.1002/2014WR016146
NPC (National Petroleum Council). (2011a). Management of produced water from oil and gas wells.
(Paper #2-17). Washington, D.C. http://www.npc.org/Prudent Development-
Topic Papers/2-17 Management of Produced Water Paper.pdf
11-41
-------
Chapter 11 - References
NPC (National Petroleum Council). (2011b). Plugging and abandonment of oil and gas wells. (Paper
#2-25). Washington, D.C.
NRC (National Research Council). (2001). Energy research at DOE: Was it worth it? Energy
efficiency and fossil energy research 1978 to 2000. Committee on Benefits of DOE R&D on
Energy Efficiency and Fossil Energy.Washington, D.C.: The National Academies Press.
NRC (National Research Council). (2010). Management and effects of coalbed methane produced
water in the western United States. Washington, D.C.: The National Academies Press.
NRC (National Research Council). (2012). Water reuse: Potential for expanding the nations water
supply through reuse of municipal wastewater. Committee on the Assessment of Water
Reuse as an Approach for Meeting Future Water Supply Need. Washington, D.C.: The
National Academies Press.
NRDC (National Resources Defense Council). (2015). Petition for Review: Before the Environmental
Appeals Board, United States Environmental Protection Agency, Washington, D.C. In re:
Phoenix Production Company, WESCO Operating, Inc. NPDES Permit Nos. WY-0024945
(Rolff Lake Unit), WY-0024953 (Sheldon Dome Field), and WY-0025232 (Tensleep #1
[Winkleman Dome Field]), Wind River. Washington, DC and Chicago, IL: Attorneys for
Natural Resources Defense Council, Inc.
https://yosemite.epa.gov/oa/EAB Web Docket.nsf/fFilingsVBFECED6BDC00B0AB85257E
2800480CEC/$File/2015-04-14%20NRDC%20Petition%20for%2 0Review.pdf
NSTC (National Science and Technology Council). (2000). Integrated assessment of hypoxia in the
Northern Gulf of Mexico. Washington, D.C.: U.S. National Science and Technology Council,
Committee on Environment and Natural Resources.
http://oceanservice.noaa.gov/products/hypox final.pdf
NYSDEC (New York State Department of Environmental Conservation). (2011). Revised draft
supplemental generic environmental impact statement (SGEIS) on the oil, gas and solution
mining regulatory program: Well permit issuance for horizontal drilling and high-volume
hydraulic fracturing to develop the Marcellus shale and other low-permeability gas
reservoirs. Albany, NY. http: //www.dec.nv.gOv/energv/7537O.html
NYSDEC (New York State Department of Environmental Conservation). (2015). Final supplemental
generic environmental impact statement Albany, NY.
http://www.dec.ny.gov/energy/75370.html
O'Mallev. D: Karra. S: Currier. RP: Makedonska. N: Hvman. ID: Viswanathan. HS. (2015). Where does
water go during hydraulic fracturing? Ground water.
http: / /dx. doi.org/10.1111 /gwat. 12380
Obolenskv. A: Singer. PC. (2008). Development and interpretation of disinfection byproduct
formation models using the Information Collection Rule database. Environ Sci Technol 42:
5654-5660. http://dx.doi.org/10.1021/es702974f
OCC (Oklahoma Corporation Commission). (2013). Personal communication, Baker, T. April 11,
2013.
OCC OGCD (Oklahoma Corporation Commission Oil and Gas Conservation Division). (2015) Title
165: Corporation commission: Chapter 10: Oil and gas conservation. Oklahoma City, OK.
http://www.occeweb.eom/rules/CH10eff08-27-15searchable.pdf
OCC OGCD (Oklahoma Corporation Commission, Oil and Gas Conservation Division). (2016). Media
advisory - Regional earthquake response plan for central Oklahoma and expansion of the
11-42
-------
Chapter 11 - References
are of interest. https: //www.occeweb.eom/News/2016/03-07-16ADVISORY-
A0I.%20V0LUME%20REDUCTI0N.pdf
Odegard. M: Daniels. T: Huck. T: Forster. T. (2015). Active geosteering strategies in structurally
complex Niobrara unconventional reservoirs. Presented at AAPG Annual Convention &
Exhibition, May 31-June 3, 2015, Denver, CO.
http://www.searchanddiscovery.eom/pdfz/documents/2015/80476odegard/ndx odegard
.pdf.html
ODNR. DMRM (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. Columbus, OH.
http://oilandgas.ohiodnr.gov/portals/oilgas/pdf/bainbridge/report.pdf
OEHHA (Office of Environmental Health Hazard Assessment). (2012). Title 27, California Code of
Regulations Article 8. No Observable Effect Levels.
http://oehha.ca.gOv/media/downloads/crnr/regulation022610.pdf
Ohio EPA (Ohio Environmental Protection Agency). (2012a). Considerations for public water
systems prior to providing raw or treated water to oil and natural gas companies.
http://www.epa.state.oh.us/Portals/0/general%20pdfs/Considerations%20for%20Public
%20Water%20Systems%20Prior%20to%20Providing%20Raw%20or%20Treated%20Wa
ter%20to%200il%20and%20Natural%20Gas%20Companies.pdf
Ohio EPA (Ohio Environmental Protection Agency). (2012b). Ohios regulations: A guide for
operators drilling in the Marcellus and Utica Shales. Columbus, OH.
http://www.epa.state.oh.us/Portals/0/general%20pdfs/Ohio%20Regulations%20-
%20A%20Guide%20for%200perators%20Drilling%20in%20the%20Marcellus%20and%
20Utica%20Shales.pdf
Oil and Gas Mineral Services. (2010). MineralWise: Oil and gas terminology.
http://www.mineralweb.com/library/oil-and-gas-terms/
Olawovin. R: Wang. TY: Ovewole. SA. (2013). Environmental safety assessment of drilling operations
in the Marcellus-shale gas development SPE Drilling & Completion 28: 212-220.
http://dx.doi.org/10.2118/163095-PA
Olmstead. SM: Muehlenbachs. LA: Shih. IS: Chu. Z: Krupnick. AT. (2013). Shale gas development
impacts on surface water quality in Pennsylvania. Proc Natl Acad Sci USA 110: 4962-4967.
http://dx.doi.org/10.1073/pnas.1213871110
Olson. IE. (2011). Hydraulic fracturing overview. Presented at Summer Institute B: Energy, Climate
and Water in the 21st Century, TXESS Revolution, Texas Earth and Space Science Revolution
Professional Development for Educators, June 2011, Austin, TX.
0MB (U.S. Office of Management and Budget). (2004). Final information quality bulletin for peer
review. (Memorandum M-05-03). Washington, D.C.: US Office of Management and Budget
(0MB).
http://www.whitehouse.gOv/sites/default/files/omb/assets/omb/memoranda/fv2005/m
05-03.pdf
Orem. W: Tatu. C: Varonka. M: Lerch. H: Bates. A: Engle. M: Crosby. L: Mcintosh. 1. (2014). Organic
substances in produced and formation water from unconventional natural gas extraction in
coal and shale. Int J Coal Geol 126: 20-31. http://dx.doi.Org/10.1016/i.coal.2014.01.003
Orem. WH: Tatu. CA: Lerch. HE: Rice. CA: Bartos. TT: Bates. AL: Tewalt. S: Corum. MP. (2007).
Organic compounds in produced waters from coalbed natural gas wells in the Powder River
11-43
-------
Chapter 11 - References
Basin, Wyoming, USA. Appl Geochem 22: 2240-2256.
http://dx.doi.Org/10.1016/i.apgeochem.2007.04.010
Osborn. S: Mcintosh. T. (2011). Distinguishing the source of natural gas accumulations with a
combined gas and coproduced formation water geochemical approach. In Proceedings of
the Technical Workshops for the Hydraulic Fracturing Study: Chemical and Analytical
Methods (pp. 60-62). Washington, D.C.: U.S. Environmental Protection Agency.
https://www.epa.gov/sites/production/files/2015-
10/documents /proceedingsofhfchemanalmethodsfinalmay2011. pdf
Osborn. SG: Vengosh. A: Warner. NR: Tackson. RB. (2011). Methane contamination of drinking water
accompanying gas-well drilling and hydraulic fracturing. Proc Natl Acad Sci USA 108: 8172-
8176. http://dx.doi.Org/10.1073/pnas.1100682108
OSHA (Occupational Safety & Health Administration). (2014a). Personal communication: Email
exchanges between Tandy Zitkus, OSHA and Rebecca Daiss, U.S. EPA.
OSHA (Occupational Safety & Health Administration). (2014b). Personal communication: Phone
conversation between Tandy Zitkus, OSHA and Rebecca Daiss, U.S. EPA.
OSHA (Occupational Safety & Health Administration). (2015). Oil and gas well drilling and servicing
etool: Well completion.
https://www.osha.gov/SLTC/etools/oilandgas/well completion/well completion.html
Otton. IK: Zielinski. RA: Smith. BP: Abbott. MM. (2007). Geologic controls on movement of
produced-water releases at US geological survey research Site A, Skiatook Lake, Osage
County, Oklahoma. Appl Geochem 22: 2138-2154.
http://dx.doi.Org/10.1016/i.apgeochem.2007.04.015
PA PEP (Pennsylvania Pepartment of Environmental Protection). (2000). Pennsylvanias plan for
addressing problem abandoned wells and orphaned wells. Harrisburg, PA: Commonwealth
of Pennsylvania Pepartment of Environmental Protection (PA PEP).
PA PEP (Pennsylvania Pepartment of Environmental Protection). (2009a). Consent order and
agreement in the matter of: Cabot Oil and Gas Corporation, Pimock and Springville
Townships, Susquehanna County. Harrisburg, PA: Commonwealth of Pennsylvania
Pepartment of Environmental Protection (PA PEP).
PA PEP (Pennsylvania Pepartment of Environmental Protection). (2009b). Inspection Report,
inspection record #1835041, enforcement record #251134. Harrisburg, PA:
Commonwealth of Pennsylvania Pepartment of Environmental Protection, Oil and Gas
Management Program.
PA PEP (Pennsylvania Pepartment of Environmental Protection). (2009c). Stray natural gas
migration associated with oil and gas wells [draft report], Harrisburg, PA: Commonwealth
of Pennsylvania Pepartment of Environmental Protection (PA PEP).
http://www.dep.state.pa.us/dep/subiect/advcoun/oil gas/2009/Stray%20Gas%20Migrati
on%2 OCases.pdf
PA PEP (Pennsylvania Pepartment of Environmental Protection). (2010a). Chemicals used by
hydraulic fracturing companies in Pennsylvania for surface and hydraulic fracturing
activities. Harrisburg, PA: Commonwealth of Pennsylvania Pepartment of Environmental
Protection (PA PEP).
http://files.dep.state.pa.us/oilgas/bogm/bogmportalfiles/MarcellusShale/Frac%201ist%20
6-30-2010.pdf
11-44
-------
Chapter 11 - References
PA PEP (Pennsylvania Department of Environmental Protection). (2010b). Consent order and
agreement in the matter of: Cabot Oil and Gas Corporation, Dimock and Springville
Townships, Susquehanna County. Harrisburg, PA: Commonwealth of Pennsylvania
Department of Environmental Protection (PA DEP).
http://files.dep.state.pa.us/oilgas/OilGasLandingPageFiles/FinalCO&Al 21510.pdf
PA DEP (Pennsylvania Department of Environmental Protection). (2010c). DEP fines Atlas
Resources for drilling wastewater spill in Washington County, (accessed February 13,
2014).
PA DEP (Pennsylvania Department of Environmental Protection). (2010d). Modification to consent
order and agreement dated November 4, 2009, in the matter of: Cabot Oil and Gas
Corporation, Dimock and Springville Townships, Susquehanna County. Harrisburg, PA:
Commonwealth of Pennsylvania Department of Environmental Protection (PA DEP).
http://s3.documentcloud.org/documents/1688/pennsylvanias-dep-orders-cabot-to-plug-
three-gas-wells.pdf
PA DEP (Pennsylvania Department of Environmental Protection). (2011a). Letter from
Pennsylvania Department of the Environment to US EPA Region 3 Administrator Shawn
Garvin, https://www.epa.gov/foia/letter-padep-regional-administrator-shawn-m-garvin-0
PA DEP (Pennsylvania Department of Environmental Protection). (2011b). Road-spreading of brine
for dust control and road stabilization, 43 Pa.B. 7377
http://www.pabulletin.com/secure/data/vol43/43-50/2362b.html
PA DEP (Pennsylvania Department of Environmental Protection). (2011c). Surface water sample
analytical results from XTO 308 response data from XTO February 3, 2011 CAWP
addendum. Indiana, PA: XTO Energy.
PA DEP (Pennsylvania Department of Environmental Protection). (2012). Oil and gas compliance
report database [Database], Harrisburg, PA: Commonwealth of Pennsylvania Department of
Environmental Protection (PA DEP).
http://www.depreportingservices.state.pa.us/ReportServer/Pages/ReportViewer.aspx7/0i
1 Gas/OG Compliance
PA DEP (Pennsylvania Department of Environmental Protection). (2015a). PA DEP oil & gas
reporting website - Statewide data downloads by reporting period. Waste and production
files downloaded for Marcellus/unconventional wells, July 2009-December 2014.
Harrisburg, PA: Commonwealth of Pennsylvania Department of Environmental Protection.
https://www.paoilandgasreporting.state.pa.us/publicreports/Modules/DataExports/DataE
xports.aspx
PA DEP (Pennsylvania Department of Environmental Protection). (2015b). Technologically
enhanced naturally occurring radioactive materials (TENORM) study report Harrisburg,
PA: Commonwealth of Pennsylvania Department of Environmental Protection.
PA DEP (Pennsylvania Department of Environmental Protection). (2016a). New regulations of shale
drilling take effect. Chapter 78A regulations strengthen water protections and improve
public data access, (accessed October 19, 2016).
PA£EP (Pennsylvania Department of Environmental Protection). (2016b). PADEP oil&gas
reporting website - Statewide data downloads by reporting period. Waste and production
files downloaded for Marcellus/unconventional wells, July 2009-June 2016. Harrisburg, PA:
Commonwealth of Pennsylvania Department of Environmental Protection.
https://www.paoilandgasreporting.state.pa.us/publicreports/Modules/DataExports/DataE
xports.aspx
11-45
-------
Chapter 11 - References
Palmer. I. (2010). Coalbed methane completions: A world view. Int J Coal Geol 82: 184-195.
http://dx.doi.Org/10.1016/i.coal.2009.12.010
Palmer. ID: Moschovidis. ZA: Cameron. TR. (2005). Coal failure and consequences for coalbed
methane wells. Presented at SPE annual technical conference and exhibition, October 9-12,
2005, Dallas, TX. https://www.onepetro.org/conference-paper/SPE-96872-MS
Papoulias. DM: Velasco. AL. (2013). Histopathological analysis of fish from Acorn Fork Creek,
Kentucky, exposed to hydraulic fracturing fluid releases. Southeastern Naturalist 12: 92-
111.
Parker. KM: Zeng. T: Harkness. 1: Vengosh. A: Mitch. WA. (2014). Enhanced formation of disinfection
byproducts in shale gas wastewater-impacted drinking water supplies. Environ Sci Technol
48: 11161-11169. http://dx.doi.org/10'l021/es5028184
Pashin. TC: Mcintvre-Redden. MR: Mann. SD: Kopaska-Merkel. DC: Varonka. M: Orem. W. (2014).
Relationships between water and gas chemistry in mature coalbed methane reservoirs of
the Black Warrior Basin. Int J Coal Geol 126: 92-105.
http://dx.doi.Org/10.1016/i.coal.2013.10.002
Patel. PS: Robart. CI: Ruegamer. M: Yang. A. (2014). Analysis of US hydraulic fracturing fluid system
and proppant trends. In 2014 SPE Hydraulic Fracturing Technology Conference, 4-6
February, 2014, The Woodlands, TX: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/168645-MS
Paterra. P. (2011). DEP shuts down Tri-County Waste Water over illegal dumping.
http://triblive.eom/x/pittsburghtrib/news/regional/s 728516.html#axzz3UCvkvM7t
(accessed March 12, 2015).
Patterson. LA: Tordaan. SM: Anadon. LP. (2016). A spatiotemporal exploration of water
consumption changes resulting from the coal-to-gas transition in Pennsylvania. (NI WP 16-
01). Durham, NC: Duke Unviersity.
https://nicholasinstitute.duke.edu/sites/default/files/publications/ni wp 16-01 pdf.pdf
Patzek. TW: Male. F: Marder. M. (2013). Gas production in the Barnett Shale obeys a simple scaling
theory. Proc Natl Acad Sci USA 110: 19731-19736.
http://dx.doi.org/10.1073/pnas.1313380110
Pearson. CM: Griffin. L: Wright. CA: Weiiers. L. (2013). Breaking up is hard to do: Creating hydraulic
fracture complexity in the Bakken Central Basin. In 2013 SPE Hydraulic Fracturing
Technology Conference, 4-6 February, 2013, The Woodlands, TX: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/163827-MS
Pearson. K. (2011). Geologic models and evaluation of undiscovered conventional and continuous
oil and gas resources- Upper Cretaceous Austin Chalk, U.S. Gulf Coast. In U.S. Geological
Survey Scientific Investigations Report (20125159). U.S. Geological Survey.
http://pubs.usgs.gov/sir/2012/5159/
PEER (Public Employees for Environmental Responsibility). (2015). Re: Petition for review of
national pollutant discharge elimination system (NPDES) permits.
http://www.peer.Org/assets/docs/wy/4 14 15 PEER appeal Wind River.pdf
Penttila. B: Heine. L: Craft. E. (2013). Manuscript in preparation assessing the hazard data gap for
hydraulic fracturing chemicals [Technical Report],
Peraki. M: Ghazanfari. E. (2014). Electrodialysis treatment of flow-back water for environmental
protection in shale gas development In Shale Gas Development: Shale Energy Engineering
11-46
-------
Chapter 11 - References
2014. Reston, VA: American Society of Civil Engineers.
http://dx.doi.org/10.1061/9780784413654.008
Peterman. ZE: Thamke. T: Futa. K: Oliver. T. (2012). Strontium isotope evolution of produced water
in the East Poplar Oil Field, Montana. Presented at US Geological Survey AAPG Annual
Convention and Exhibition, April 23, 2012, Long Beach, California. http://wv-
mt.water.usgs.gov/projects/east poplar/pdf/EPOF%20Strontium%20isotopes AAPG%20
meeting.pdf
Pfunt. H: Houben. G: Himmelsbach. T. (2016). Numerical modeling of fracking fluid migration
through fault zones and fractures in the North German Basin. Hydrogeo J April: 1-16.
http://dx.d0i.0rg/l 0.1007/si 0040-016-1418-7
Phillips. A. (2014). Frackers spill Olympic pools worth of hydrochloric acid in Oklahoma.
http://thinkprogress.org/climate/2014/07/31/3466283/olympic-pool-sized-
hydrochloric-acid-spill-oklahoma/
Pinder. GF: Celia. MA. (2006). Subsurface hydrology. Hoboken, NJ: John Wiley & Sons, Inc.
http://dx.d0i.0rg/l 0.1002/0470044209
Pinder. GF: Gray. WG. (2008). Flow equations. In Essentials of Multiphase Flow and Transport in
Porous Media. Hoboken, NJ: John Wiley & Sons.
http://www.wiley.com/WileyCDA/WileyTitle/productCd-047031762Q.html
Plewa. MI: Simmons. IE: Richardson. SD: Wagner. ED. (2010). Mammalian cell cytotoxicity and
genotoxicity of the haloacetic acids, a major class of drinking water disinfection by-
products. Environ Mol Mutagen 51: 871-878. http://dx.doi.org/10.1002/em.20585
Plewa. Ml: Wagner. ED. (2009). Quantitative comparative mammalian cell cytotoxicity and
genotoxicity of selected classes of drinking water disinfection by-products.
http://www.waterrf.org/ExecutiveSummaryLibrary/91249 3089 profile.pdf
Plumlee. MH: Debroux. IF: Taffler. D: Gravdon. TW: Mayer. X: Dahm. KG: Hancock. NT: Guerra. KL:
Xu. P: Drewes. IE: Cath. TY. (2014). Coalbed methane produced water screening tool for
treatment technology and beneficial use. J Unconven Oil Gas Resour 5: 22-34.
http://dx.doi.0rg/lO.lOl6/i.iuogr.2Oi3.i2.OO2
Plummer. M: Wood. T: Huang. H: Guo. L: Reiten. I: Chandler. K: Metesh. I. (2013). Water needs and
availability for hydraulic fracturing in the Bakken formation, eastern Montana. Presented at
2013 Technical Workshop, Water Acquisition Modeling: Assessing Impacts Through
Modeling and Other Means, June 4, 2013, Arlington, VA.
http://www2.epa.gov/hfstudy/water-needs-and-availability-hydraulic-fracturing-bakken-
formation-eastern-montana
Pope. PG: Martin-Poole. M: Speitel. GE: Collins. MR. (2007). Relative significance of factors
influencing DXAA formation during chloramination. JAWWA 99: 144-156.
Porcelli. D: Kim. CK: Martin. P: Moore. WS: Phaneuf. M. (2014). Properties of radium. In The
Environmental Behaviour of Radium: Revised edition. Vienna, Austria: International Atomic
Energy Agency. http: //www-pub.iaea.org/MTCD /Publications/PDF/trs476 web.pdf
Powell. B. (2013). Secondary containment: Regulations and best management practices in the
Appalachian Basin. (AADE-13-FTCE-18). Houston, Texas: American Association of Drilling
Engineers.
Powers. SE: Hunt. CS: Heermann. SE: Corseuil. HX: Rice. D: Alvarez. PIT. (2001). The transport and
fate of ethanol BTEX in groundwater contaminated by gasohol. Environ Sci Technol 31: 79-
123. http://dx.doi.Org/10.1080/20016491089181
11-47
-------
Chapter 11 - References
Purestream (Purestream Services). (2013). Purestream services will begin commercial operations
to treat Eagle Ford Shale produced and frac flow-back water in Gonzalez County, Texas.
Ouaranta. 1: Wise. R: Darnell. A. (2012). Pits and impoundments final report for assessing
environmental impacts of horizontal gas well drilling operations. (ETD-10 project).
http://www.dep.wv.gov/oil-and-gas/Horizontal-
Permits/legislativestudies/Documents/WVU%20Pit%20and%20lmpoundment%20Final%
20Report%20Dec%2017.%202012%20%20submitted%20Feb%201 S.%202013.pdf
Ouiroga. C: Tsapakis. I. (2015). Oil and gas energy developments and changes in crash trends in
Texas. (PRC 15-35 F). College Station, TX: Texas A&M Transportation Institute.
http://d2dtl5nnlpfr0r.cloudfront.net/tti.tamu.edu/documents/PRC-15-35-F.pdf
Rabinowitz. PM: Slizovskiv. IB: Lamers. V: Trufan. ST: Holford. TR: Dziura. ID: Peduzzi. PN: Kane. Ml:
Reif. IS: Weiss. TR: Stowe. MH. (2015). Proximity to natural gas wells and reported health
status: Results of a household survey in Washington County, Pennsylvania. Environ Health
Perspect 123: 21-26. http://dx.doi.org/10.1289/ehp.1307732
Rahm. BG: Bates. IT: Bertoia. LR: Galford. AE: Yoxtheimer. DA: Riha. ST. (2013). Wastewater
management and Marcellus Shale gas development: Trends, drivers, and planning
implications. J Environ Manage 120: 105-113.
http: / /dx. doi. or g/10.1016 /i ,i envman. 2013.02.029
Rahm. BG: Riha. ST. (2012). Toward strategic management of shale gas development: Regional,
collective impacts on water resources. Environ Sci Pol 17: 12-23.
http://dx.doi.Org/10.1016/i.envsci.2011.12.004
Rahm. BG: Riha. ST. (2014). Evolving shale gas management: Water resource risks, impacts, and
lessons learned [Review], Environ Sci Process Impacts 16: 1400-1412.
http://dx.doi.org/10.1039/c4em00Q18h
Rahm. BG: Vedachalam. S: Bertoia. LR: Mehta. D: Vanka. VS: Riha. ST. (2015). Shale gas operator
violations in the Marcellus and what they tell us about water resource risks. Energy Policy
82: 1-11. http://dx.doi.Org/10.1016/i.enpol.2015.02.033
Ramirez. P. (2002). Oil field produced water discharges into wetlands in Wyoming. (97-6- 6F34).
Cheyenne, WY: U.S. Fish and Wildlife Service, http: / /www.fws.gov/mountain-
prairie/contaminants /papers Zr6718c02.pdf
Rasa. E: Bekins. BA: Mackav. DM: de Sieves. NR: Wilson. IT: Feris. KP: Wood. IA: Scow. KM. (2013).
Impacts of an ethanol-blended fuel release on groundwater and fate of produced methane:
Simulation of field observations. Water Resour Res 49: 4907-4926.
http: / /dx. doi. or g/10.10 0 2 /wrcr. 20382
Rassenfoss. S. (2011). From flowback to fracturing: Water recycling grows in the Marcellus shale. J
Pet Tech 63: 48-51.
Reagan. MT: Moridis. GT: Tohnson. IN: Keen. ND. (2015). Numerical simulation of the environmental
impact of hydraulic fracturing of tight/shale gas reservoirs on near-surface groundwater:
Background, base cases, shallow reservoirs, short-term gas and water transport Water
Resour Res 51: 1-31. http://dx.doi.org/10.1002/2014WR016086
Regional Tier 3 Toxicity Value Workgroup. OSWER Human Health Regional Risk Assessors Forum..
(2013). Tier 3 toxicity value white paper. (OSWER 9285.7-86). Washington, D.C.: U.S.
Environmental Protection Agency.
https://nepis.epa.gov/Exe /ZyPURL.cgi?Dockey=P 10 0NTU6.txt
11-48
-------
Chapter 11 - References
Renpu. W. (2011). Advanced well completion engineering (3rd ed.). Houston, TX: Gulf Professional
Publishing.
Resnikoff. M: Alexandrova. E: Travers. 1. (2010). Radioactivity in Marcellus shale: Report prepared
for Residents for the Preservation of Lowman and Chemung (RFPLC). New York, NY:
Radioactive Waste Management Associates.
Reuters. (2014). Update 2-oil well in North Dakota out of control, leaking.
http://www.reuters.eom/article/2014/02/14/energy-crude-blowout-
idIJST,2N0TJl 5820140214 (accessed March 2, 2015).
Revesz. KM: Breen. KT: Baldassare. AT: Burruss. RC. (2012). Carbon and hydrogen isotopic evidence
for the origin of combustible gases in water-supply wells in north-central Pennsylvania.
Appl Geochem 27: 361-375. http://dx.doi.Org/10.1016/i.apgeochem.2011.12.002
Reynolds. RR: Kiker. RD. (2003). Produced water and associated issues a manual for the
independent operator. (Oklahoma Geological Survey Open-File Report 6-2003). Tulsa, OK:
Oklahoma Geological Survey. http: //www.ogs.ou.edu/PTTC/pwm/produced water.pdf
Rice. CA. (1999). Waters co-produced with coalbed methane from the Ferron Sandstone in east-
central Utah: Chemical and isotopic composition, volumes, and impacts of disposal
[Abstract], Geological Society of America Abstracts with Programs 31: A385.
Rice. CA: Ellis. MS: Bullock. TH. (2000). Water co-produced with coalbed methane in the Powder
River basin, Wyoming: Preliminary compositional data. (Open File-Report 00-372). Denver,
CO: U.S. Department of the Interior, U.S. Geological Survey.
Rich. AL: Crosby. EC. (2013). Analysis of reserve pit sludge from unconventional natural gas
hydraulic fracturing and drilling operations for the presence of technologically enhanced
naturally occurring radioactive material (TENORM). New Solutions: A Journal of
Environmental and Occupational Health Policy 23: 117-135.
http://dx.doi.Org/10.2190/NS.23.l.h
Richardson. N: Gottlieb. M: Krupnick. A: Wiseman. H. (2013). The state of state shale gas regulation.
Washington, D.C.: Resources for the Future.
http://www.rff.org/files/sharepoint/WorkImages/Download/RFF-Rpt-
StateofStateRegs Reportpdf
Richardson. SD: Plewa. Ml: Wagner. ED: Schoenv. R: Demarini. DM. (2007). Occurrence,
genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in
drinking water: A review and roadmap for research [Review], Mutat Res 636: 178-242.
https://www.ncbi.nlm.nih.gov/pubmed/17980649
Rickman. R: Mullen. Ml: Petre. IE: Grieser. WV: Kundert. D. (2008). A practical use of shale
petrophysics for stimulation design optimization: All shale plays are not clones of the
Barnett shale. In 2008 SPE Annual Technical Conference & Exhibition, 12-14 September,
2008, Denver, CO: Society of Petroleum Engineers. http: //dx.doi.org/10.2118 /115258-MS
Rivett. M0: Wealthall. GP: Dearden. RA: McAlarv. TA. (2011). Review of unsaturated-zone transport
and attenuation of volatile organic compound (V0C) plumes leached from shallow source
zones [Review], J Contam Hydrol 123: 130-156.
http://dx.doi.Org/10.1016/i.iconhyd.2010.12.013
Robertson. 10: Chilingar. GV: Khilvuk. LF: Endres. B. (2012). Migration of gas from oil/gas fields.
Energ Source Part A 34: 1436-1447. http: //dx.doi.org/10.1080 /15567030903077899
Rodvelt. GD: Yeager. VI: Hyatt. MA. (2013). Case history: challenges using ultraviolet light to control
bacteria in Marcellus completions. In 2011 SPE Eastern Regional Meeting, 17-19 August,
11-49
-------
Chapter 11 - References
2011, Columbus, OH: Society of Petroleum Engineers. http: //dx.doi.org/lQ.2118 /149445-
MS
Rolls. Rl: Leigh. C: Sheldon. F. (2012). Mechanistic effects of low-flow hydrology on riverine
ecosystems: Ecological principles and consequences of alteration. Freshwater Science 31:
1163-1186. http://dx.doi.org/10.1899/12-002.1
Ross. D: King. G. (2007). Well completions. In MJ Economides; T Martin (Eds.), Modern Fracturing:
Enhancing Natural Gas Production (1 ed., pp. 169-198). Houston, Texas: ET Publishing.
Rowan. EL: Engle. MA: Kirbv. CS: Kraemer. TF. (2011). Radium content of oil- and gas-field
produced waters in the northern Appalachian Basin (USA): Summary and discussion of
data. (Scientific Investigations Report 20115135). Reston, VA: U.S. Geological Survey.
http://pubs.usgs.gov/sir/2011/5135/
Rowe. D: Muehlenbachs. K. (1999). Isotopic fingerprints of shallow gases in the western Canadian
sedimentary basin: Tools for remediation of leaking heavy oil wells. Organic Geochemistry
30: 861 -871. http://dx.doi.org/l0.1016/S0146-6380r99100068-6
Roy. P: Morris. TP: Walsh. SDC: Toraster. M. (2016). Improved cementing practice prevents gas
migration and surface casing vent flows. (ARMA 16-0668). Alexandria, VA: American Rock
Mechanics Association. http://dx.doi.Org/10.13140/RG.2.l.1457.2404
Roy. SB: Ricci. PF: Summers. KV: Chung. CF: Goldstein. RA. (2005). Evaluation of the sustainability of
water withdrawals in the United States, 1995 to 2025. J Am Water Resour Assoc 41: 1091-
1108. http://dx.doi.Org/10.llll/i.1752-1688.2005.tb03787.x
Rovchaudhuri. B: Tsotsis. TT: lessen. K. (2011). An experimental and numerical investigation of
spontaneous imbibition in gas shales. Presented at SPE Annual Technical Conference and
Exhibition, October 30-November 2, 2011, Denver, Colorado.
https://www.onepetro.org/conference-paper/SPE-147652-MS
Rozell. DT: Reaven. ST. (2012). Water pollution risk associated with natural gas extraction from the
Marcellus Shale. Risk Anal 32: 13821393. http://dx.doi.0rg/lO.llll/i. 1539-
6924.2011.01757.x
Rubinstein. TL: Ellsworth. WL: McGarr. A: Benz. HM. (2014). The 2001-present induced earthquake
sequence in the Raton Basin of northern New Mexico and southern Colorado. Seismological
Society of America Bulletin 104: 2162-2181. http://dx.doi.org/10.1785/0120140009
Rushton. L: Castaneda. C. (2014). Drilling into hydraulic fracturing and the associated wastewater
management issues. Washington, D.C.: Paul Hastings, LLP.
http://www.paulhastings.com/docs/default-source/PDFs/stay-current-hydraulic-
fracturing-wastewater-managementpdf
Rutledge. TT: Phillips. WS. (2003). Hydraulic stimulation of natural fractures as revealed by induced
microearthquakes, Carthage Cotton Valley gas field, eastTexas. Geophysics 68: 441-452.
http://dx.doi.Org/10.1190/l.1567214
Rutqvist. 1: Rinaldi. AP: Cappa. F: Moridis. GT. (2013). Modeling of fault reactivation and induced
seismicity during hydraulic fracturing of shale-gas reservoirs. Journal of Petroleum Science
and Engineering 107: 31-44. http://dx.doi.Org/10.1016/i.petrol.2013.04.023
Rutqvist. 1: Rinaldi. AP: Cappa. F: Moridis. GT. (2015). Modeling of fault activation and seismicity by
injection directly into a fault zone associated with hydraulic fracturing of shale-gas
reservoirs. Journal of Petroleum Science and Engineering 127: 377-386.
http://dx.doi.Org/10.1016/j.petrol.2015.01.019
11-50
-------
Chapter 11 - References
Ryan. MC: Alessi. D: Mahani. AB: Cahill. A: Cherry. T: Eaton. D: Evans. R: Farah. N: Fernandes. A:
Forde. 0: Humez. P: Kletke. S: Ladd. B: Lemieux. TM: Mayer. B: Mayer. KU: Molson. T:
Muehlenbachs. L: Nowamooz. A: Parker. B. (2015). Subsurface impacts of hydraulic
fracturing: Contamination, seismic sensitivity, and groundwater use and demand
management. (CWN HF-KI Subsurface Impacts Report). Calgary, ON, Canada: Canadian
Water Network. http: //www.cwn-rce.ca/assets/resources/pdf/Hvdraulic-Fracturing-
Research-Reports/Ryan-et-al-2015-CWN-Report-Subsurface-Impacts-of-Hydraulic-
Fracturing.pdf
S.S. Papadopulos & Associates. Inc. (2008). Phase II hydrogeologic characterization of the Mamm
Creek Field area, Garfield County, Colorado, http://www.garfield-county.com/oil-
gas/phase-II-hydrogeologic-characterization-mamm-creek.aspx
SAB (Science Advisory Board). (2016). SAB review of the EPAs draft assessment of the potential
impacts of hydraulic fracturing for oil and gas on drinking water resources [EPA Report],
(EPA-SAB-16-005). Washington, D.C.: U.S. Environmental Protection Agency.
Saba. T: Mohsen. F: Garry. M: Murphy. B: Hilbert. B. (2012). White paper: Methanol use in hydraulic
fracturing fluids. (1103844.000 0101 0711 TS26). Maynard, MA: Exponent.
http://www.methanol.org/wp-content/uploads/2016/06/White-Paper-Methanol-Use-in-
Hydraulic-Fracturing-Ian-ll.pdf
Sabins. F. (1990). Problems in cementing horizontal wells. J Pet Tech 42: 398-400.
http://dx.doi.org/10.2118/20005-PA
SAFER PA (Shale Alliance for Energy Research Pennsylvania). (2015). Life cycle management of
future waste challenges for shale development in Pennsylvania. Harrisburg, PA:
Commonwealth of Pennsylvania Department of Environmental Protection.
http://www.saferpa.org/Pages/Life-Cvcle-Management-of-Future-Waste-Challenges-4-
Shale-Development-in-PA.aspx
SAIC and GES (SAIC Energy, Environment & Infrastructure, LLC and Groundwater & Environmental
Services, Inc). (2011). ATGAS investigation initial site characterization and response, April
19, 2011 to May 2, 2011, ATGAS2H Well Pad, permitno. 37-015-21237, Leroy Township,
Bradford County, PA. Harrisburg, Pennsylvania: Chesapeake Appalachia, LLC.
http://www.chk.com/news/articles/documents/atgas initial site characterization report
final 08292011.pdf
Salehi. I: Ciezobka. I. (2013). Controlled hydraulic fracturing of naturally fractured shales - A case
study in the Marcellus Shale examining how to identify and exploit natural fractures. In
Research Partnership to Secure Energy for America (RPSEA) final report: Marcellus Shale
gas project. Sugar Land, TX:. http://dx.doi.org/10.2118/164524-MS
Salman. A: Kurtoglu. B: Kazemi. H. (2014). Analysis of chemical tracer flowback in unconventional
reservoirs, 30 September-2 October, 2014, Calgary, Alberta, Canada: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/171656-MS
Sando. SK: Vecchia. AY: Barnhart. EP: Sando. TR: Clark. ML: Lorenz. PL. (2014). Trends in major-ion
constituents and properties for selected sampling sites in the Tongue and Powder River
watersheds, Montana and Wyoming, based on data collected during water years 1980 -
2010. (Scientific Investigations Report 20135179). Reston, VA: U.S. Geological Survey.
http://dx.doi.org/10.3133/sir20135179
Sang. W: Stoof. CR: Zhang. W: Morales. VL: Gao. B: Kay. RW: Liu. L: Zhang. Y: Steenhuis. TS. (2014).
Effect of hydrofracking fluid on colloid transport in the unsaturated zone. Environ Sci
Technol 48: 8266-8274. http://dx.doi.org/10.1021/es501441e
11-51
-------
Chapter 11 - References
Santa Cruz Biotechnology. (2015). Sorbitane trioleate (CAS 26266-58-0).
http: / /www, scbt com /datasheet- 28115 4- Sorbitane-T rioleate.html (accessed April 6, 2015).
Scanlon. BR: Reedy. RC: Male. F: Hove. M. (2016). Managing the increasing water footprint of
hydraulic fracturing in the Bakken Play, United States. Environ Sci Technol 50: 10273-
10281. http://dx.doi.org/10.1021/acs.est6b01375
Scanlon. BR: Reedy. RC: Nicot. TP. (2014a). Comparison of water use for hydraulic fracturing for
unconventional oil and gas versus conventional oil. Environ Sci Technol 48: 12386-12393.
http://dx.doi.org/10.1021/es5025Q6v
Scanlon. BR: Reedy. RC: Nicot. TP. (2014b). Will water scarcity in semiarid regions limit hydraulic
fracturing of shale plays? Environmental Research Letters 9.
http://dx.doi.Org/10.1088/1748-9326/9/12/124011
Schenk. CI: Pollastro. RM. (2002). Natural gas production in the United States [Fact Sheet], (FS-113-
01). Denver, CO: U.S. Geological Survey, https://pubs.er.usgs.gov/publication/fsll301
Schlumberger (Schlumberger Limited). (2006). Fundamentals of formation testing. Sugar Land,
Texas.
http://www.slb.eom/~/media/Files/evaluation/books/fundamentals formation testing o
verview.pdf
Schlumberger (Schlumberger Limited). (2014). Schlumberger oilfield glossary.
http://www.glossary.oilfield.slb.com/
Schlumberger (Schlumberger Limited). (2015). Stimulation.
http://www.slb.com/services/completions/stimulation.aspx
Schmid. K: Yoxtheimer. D. (2015). Wastewater recycling and reuse trends in Pennsylvania shale gas
wells. Environmental Geosciences 22: 115-125.
Schmidlev. EB: Smith. BT. (2011). Personal communication from Schmidley and Smith to DiCello:
CAWP Addendum EM Survey & Well Location; XTO Energy, Inc. Marquardt Release.
Schmidt. CW. (2013). Estimating wastewater impacts from fracking. Environ Health Perspect 121:
A117. http://dx.doi.org/10.1289/ehp.121-all7
Schnoor. TL. (1996). Environmental modeling: Fate and transport of pollutants in water, air, and soil
(lsted.). Hoboken, NJ: Wiley-Interscience.
Schrodinger. (2012). Qikprop [Computer Program], New York, New York: Schrodinger, LLC.
Retrieved from http://www.schrodinger.com/products/14/17
Schubert. 1: Rosenmeier. 1: Zatezalo. M. (2014). A review of NORM/TENORM in wastes and waters
associated with Marcellus shale gas development and production. In CL Meehan; JM
Vanbriesen; F Vahedifard; XYu; C Quiroga (Eds.), Shale energy engineering 2014: technical
challenges, environmental issues, and public policy (pp. 492-501). Reston, VA: American
Society of Civil Engineers. http: //dx.doi.org/10.1061/9780784413654.Q52
Schwartz. MO. (2015). Modelling the hypothetical methane-leakage in a shale-gas project and the
impact on groundwater quality. Environ Earth Sci 73: 4619-4632.
http: //dx.doi.org/10.1007/sl 2665-014-3787-3
Schwarzenbach. RP: Gschwend. PM: Imboden. DM. (2002). Environmental Organic Chemistry (2nd
ed.). Hoboken, NJ: John Wiley & Sons, Inc.
Science Applications International Corporation. (2010). XTO - Marquardt pad soil and water
corrective action workplan. (XTO-EPA0001074). Indiana, PA: XTO Energy.
11-52
-------
Chapter 11 - References
Science Based Solutions LLC. (2014). Summary of hydrogeology investigations in the Mamm Creek
field area, Garfield County. Laramie, Wyoming, http://www.garfield-countv.com/oil-
gas/documents/Summary-Hydrogeologic-Studies-Mamm%20Creek-Area-Feb-10-2014.pdf
Senior. LA. (2014). A reconnaissance spatial and temporal baseline assessment of methane and
inorganic constituents in groundwater in bedrock aquifers, Pike County, Pennsylvania,
201213 (pp. i-106). (2014-5117). Senior, LA.
http://pubs.usgs.gOv/sir/2014/5117/support/sir2014-5117.pdf
Shaffer. PL: Arias Chavez. LH: Ben-Sasson. M: Romero-Vargas Castrillon. S: Yip. NY: Elimelech. M.
(2013). Desalination and reuse of high-salinity shale gas produced water: Drivers,
technologies, and future directions. Environ Sci Technol 47: 9569-9583.
Shammas. NK. (2010). Wastewater renovation by flotation. In LK Wang; NK Shammas; WA Selke;
DB Aulenbach (Eds.), Flotation Technology (pp. 327-345). New York, NY: Humana Press.
http://dx.d0i.0rg/l0.1007/978-1 -60327-133-2 9
Shank. MK: Stauffer. 1. avR. T..r. (2015). Land use and surface water withdrawal effects on fish and
macroinvertebrate assemblages in the Susquehanna River basin, USA. J Freshwat Ecol 30:
229-248. http://dx.d0i.0rg/l 0.1080/02705060.2014.959082
Shapiro. SA: Kriiger. OS: Dinske. C: Langenbruch. C. (2011). Magnitudes of induced earthquakes and
geometric scales of fluid-stimulated rock volumes. Geophysics 76: WC55-WC63.
http://dx.doi.Org/10.1190/geo2010-0349.l
Sharma. S: Boardman. G: Swann. C. Parks. 1.: Countess. S. (2015). Leachability of elements from
hydraulic fracturing residuals. U.S. Department of Energy, National Energy Technology
Laboratory. http://www.tsphosting.eom/testbed3.tsphosting.com/content/ECEP2015-
005.pdf
Sharma. S: Bowman. L: Schroeder. K: Hammack. R. (2014a). Assessing changes in gas migration
pathways at a hydraulic fracturing site: Example from Greene County, Pennsylvania, USA.
Appl Geochem. http://dx.doi.Org/10.1016/j.apgeochem.2014.07.018
Sharma. S: Mulder. ML: Sack. A: Schroeder. K: Hammack. R. (2014b). Isotope approach to assess
hydrologic connections during Marcellus Shale drilling. Ground Water 52: 424433.
http: / /dx. doi. or g/10.1111 /gwat. 12083
Sheng. Z. (2005). An aquifer storage and recovery system with reclaimed wastewater to preserve
native groundwater resources in El Paso, Texas. J Environ Manage 75: 367-377.
http://dx.doi.Org/10.1016/j.jenvman.2004.10.007
Sherwood. OA: Rogers. ID: Lackey. G: Burke. TL: Osborn. SG: Ryan. IN. (2016). Groundwater
methane in relation to oil and gas development and shallow coal seams in the Denver-
Julesburg Basin of Colorado. Proc Natl Acad Sci USA.
http://dx.doi.org/10.1073/pnas.1523267113
Shiklomanov. I. (1993). World fresh water resources. In PH Gleick (Ed.), Water in crisis: A guide to
the world's fresh water resources (1st ed.). New York, NY: Oxford University Press.
Shires. T: Lev-On. M. (2012). Characterizing pivotal sources of methane emissions from
unconventional natural gas production - Summary and analysis of API and ANGA survey
responses. Washington, D.C.: American Petroleum Institute, American Natural Gas Alliance.
Siegel. PI: Azzolina. NA: Smith. BT: Perry. AE: Bothun. RL. (2015). Methane concentrations in water
wells unrelated to proximity to existing oil and gas wells in northeastern Pennsylvania.
Environ Sci Technol 49: 4106-4112. http://dx.doi.org/10.1021/es505775c
11-53
-------
Chapter 11 - References
Silber. R: Martin. T: Willis. S: Kozera. G. (2003). Comparing fracture simulation design to radioactive
tracer field results: A case history. Presented at SPE Eastern Regional Meeting, 6-10
September, 2003, Pittsburgh, PA: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/84842-MS
Silva. TM: Gettings. RM: Kostedt. WL: Watkins. VH. (2014a). Produced water from hydrofracturing:
Challenges and opportunities for reuse and recovery. The Bridge (Washington) 44: 34-40.
Silva. TM: Gettings. RM. Kostedt. W. L.. IV: Watkins. VH: Hardy. A: Shapiro. A. (2014b). NORM
mitigation and clean water recovery from Marcellus produced water. Research partnership
to secure energy for America. (10122.07). GE Global Research.
https://www.netl.doe.gov/File%20Library/Research/Oil-
Gas/Natural%20Gas/environmental/10122-07-final-reportpdf
Silva. TM: Matis. H: Kostedt. WL: Watkins. V. (2012). Produced water pretreatment for water
recovery and salt production. (08122-36). Niskayuna, NY: Research Partnership to Secure
Energy for America. http://www.rpsea.org/media/files/proiect/18621900/08122-36-FR-
Pretreatment Water Mgt Frac Water Reuse Salt-01-26-12.pdf
Sionix (Sionix Corporation). (2011). Sionix to build Bakken water treatment plant.
http://www.rigzone.com/news/article pf.asp?a id=l 10613
Sirivedhin. T: Dallbauman. L. (2004). Organic matrix in produced water from the Osage-Skiatook
petroleum environmental research site, Osage County, Oklahoma. Chemosphere 57: 463-
469.
Siolander. SA: Clark. 1: Rizzo. D: Turack. 1. (2011). Water facts #31: Introduction to hydrofracturing.
University Park, PA: Penn State College of Agricultural Sciences - Cooperative Extension.
http://www.shale-gas-information-
platform.org/fileadmin/ship/dokumente/introduction to hvdrofracturing-2.pdf
Skalak. KT: Engle. MA: Rowan. EL: lolly. GD: Conko. KM: Benthem. AT: Kraemer. TF. (2014). Surface
disposal of produced waters in western and southwestern Pennsylvania: Potential for
accumulation of alkali-earth elements in sediments. Int J Coal Geol 126: 162-170.
http://dx.doi.Org/10.1016/i.coal.2013.12.001
Skierven. T: Lunde. 0: Perander. M: Williams. B: Farquhar. R: Sinet. 1: Saebv. I: Haga. HB: Finnseth. 0:
Tohnsen. S. (2011). Norwegian Oil and Gas Association recommended guidelines for well
integrity. (117, Revision 4). Norway: Norwegian Oil and Gas Association.
Skoumal. RT: Brudzinski. MR: Currie. BS. (2015). Earthquakes induced by hydraulic fracturing in
Poland Township, Ohio. Seismological Society of America Bulletin 105: 189-197.
http://dx.doi. org/10.1785/0120140168
Slutz. I: Anderson. 1: Broderick. R: Horner. P. (2012). Key shale gas water management strategies:
An economic assessment tool. Paper presented at International Conference on Health,
Safety and Environment in Oil and Gas Exploration and Production, September 11-13, 2012,
Perth, Australia. http://www.onepetro.org/mslib/app/Preview.do?paperNumber=SPE-
157532-MS&societyCode=SPE
Sminchak. TR. IR: Moody. M: Theodos. A: Larsen. G: Gupta. N. (2014). Investigation of wellbore
integrity factors in historical oil and gas wells for C02 geosequestration in the Midwestern
US. Energy Procedia 63: 5787-5797. http://dx.doi.Org/10.1016/j.egypro.2014.ll.611
Smith. KP: Arnish. II: Blunt. PL: Haffenden. RA: Herbert. I: Pfingston. M: Williams. GP. (1999).
Assessment of the disposal of petroleum industry NORM in nonhazardous landfills.
11-54
-------
Chapter 11 - References
(DOE/BC/W-31-1 09-ENG-388). Washington, D.C.: Department of Energy.
http://www.evs.anl.gov/downloads/Smith-et-al-1999-Landfill-Disposal-of-N0RM.pdf
Smith. KP: Blunt. PL: Arnish. 11. (1998). Potential radiological doses associated with the disposal of
petroleum industry NORM via landspreading. Final report, September 1998.
(NTIS/03010458_a; DOE/BC/W-31-109-ENG-38-5). Washington, D.C.: Department of
Energy.
Smith. MB: Montgomery. C. (2015). Hydraulic fracturing. Boca Raton, FL: CRC Press, Taylor &
Francis Group.
Smolen. 11. (2006). Cased hole and production log evaluation. Tulsa, OK: PennWell Books.
Soeder. DT: Kappel. WM. (2009). Water resources and natural gas production from the Marcellus
shale [Fact Sheet] (pp. 6). (U.S. Geological Survey, Fact Sheet 20093032).
http://pubs.usgs.gOv/fs/2009/3032/pdf/FS2009-3032.pdf
Sollev. WB: Pierce. RR: Perlman. HA. (1998). Estimated use of water in the United States in 1995.
(USGS Circular: 1200). U.S. Geological Survey, http: //pubs.er.usgs.gov/publication/cirl200
Sonnenfield. M: Katz. D: Odegard. M: Ohlson. C: Zahm. C. (2016). Niobrara core poster highlighting
bentonite distribution and their impacts on proppant placement Presented at AAPG 2015
Annual Convention and Exhibition, May 31 - June 3, 2015, Denver, CO.
http://www.searchanddiscovery.eom/documents/2016/41803sonnenfeld/ndx sonnenfeld
.pdf
Spellman. FR. (2012). Chapter 7. Chemicals Used in Hydraulic Fracturing. In Environmental Impacts
of Hydraulic Fracturing. Boca Raton, Florida: CRC Press.
SRBC (Susquehanna River Basin Commission). (2013). Comprehensive plan for the water resources
of the Susquehanna River basin. Harrisburg, PA.
http://www.srbc.net/planning/comprehensiveplan.htm
SRBC (Susquehanna River Basin Commission). (2016). Water use associated with natural gas
development: An assessment of activities managed by the Susquehanna river basin
commission July 2008 - December 2013. (Publication No. 299). Harrisburg, PA.
http://www.srbc.net/pubinfo/techdocs/NaturalGasReport/docs/SRBC Full Gas Report fs
306397vl 20160408.pdf
SRP (Salt River Project). (2016). Daily water reports. Available online at
http://data.hydrometdataservice.info/dwr/
Stacy. SL: Brink. LL: Larkin. TC: Sadovskv. Y: Goldstein. BP: Pitt. BR: Talbott. EO. (2015). Perinatal
outcomes and unconventional natural gas operations in Southwest Pennsylvania. PLoS ONE
10: e0126425. http://dx.doi.org/10.1371/journal.pone.0126425
States First Initiative. (2014). States team up to assess risk of induced seismicity.
http://www.statesfirstinitiative.org/single-post/2014/04/29/States-Team-Up-to-Assess-
Risk-of-Induced-Seismicitv (accessed September 12, 2016).
States. S: Cvprvch. G: Stoner. M: Wvdra. F: Kuchta. 1: Monnell. I: Casson. L. (2013). Marcellus Shale
drilling and brominated THMs in Pittsburgh, Pa., drinking water. J Am Water Works Assoc
105: E432-E448. http://dx.doi.org/10.5942/iawwa.2013.105.0093
Steadman. MP: Arnett. B: Healv. K: Tiang. Z: LeClare. D: Mclaughlin. L: Roberts. 1. (2015).
Groundwater use in the Eagle Ford Shale: Some policy recommendations. Texas Water
Journal 6: 67-78.
11-55
-------
Chapter 11 - References
Stepan. DT: Shockev. RE: Kurz. BA: Kalenze. NS: Cowan. RM: Ziman. TT: Hariu. TA. (2010). Bakken
water opportunities assessment: Phase I. (2010-EERC-04-03). Bismarck, ND: North Dakota
Industrial Commission, http: //www.nd.gov/ndic/ogrp/info/g-018-036-fi.pdf
Stinger Wellhead Protection. Inc. (2010). Stinger Wellhead Protection. Houston, TX: Stinger
Wellhead Protection, Inc. http: //etdevelopers.com/design-preview/STS/product-
catalog/STS Product Catalog 2010-SWP.pdf
STO (Statoil). (2013). Shale facts: Drilling and hydraulic fracturing, how it's done, responsibly.
(Global Version, April 2013). Stavanger, Norway.
http://www.statoil.eom/no/OurOperations/ExplorationProd/ShaleGas/FactSheets/Downl
oads/Shale DrillingHydraulicFacturing.pdf
Stringfellow. WT: Domen. IK: Camarillo. MK: Sandelin. WL: Borglin. S. (2014). Physical, chemical,
and biological characteristics of compounds used in hydraulic fracturing. J Hazard Mater
275: 37-54. http://dx.doi.Org/10.1016/j.jhazmat.2014.04.040
STRONGER (State Review of Oil and Natural Gas Environmental Regulations). (2011a). Louisiana
hydraulic fracturing state review. Oklahoma City, OK.
http://www.shalegas.energy.gov/resources/071311 stronger louisiana hfreview.pdf
STRONGER (State Review of Oil and Natural Gas Environmental Regulations). (2011b). Ohio
hydraulic fracturing state review. Oklahoma City, OK.
https://oilandgas.ohiodnr.gov/portals/oilgas/pdf/stronger reviewll.pdf
STRONGER (State Review of Oil and Natural Gas Environmental Regulations). (2012). Arkansas
hydraulic fracturing state review. Oklahoma City, OK.
http://www.aogc.state.ar.us/notices/AR HFR FINAL.pdf
Sturchio. NC: Banner. TL: Binz. CM: Heratv. LB: Musgrove. M. (2001). Radium geochemistry of
ground waters in Paleozoic carbonate aquifers, midcontinent, USA. Appl Geochem 16: 109-
122.
Sugden. C: Tohnson. 1: Chambers. M: Ring. G: Suryanaravana. PV. (2013). Special considerations in
the design optimization of the production casing in high-rate, multistage-fractured shale
wells. SPE Drilling & Completion 27: 459-472. http://dx.doi.org/10.2118/151470-PA
Sumi. L. (2004). Pit pollution: Backgrounder on the issues, with a New Mexico case study.
Washington, D.C.: Earthworks: Oil and Gas Accountability Project
http://www.earthworksaction.org/files/publications/PitReportpdf
Sumi. L. (2015). Gaps in water and waste reporting for oil & gas production: A five-state review.
Washington, D.C.: Earthworks.
https://www.earthworksaction.org/files/publications/GapsInWaterAndWasteReportingIn
5States-FINAL.pdf
Sun. M: Lowrv. GV: Gregory. KB. (2013). Selective oxidation of bromide in wastewater brines from
hydraulic fracturing. Water Res 47: 3723-3731.
http://dx.doi.Org/10.1016/j.watres.2013.04.041
Swann. C: Matthews. 1: Ericksen. R: Kuszaul. 1. (2004). Evaluations of radionuclides of uranium,
thorium, and radium with produced fluids, precipitates, and sludges from oil, gas, and
oilfield brine injections wells. (DE-FG26-02NT 15227). Washington, D.C.: U.S. Department of
Energy, http://www.olemiss.edu/depts/mmri/programs/norm final.pdf
Swanson. YE. (1955). Uranium in marine black shales of the United States. In Contributions to the
Geology of Uranium and Thorium by the United States Geological Survey and Atomic Energy
Commission for the United Nations International Conference on Peaceful Uses of Atomic
11-56
-------
Chapter 11 - References
Energy, Geneva, Switzerland, 1955 (pp. 451-456). Reston, VA: U.S. Geological Survey.
http://pubs.usgs.gOv/pp/0300/reportpdf
SWN (Southwestern Energy). (2011). Frac fluid-Whats in it? Houston, TX.
http://www.swn.com/operations/documents/frac fluid fact sheetpdf
SWN (Southwestern Energy). (2014). Field Site Visit at Southwestern Energy.
Sved. T: Cutler. T. (2010). Well integrity technical and regulatory considerations for C02 injection
wells. In 2010 SPE International Conference on Health, Safety & Environment in Oil and Gas
Exploration and Production, 12-14 April, 2010, Rio de Janeiro, Brazil: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/125839-MS
Tasker. TL: Piotrowski. PK: Dorman. FL: Burgos. WD. (2016). Metal associations in Marcellus Shale
and fate of synthetic hydraulic fracturing fluids reacted at high pressure and temperature.
Environ Eng Sci 33: 753-765. http://dx.doi.org/10.1089/ees.2015.0605
Tavlor. A. (2012). Watering the boom in Oklahoma: Supplies, demands, and neighbors. Presented at
2012 Kansas Water Issues Forums, February 29-March 1, 2012, Wichita and Hays, Kansas.
Taylor. RS: Lestz. RS: Loree. D: Funkhouser. GP: Fvten. G: Attawav. D: Watkins. H. (2006). Optimized
C02 miscible hydrocarbon fracturing fluids. Calgary, Alberta, Canada: Petroleum Society of
Canada, http://dx.doi.org/10.2118/2006-168
Thacker. IB: Carlton. DP. Tr: Hildenbrand. ZL: Kadio. AF: Schug. KA. (2015). Chemical analysis of
wastewater from unconventional drilling operations. Water 7: 1568-1579.
http://dx.doi.org/10.3390/w7041568
The Royal Society and the Royal Academy of Engineering. (2012). Shale gas extraction in the UK: A
review of hydraulic fracturing. London.
http://www.raeng.org.uk/news/publications/list/reports/Shale Gas.pdf
Thompson. AM. (2010) Induced fracture detection in the Barnett Shale, Ft Worth Basin, Texas.
(Master's Thesis). University of Oklahoma, Norman, OK.
Thordsen. II: Kharaka. YK: Ambats. G: Kakouros. E: Abbott. MM. (2007). Geochemical data from
produced water contamination investigations: Osage-Skiatook Petroleum Environmental
Research (OSPER) sites, Osage County, Oklahoma. (Open-File Report 2007-1055). Reston,
VA: United States Geological Survey.
https://pubs.usgs.gOv/of/2007/1055/pdf/ofr20071055.pdf
Thurman. EM: Ferrer. I: Blotevogel. I: Borch. T. (2014). Analysis of hydraulic fracturing flowback
and produced waters using accurate mass: Identification of ethoxylated surfactants. Anal
Chem 86: 9653-9661. http://dx.doi.org/l 0.1021 /ac502163k
Thurman. EM: Ferrer. I: Rosenblum. 1: Linden. K: Ryan. IN. (2016). Identification of polypropylene
glycols and polyethylene glycol carboxylates in flowback and produced water from
hydraulic fracturing. J Hazard Mater. http://dx.doi.Org/10.1016/i.ihazmat.2016.02.041
Tian. Z: Shi. L: Qiao. L. (2015). Problems in the wellbore integrity of a shale gas horizontal well and
corresponding countermeasures. Nat Gas Industry B 2: 522-529.
http://dx.doi.Org/10.1016/i.ngib.2015.12.006
Tidwell. VC: Kobos. PH: Malczvnski. LA: Klise. G: Castillo. CR. (2012). Exploring the water-
thermoelectric power nexus. J Water Resour Plann Manag 138: 491-501.
http://dx.doi.Org/10.1061/rASCElWR.1943-5452.0000222
11-57
-------
Chapter 11 - References
Tidwell. VC: Zemlick. K: Klise. G. (2013). Nationwide water availability data for energy-water
modeling. Albuquerque, New Mexico: Sandia National Laboratories.
http://prod.sandia.gov/techlib/access-control.cgi/2013/139968.pdf
Tiemann. M: Folger. P: Carter. NT. (2014). Shale energy technology assessment: Current and
emerging water practices. Washington, D.C.: Congressional Research Service.
http://nationalaglawcenter.Org/wp-content/uploads//assets/crs/R43635.pdf
Tillev. BT: Muehlenbachs. K. (2012). Fingerprinting of gas contaminating groundwater and soil in a
petroliferous region, Alberta, Canada. In RD Morrison; G O'Sullivan (Eds.), Environmental
forensics: Proceedings of the 211INEF Conference (pp. 115-125). London: RSC Publishing.
http://dx.d0i.0rg/l 0.1039/9781849734967-00115
TIPRO (Texas Independent Producers and Royalty Owners Association). (2012). Bradenhead
pressure management. Austin, TX.
http://www.tipro.org/UserFiles/BHP Guidance Final 071812.pdf
Titler. RV: Curry. P. (2011). Chemical analysis of major constituents and trace contaminants of rock
salt Harrisburg, PA: Pennsylvania Department of Environmental Protection.
http://files.dep.state.pa.us/water/Wastewater%20Management/WastewaterPortalFiles/R
ock%20Salt%20Paper%20final%2005 2711.pdf
Tomastik. T: Bair. ES. (2010). The Bainbridge Natural Gas Incident in Geauga County OH: What
really happened! Oklahoma City, OK: Groundwater Protection Council.
http://www.gwpc.org/sites/default/files/event-sessions/16Tomastik Tom O.pdf
Tucker. H. (2016). Letter from Holly Tucker, Chief, Ohio EPA Southeast District to Mayor Dale
Bunting, Barnesville, Ohio.
Tudor. EH: Nevison. GW: Allen. S: Pike. B. (2009). Case study of a novel hydraulic fracturing method
that maximizes effective hydraulic fracture length. In 2009 SPE Annual Technical
Conference and Exhibition, 4-7 October, 2009, New Orleans, LA: Society of Petroleum
Engineers. http://dx.doi.org/10.2118/124480-MS
Tustin. AW: Hirsch. AG: Rasmussen. SG: Casey. TA: Bandeen-Roche. K: Schwartz. BS. (2016).
Associations between unconventional natural gas development and nasal and sinus,
migraine headache, and fatigue symptoms in Pennsylvania. Environ Health Perspect.
http://dx.doi.org/10.1289/EHP281
TWDB (Texas Water Development Board). (2012). Water for Texas: 2012 State Water Plan. Austin,
TX. http://www.twdb.texas.gov/waterplanning/swp/2012/index.asp
TWDB (Texas Water Development Board). (2016). Water for Texas: 2017 State Water Plan. Austin,
TX. http://www.twdb.texas.gov/waterplanning/swp/2017/index.asp
Tyrrell. P. (2012). Water needs for oil & gas well drilling and fracturing. Presented at 85th Annual
AWSE Fall Conference, 23-26 September, 2012, Omaha, Nebraska.
http://westernstateengineers.files.wordpress.eom/2012/10/3 tvrrell awse 2012 omaha f
inal.pdf
U.S. Army Corps of Engineers. (2011). Final Garrison Dam/Lake Sakakawea project, North Dakota
surplus water report Volume 1 with addendum. Omaha, NE: U.S. Army Corps of Engineers,
Omaha District
http://www.swc.nd.gov/pdfs/corps final surplus storage lake sakakawea.pdf
U.S. Census Bureau. (2012). State area measurements and internal point coordinates.
http://www.census.gov/geo/reference/state-area.html
11-58
-------
Chapter 11 - References
U.S. Census Bureau. (2013). Annual estimates of the resident population: April 1, 2010 to July 1,
2013. Suitland, MD: U.S. Census Bureau, Population Division.
https://factfinder.census.gOv/bkmk/table/l.0/en/PEP/2013/PEPANNRES/0400000US05.
05000
U.S. Census Bureau. (2014). American FactFinder. Available online at
http: //factfinder.census. gov/faces/nav/i sf/pages/index.xhtml
U.S. Department of justice. (2014). Company owner sentenced to more than two years in prison for
dumping fracking waste in Mahoning River tributary.
http://www.iustice.gov/usao/ohn/news/2014/Q5auglupo.html (accessed March 4, 2015).
U.S. DPI. Bureau of Reclamation. (U.S. Department of the Interior, Bureau of Reclamation). (2016a).
Lower Colorado River operations schedule.
http://www.usbr.gov/rsvrWater/faces/rsvr40Day.xhtml
U.S. DPI. Bureau of Reclamation. (U.S. Department of the Interior, Bureau of Reclamation). (2016b).
Upper Colorado Region Water Pperations: 40-Day Datasets.
http://www.usbr.gov/rsvrWater/faces/rsvr40Day.xhtml
U.S. EPA (U.S. Environmental Protection Agency). (1989). Risk assessment guidance for superfund:
Volume 1: Human health evaluation manual (Part A): Interim final [EPA Report],
(EPA/540/1-89/002). Washington, D.C.: U.S. Environmental Protection Agency, Pffice of
Emergency and Remedial Response.
http://www.epa.gov/oswer/riskassessment/ragsa/index.htm
U.S. EPA (U.S. Environmental Protection Agency). (1991). Manual of individual and non-public
water supply systems [EPA Report], (EPA 570/9-91-004). Washington, D.C.
https://nepis.epa. gov/Exe/ZyPURL.cgi?Dockey=2000U9HN.txt
U.S. EPA (U.S. Environmental Protection Agency). (1996). Soil screening guidance: Technical
background document, Part 2 [EPA Report] (2nd ed.). (EPA/540/R-95/128). Washington,
D.C.: U.S. Environmental Protection Agency, Pffice of Solid Waste and Emergency Response.
U.S. EPA (U.S. Environmental Protection Agency). (1998). National primary drinking water
regulations: Disinfectants and disinfection byproducts. Fed Reg 63: 69390-69476.
https://www.epa.gov/dwreginfo/stage-l-and-stage-2-disinfectants-and-disinfection-
bvproducts-rules
U.S. EPA (Environmental Protection Agency). (2000). Development document for effluent
limitations guidelines and standards for the centralized waste treatment industry.
(821R00020). Washington, D.C.
U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes (pp. 1-192). (EPA/630/P-02/002F). Washington, D.C.:
U.S. Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/osa/review-reference-dose-and-reference-concentration-processes
U.S. EPA (U.S. Environmental Protection Agency). (2003a). Human health toxicity values in
superfund risk assessments, https: //rais.ornl.gov/documents/hhmemo.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2003b). Pil and gas environmental assessment
effort 1996 - 2002. Washington, D.C.: U.S. Environmental Protection Agency, Region 8.
https://www.earthworksaction.org/files/pubs-others/PGEA EPA landwaste 2003.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2003c). A summary of general assessment
factors for evaluating the quality of scientific and technical information [EPA Report],
(EPA/100/B-03/001). Washington, D.C.: U.S. Environmental Protection Agency, Pffice of
11-59
-------
Chapter 11 - References
Research and Development, http://www2.epa.gov/osa/summary-general-assessment-
factors-evaluating-quality-scientific-and-technical-information
U.S. EPA (U.S. Environmental Protection Agency). (2004a). Evaluation of impacts to underground
sources of drinking water by hydraulic fracturing of coalbed methane reservoirs [EPA
Report], (EPA/816/R-04/003). Washington, D.C.: U.S. Environmental Protection Agency,
Office of Solid Waste. https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100A99N.txt
U.S. EPA (U.S. Environmental Protection Agency). (2004b). Multi-agency radiological laboratory
analytical protocols (MARLAP) manual: part 1, chapters 1-9, appendices a-e (volume 1)
[EPA Report], (EPA 402-B-04-001A).
https://www.orau.org/ptp/PTP%20Library/library/Nrc/Nureg/1576vl.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2006). Technical support document for the 2006
Effluent Guidelines Program Plan. (EPA-821 R-06-018). Washington, D.C.
http://nepis.epa.gov/exe/ZyPURL.cgi?Dockey=P10099DV.txt
U.S. EPA (U.S. Environmental Protection Agency). (2010a). Coalbed methane extraction: Detailed
study report [EPA Report], (EPA-820-R-10-022). Washington, D.C.
https://nepis.epa.gov/Exe /ZyPURL.cgi?Dockey=P100MOTE.txt
U.S. EPA (U.S. Environmental Protection Agency). (2010b). Provisional peer-reviewed toxicity
values for formic acid (CASRN 64-18-6) [EPA Report], Cincinnati, OH.
http://hhpprtv.ornl.gov/issue papers/FormicAcid.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2011a). Data received from oil and gas
exploration and production companies, including hydraulic fracturing service companies,
2005 to 2010. http://www.regulations. gov /
U.S. EPA (U.S. Environmental Protection Agency). (2011b). Design for the Environment Program:
Alternatives assessment criteria for hazard evaluation (version 2.0). Washington, D.C.
http://www2.epa.gov/saferchoice/alternatives-assessment-criteria-hazard-evaluation
U.S. EPA (U.S. Environmental Protection Agency). (2011c). Ground water cleanup at superfund sites
[EPA Report], (EPA 540-K-96 008). Washington, D.C.: U. S. Environmental Protection
Agency, Office of Water.
U.S. EPA (U.S. Environmental Protection Agency). (201 Id). Plan to study the potential impacts of
hydraulic fracturing on drinking water resources [EPA Report], (EPA/600/R-11/122).
Washington, D.C.: U.S. Environmental Protection Agency, Office of Research and
Development, http://www2.epa.gov/hfstudy/plan-study-potential-impacts-hydraulic-
fracturing-drinking-water-resources-epa600r-11122
U.S. EPA (U.S. Environmental Protection Agency). (2011e). Sampling data for flowback and
produced water provided to EPA by nine oil and gas well operators (non-confidential
business information).
http://www.regulations.gov/#!docketDetail:rpp=100:so=DESC:sb=docId:po=0:D=EPA-HO-
QRD-2010-0674
U.S. EPA (U.S. Environmental Protection Agency). (201 If). Terminology services (TS): Vocabulary
catalog - IRIS glossary.
http://ofmpub.epa.gov/sor internet/registry/termreg/searchandretrieve/glossariesandke
ywordlists/search.do?details=&glossaryName=IRIS%20Glossary (accessed May 21, 2015).
U.S. EPA (U.S. Environmental Protection Agency). (2012a). 5.2 Dissolved oxygen and biochemical
oxygen demand. In Water Monitoring and Assessment Washington, D.C.
https://archive.epa.gov/water/archive/web/html/vms52.html
11-60
-------
Chapter 11 - References
U.S. EPA (U.S. Environmental Protection Agency). (2012b). 2012 Edition of the drinking water
standards and health advisories [EPA Report], (EPA/822/S-12/001). Washington, D.C.: U.S.
Environmental Protection Agency, Office of Water.
http://www.epa.gov/sites/production/files/2015-09/documents/dwstandards2012.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2012c). Estimation Programs Interface Suite for
Microsoft Windows (EPI Suite) [Computer Program], Washington D.C.
https://www.epa.gov/tsca-screening-tools/epi-suitetm-estimation-program-interface
U.S. EPA (U.S. Environmental Protection Agency). (2012d). Geologic sequestration of carbon
dioxide: Underground injection control (UIC) program class VI well construction guidance
[EPA Report], (EPA 816-R-11-020). Washington, D.C.
http://water.epa.gOv/type/groundwater/uic/class6/upload/epa816rll020.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2012e). Guidance for the determination of
underground sources of drinking water.
https://www.epa.gov/sites/production/files/2015-09/documents/r5-deepwell-
guidance3-determination-underground-sources-drinking-water-19870205.pdf (accessed
January 12, 2015).
U.S. EPA (U.S. Environmental Protection Agency). (2012f). Oil and natural gas sector: Standards of
performance for crude oil and natural gas production, transmission, and distribution.
Background supplemental technical support document for the final new source
performance standards. Washington, D.C. https://www.regulations.gov/document?D=EPA-
HO-QAR-2010-0505-45 50
U.S. EPA (U.S. Environmental Protection Agency). (2012g). Public drinking water systems: Facts and
figures. Washington, D.C.: U.S. Environmental Protection Agency, Office of Water.
http://water.epa.gov/infrastructure/drinkingwater/pws/factoids.cfm
U.S. EPA (U.S. Environmental Protection Agency). (2012h). Study of the potential impacts of
hydraulic fracturing on drinking water resources: Progress report (EPA/601/R-12/011).
Washington, D.C.: U.S. Environmental Protection Agency, Office of Research and
Development. http://nepis.epa.gov/exe/ZyPURL.cgi?Dockey=P100FH8M.txt
U.S. EPA (U.S. Environmental Protection Agency). (2012i). Sustainable futures/pollution prevention
(P2) framework manual [EPA Report], (EPA-748-B12-001). Washington, D.C.: U.S.
Environmental Protection Agency, Office of Pollution Prevention and Toxics.
https://www.epa.gov/sustainable-futures/sustainable-futures-p2-framework-manual
U.S. EPA (U.S. Environmental Protection Agency). (2013a). Data received from oil and gas
exploration and production companies, including hydraulic fracturing service companies
2011 to 2013. Non-confidential business information source documents are located in
Federal Docket ID: EPA-HQ-ORD2010-0674. Available at http: //www.regulations.gov.
U.S. EPA (U.S. Environmental Protection Agency). (2013b). Fiscal year 2011: Drinking water and
groundwater statistics [EPA Report], (EPA 816-R-13-003). Washington, D.C.: U.S.
Environmental Protection Agency, Office of Water.
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P10056TW.txt
U.S. EPA (U.S. Environmental Protection Agency). (2013c). Inventory of U.S. greenhouse gas
emissions and sinks: 1990-2011. (430R13001). Washington, D.C.: U.S. Environmental
Protection Agency, Office of Atmospheric Programs.
https://www3.epa.gov/climatechange/Downloads/ghgemissions/US-GHG-Inventory-
2013-Main-Text.pdf
11-61
-------
Chapter 11 - References
U.S. EPA (U.S. Environmental Protection Agency). (2013d). Supplemental programmatic quality
assurance project plan for work assignment 5-83 technical support for the hydraulic
fracturing drinking water assessment. Washington, D.C.
http://www2.epa.gov/sites/production/files/documents/literature-review-qappl.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2013e). Technical development document for
the coalbed methane (CBM) extraction industry [EPA Report], (EPA-820-R-13-009).
Washington, D.C. https://www.epa.gov/sites/production/files/2015-
06/documents/cbm tdd final 2013.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2013f). Terminology services (TS): Terms and
acronyms.
http://iaspub.epa.gov/sor internet/registrv/termreg/searchandretrieve/termsandacrony
ms/search.do
U.S. EPA (U.S. Environmental Protection Agency). (2014a). Drinking water contaminants.
http: //water.epa. gov/drink/contaminants /
U.S. EPA (U.S. Environmental Protection Agency). (2014b). Greenhouse gas reporting program,
Subpart W - Petroleum and natural gas systems.
https://www.epa.gov/ghgreporting/subpart-w-petroleum-and-natural-gas-systems
U.S. EPA (U.S. Environmental Protection Agency). (2014c). Minimizing and managing potential
impacts of injection-induced seismicity from class II disposal wells: Practical approaches
[EPA Report], Washington, D.C. https://www.epa.gov/sites/production/files/2015-
08/documents/induced-seismicity-2 015 02.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2014d). Quality assurance project plan - Revision
no. 2: Data and literature evaluation for the EPA's study of the potential impacts of
hydraulic fracturing (HF) on drinking water resources [EPA Report], Washington, D.C.
https://www.epa.gov/hfstudy/qapp-revision-no-2-data-and-literature-evaluation-epas-
studv-potential-impacts-hydraulic
U.S. EPA (U.S. Environmental Protection Agency). (2014e). Quality management plan- Revision no.
2: Study of the potential impacts of hydraulic fracturing for oil and gas on drinking water
resources [EPA Report], Washington, D.C. http ://www2 .epa.gov/hfstudy/qualitv-
management-plan-revision-no-2-study-potential-impacts-hydraulic-fracturing-oil-and
U.S. EPA (U.S. Environmental Protection Agency). (2014f). Retrospective case study in northeastern
Pennsylvania: Study of the potential impacts of hydraulic fracturing on drinking water
resources [EPA Report], (EPA 600/R-14/088). Washington, D.C.
http://www2.epa.gov/hfstudv/retrospective-case-study-northeastern-pennsylvania
U.S. EPA (U.S. Environmental Protection Agency). (2014g). Retrospective case study in the Raton
Basin, Colorado: Study of the potential impacts of hydraulic fracturing on drinking water
resources [EPA Report], (EPA 600/R-14/091). Washington, D.C.
U.S. EPA (U.S. Environmental Protection Agency). (2014h). Safe drinking water information system
(SDWIS). Data obtained from the Office of Water [Database], Washington, D.C.: U.S.
Environmental Protection Agency, Office of Water.
https://www3.epa.gov/enviro/facts/sdwis/search.html
U.S. EPA (U.S. Environmental Protection Agency). (2014i). The verification of a method for detecting
and quantifying diethylene glycol, triethylene glycol, tetraethylene glycol, 2-butoxyethanol
and 2-methoxyethanol in ground and surface waters [EPA Report], (EPA/600/R-14/008).
Washington, D.C.: U.S. Environmental Protection Agency, Office of Research and
11-62
-------
Chapter 11 - References
Development, http://www2.epa.gov /hfstudv/verification-method-detecting-and-
quantifying-diethvlene-glvcol-triethvlene-glycol
U.S. EPA (U.S. Environmental Protection Agency). (2015a). Analysis of hydraulic fracturing fluid
data from the FracFocus chemical disclosure registry 1.0 [EPA Report], (EPA/601/R-
14/003). Washington, D.C.: U.S. Environmental Protection Agency, Office of Research and
Development, http: / /www2.epa.gov/hfstudv/analysis-hydraulic-fracturing-fluid-data-
fracfocus-chemical-disclosure-registry-l-pdf
U.S. EPA (U.S. Environmental Protection Agency). (2015b). Analysis of hydraulic fracturing fluid
data from the FracFocus chemical disclosure registry 1.0: Data management and quality
assessment report [EPA Report], (EPA/601/R-14/006). Washington, D.C.: U.S.
Environmental Protection Agency, Office of Research and Development
http ://www2 .epa. gov/sites /production/files /2 015-
03/documents/fracfocus data management report final 032015 508.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2015c). Analysis of hydraulic fracturing fluid
data from the FracFocus chemical disclosure registry 1.0: Project database [EPA Report],
(EPA/601/R-14/003). Washington, D.C.: U.S. Environmental Protection Agency, Office of
Research and Development, http: //www2.epa.gov/hfstudv/epa-proiect-database-
developed-fracfocus-l-disclosures
U.S. EPA (U.S. Environmental Protection Agency). (2015d). Assessment of the potential impacts of
hydraulic fracturing for oil and gas on drinking water resources: External review draft [EPA
Report], (EPA/600/R-15/047a). Washington, D.C.: U.S. Environmental Protection AGENCY,
Office of Research and Development
https://cfpub.epa. gov/ncea/hfstudv/recordisplay.cfm?deid=244651
U.S. EPA (U.S. Environmental Protection Agency). (2015e). Case study analysis of the impacts of
water acquisition for hydraulic fracturing on local water availability [EPA Report],
(EPA/600/R-14/179). Washington, D.C.: U.S. Environmental Protection Agency, Office of
Research and Development, https: //www.epa.gov/hfstudv/case-study-analysis-impacts-
water-acquisition-hydraulic-fracturing-local-water-availabilitv
U.S. EPA (U.S. Environmental Protection Agency). (2015f). DMR spreadsheet Pennsylvania
wastewater treatment plants per Region 3 Information Request Data provided by request
Washington, D.C.: Region 3, U.S. Environmental Protection Agency.
U.S. EPA (U.S. Environmental Protection Agency). (2015g). Effluent data from Pennsylvania
wastewater treatment plants per Region 3 Information Request Data provided by request
Washington, D.C.: Region 3, U.S. Environmental Protection Agency.
U.S. EPA (U.S. Environmental Protection Agency). (2015h). Inventory of U.S. greenhouse gas
emissions and sinks: 1990-2013. (EPA 430-R-15-004). Washington, D.C.
https://www3.epa.gov/climatechange/Downloads/ghgemissions/US-GHG-Inventory-
2015-Main-Text.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2015i). Retrospective case study in Killdeer,
North Dakota: Study of the potential impacts of hydraulic fracturing on drinking water
resources [EPA Report], (EPA 600/R-14/103). Washington, D.C.
http://www2.epa.gov/hfstudv/retrospective-case-study-killdeer-north-dakota
U.S. EPA (U.S. Environmental Protection Agency). (2015j). Retrospective case study in
southwestern Pennsylvania: Study of the potential impacts of hydraulic fracturing on
drinking water resources [EPA Report], (EPA 600/R-14/084). Washington, D.C.
http://www2.epa.gov/hfstudv/retrospective-case-study-southwestern-pennsylvania
11-63
-------
Chapter 11 - References
U.S. EPA (U.S. Environmental Protection Agency). (2015k). Retrospective case study in the Raton
Basin, Colorado: Study of the potential impacts of hydraulic fracturing on drinking water
resources [EPA Report], (EPA 600/R-14/091). Washington, D.C.
http://www2.epa.gov/hfstudy/retrospective-case-study-raton-basin-colorado
U.S. EPA (U.S. Environmental Protection Agency). (20151). Retrospective case study in Wise County,
Texas: Study of the potential impacts of hydraulic fracturing on drinking water resources
[EPA Report], (EPA 600/R-14/090). Washington, D.C.
http://www2.epa.gov/hfstudv/retrospective-case-study-wise-county-texas
U.S. EPA (U.S. Environmental Protection Agency). (2015m). Review of state and industry spill data:
Characterization of hydraulic fracturing-related spills [EPA Report], (EPA/601/R-14/001).
Washington, D.C.: U.S. Environmental Protection Agency, Office of Research and
Development, http: //www2. epa. gov /hfstudv/r e vie w- state - and- industry- spill- data-
characterization-hydraulic-fracturing-related-spills-1
U.S. EPA (U.S. Environmental Protection Agency). (2015n). Review of well operator files for
hydraulically fractured oil and gas production wells: Well design and construction [EPA
Report], (EPA/601/R-14/002). Washington, D.C.: U.S. Environmental Protection Agency,
Office of Research and Development, http://www2.epa.gov/hfstudy/review-well-operator-
files-hydraulically-fractured-oil-and-gas-production-wells-well-design
U.S. EPA (U.S. Environmental Protection Agency). (2015o). Sources contributing bromide and
inorganic species to drinking water intakes on the Allegheny River in western Pennsylvania
[EPA Report], (EPA/600/R-14/430). Washington, D.C.
https://www.epa.gov/hfstudy/sources-contributing-inorganic-species-drinking-water-
intakes-during-low-flow-conditions
U.S. EPA (U.S. Environmental Protection Agency). (2015p). WaterSense: Water supply in the U.S.
http://www.epa.gov/WaterSense/pubs/supply.html (accessed January 12, 2015).
U.S. EPA (U.S. Environmental Protection Agency). (2016a). Human health risk assessment
https://www.epa.gov/risk/human-health-risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2016b). Regional screening levels for chemical
contaminants at superfund sites, https://www.epa.gov/risk/regional-screening-levels-rsls
U.S. EPA (U.S. Environmental Protection Agency). (2016c). Review of well operator files for
hydraulically fractured oil and gas production wells: Hydraulic fracturing operations [EPA
Report], (EPA/601/R-14/004). Washington, D.C.: U.S. Environmental Protection Agency,
Office of Research and Development, https: //www.epa.gov/sites/production/files/2016-
07/documents/wfr2 final 07-28-16 508.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2016d). Technical development document for
the effluent limitations guidelines and standards for the oil and gas extraction point source
category [EPAReport], (EPA-820-R-16-003). Washington, D.C.: U.S. Environmental
Protection Agency, Office of Water.
U.S. EPA (U.S. Environmental Protection Agency). (2016e). XTO Energy, Inc. Settlement
http://www2.epa.gov/enforcement/xto-energy-inc-settlement
U.S. EPA. USGS (U.S. Environmental Protection Agency, U.S. Geological Survey). (2016f). National
hydrography datasetplus - NHDPlus edition 2.10.
https://www.epa.gov/waterdata/nhdplus-national-hydrography-dataset-plus
U.S. FDA. HHS (U.S. Food & Drug Administration, U.S. Department of Health and Human Services)
(2016). Food for human consumption: Indirect food substances affirmed as generally
11-64
-------
Chapter 11 - References
recognized as safe: Listing of specific substances affirmed as GRAS: Formic acid, 21 CFR
http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=186.1316
U.S. GAP (U.S. Government Accountability Office). (2012). Energy-water nexus: Information on the
quantity, quality, and management of water produced during oil and gas production. (GAO-
12-156). Washington, D.C. http: / /www,gao.gov/products/GAO-12-156
U.S. GAP (U.S. Government Accountability Office). (2014). Freshwater: Supply concerns continue,
and uncertainties complicate planning. Report to Congressional requesters. (GAO-14-430).
Washington, D.C.: U.S. Government Accountability Office (GAO).
http://www.gao.gov/assets/670/663343.pdf
Upstream Pumping. (2015). Upstream pumping: Wellhead technology and services.
http://upstreampumping.com/
URS Corporation. (2011). Water-related issues associated with gas production in the Marcellus
Shale: Additives use; Flowback quality and quantities; Regulations; On-site treatment;
Green technologies; Alternate water sources; Water well-testing. (NYSERDA Contract PO
Number 10666). Fort Washington, PA.
USFWS (U.S. Fish and Wildlife Service). (2014). Migratory bird threats by surfactants and other
contaminants in evaporation ponds associated with oilfield wastewater disposal facilities.
(Environmental Contaminants Report Number: R6/729C/14). Cheyenne, Wyoming: U.S.
Fish and Wildlife Service, Region 6. https: //www.fws.gov/mountain-
prairie/contaminants /documents ZR6729C14.pdf
USGS (U.S. Geological Survey). (1999). Naturally occurring radioactive materials (NORM) in
produced water and oil-field equipment An issue for the energy industry [Fact Sheet], (FS-
142-99). Denver, CO: USGS (U.S. Geological Survey).
https://pubs.er.usgs.gov/publication/fsl4299
USGS (U.S. Geological Survey). (2000). Coal-bed methane: Potential and concerns [Fact Sheet], (Fact
Sheet 123-00). Denver, CO. http://pubs.usgs.gov/fs/fsl23-00/fsl23-00.pdf
USGS (U.S. Geological Survey). (2006). Produced Water Database [Database]: U.S. Geological Survey:
USGS.
http://energy.usgs.gov/EnvironmentalAspects/EnvironmentalAspectsofEnergyProduction
andUse /ProducedWaters.aspx#3 822110-overview
USGS (U.S. Geological Survey). (2009). Ground water atlas of the United States.
http://pubs.usgs.gov/ha/ha730/
USGS (U.S. Geological Survey). (2010). Volatile organic compounds in the nation's ground water and
drinking-water supply wells: Supporting information. Glossary.
http://water.usgs.gov/nawqa/vocs/national assessment/report/glossary.html
USGS (U.S. Geological Survey). (2012). Source Water-Quality Assessment (SWQA) Program.
http: //water.usgs. gov/nawqa/swqa/
USGS (U.S. Geological Survey). (2013a). Map of assessed shale gas in the United States, 2012: U.S.
Geological Survey Digital Data Series (GSDDS 69-Z). Reston, VA: USGS (U.S. Geological
Survey National Assessment of Oil and Gas Resources Team, and Biewick, L.R.H., compiler).
http://pubs.usgs.gov/dds/dds-069/dds-069-z/
USGS (U.S. Geological Survey). (2013b). National hydrography dataset: High-resolution flowline
data: The national map. http://viewer.nationalmap.gov/viewer/
11-65
-------
Chapter 11 - References
USGS (U.S. Geological Survey). (2014a). Energy glossary and acronym list t
http://energy.usgs.gov/GeneralInfo/HelpfulResources/EnergyGlossary.aspx
USGS (U.S. Geological Survey). (2014b). Estimated use of water in the United States, county-level
data for 2010. Reston, VA. http://water.usgs.gov/watuse/data/201Q/
USGS (U.S. Geological Survey). (2014c). WaterWatch: Computed runoff in hydrologic units.
http://waterwatch.usgs. gov/index.php?id=romap3
USGS (U.S. Geological Survey). (2016). How much of your state is wet?
http: //water.usgs. gov/edu/wetstates.html
Utah Division of Oil. Gas and Mining (2016). Utah Division of Oil, Gas and Mining approved
commercial disposal facilities.
https://oilgas.ogm.utah.gov/pub/Publications/Lists/disposal facilities locations.pdf
UWS (Universal Well Services). (2008). Environmental response plan for field operations. (PATT-
EPA-0001060). Meadville, PA: Universal Well Services, Inc.
Vaidvanathan. G. (2013a). Hydraulic fracturing: When 2 wells meet, spills can often follow.
http://www.eenews.net/energywire/stories/1059985587 (accessed September 2, 2014).
Vaidvanathan. G. (2013b). XTO comes out swinging against 'unwarranted' criminal charges in PA
[Magazine], E&E News, Sept 12, 2013. http://www.eenews.net/stories/1059987124
Vaidvanathan. G. (2014). Email communications between Gayathri Vaidyanathan and Ken Klewicki
regarding the New Mexico Oil Conservation Division District 3 Well Communication Data.
van Vliet. MTH: Zwolsman. TIG. (2008). Impact of summer droughts on the water quality of the
Meuse river. J Hydrol 353: 1-17. http://dx.doi.Org/10.1016/j.jhydrol.2008.01.001
Van Voast. WA. (2003). Geochemical signature of formation waters associated with coalbed
methane. AAPG Bulletin 87: 667-676.
Veil. TA: Puder. MG: Elock. D: Redweik. RT. (2004). A white paper describing produced water from
production of crude oil, natural gas, and coal bed methane. Morgantown, WV: Department
of Energy (DOE), National Energy Technology Laboratory (NETL).
http://www.ipd.anl.goy/anlpubs/2004/02/49109.pdf
Veil. I. (2015). U.S. Produced water volumes and management practices in 2012. Oklahoma City, OK:
Ground Water Protection Council.
http://www.gwpc.org/sites/default/files/Produced%20Water%20Report%202014-
GWPC O.pdf
Veil. TA. (2011). Water management practices used by Fayetteville shale gas producers.
(ANL/EVS/R-11/5). Washington, D.C.: U.S. Department of Energy, National Energy
Technology Laboratory, http://www.ipd.anl.gov/anlpubs/2011/06/70192.pdf
Veil. TA: Puder. MG: Elcock. D: Redweik. RT. (2004). A white paper describing produced water from
production of crude oil, natural gas, and coalbed methane. Lemont, IL: U.S. Department of
Energy, National Energy Technology Laboratory; and Argonne National Laboratory.
http://www.ipd.anl.gOv/anlpubs/2004/02/49109.pdf
Veloski. GA: Sams. I: Hammack. RW. (2015). Rapid methods for locating existing well penetrations
in unconventional well development areas of Pennsylvania. Presented at SPE
Unconventional Resources Technology Conference, 20-22 July, 2015, San Antonio, TX:
Society of Petroleum Engineers. http: //dx.doi.org/10.2118/178558-MS
11-66
-------
Chapter 11 - References
Vengosh. A: Tackson. RB: Warner. N: Darrah. TH: Kondash. A. (2014). A critical review of the risks to
water resources from unconventional shale gas development and hydraulic fracturing in the
United States. Environ SciTechnol 48: 36-52. http://dx.doi.org/10.1021/es4Q5118y
Verdegem. MCI: Bosma. RH. (2009). Water withdrawal for brackish and inland aquaculture, and
options to produce more fish in ponds with present water use. Water Policy 11: 52-68.
http://dx.doi.org/10.2166/wp.2009.0Q3
Vesterkiaer. R. (2002) Is it necessary to install a downhole safety valve in a subsea oil/gas well?: A
comparison of the risk level of a well with, and without a safety valve. (Master's Thesis).
Norwegian University of Science and Technology, Trondheim, Norway. Retrieved from
http://frigg.ivtntnu.no/ross/reports/stud/vesterkjaer.pdf
Vidic. RD: Brantley. SL: Vandenbossche. TM: Yoxtheimer. D: Abad. ID. (2013). Impact of shale gas
development on regional water quality [Review], Science 340: 1235009.
http://dx.doi.org/10.1126/science.1235009
Vinal. I. (2015). Time-lapse seismic monitoring of overburden compaction due to hydraulic
fracturing in the Montney shale. SPE EUR0PEC 2015/77th EAGE Conference and Exhibition,
June 1-4, 2015, Madrid, Spain.
http://www.earthdoc.org/publication/publicationdetails/?publication=81177
Vinal. I: Davis. T. (2015). Surface time-lapse multi-component seismology- A new technology for
hydraulic fracture monitoring? A Montney Shale gas case study. First Break 33: 65-70.
Vincent. M. (2011). Restimulation of unconventional reservoirs: When are refracs beneficial?
Journal of Canadian Petroleum Technology 50: 36-52. http: / /dx. doi. o r g/10.2118 /13 675 7-
PA
Vittitow. 1G. Sr. (2010). Well control incident analysis, EOG Resources Inc., Punxautawney Hunting
Club36H, Clearfield County, Pennsylvania. Bedrock Engineering.
http://www.sootypaws.net/gws/class/study docs/36H/2.BEDROCK ENGINEERING PHC 3
6H Incident Report.Final.709.pdf
Volz. CD: Ferrar. K: Michanowicz. D: Christen. C: Kearney. S: Kelso. M: Malone. S. (2011).
Contaminant characterization of effluent from Pennsylvania Brine Treatment Inc., Josephine
facility being released into Blacklick Creek, Indiana County, Pennsylvania: Implications for
disposal of oil and gas flowback fluids from brine treatment plants [Standard],
http://www2.epa.gov/hfstudy/contaminant-characterization-effluent-pennsylvania-brine-
treatment-inc-iosephine-facility
Vulgamore. TB: Clawson. TP: Pope. CD: Wolhart. SL: Maverhofer. MI: Machovoe. SR: Waltman. CK.
(2007). Applying hydraulic fracture diagnostics to optimize stimulations in the Woodford
Shale. Richardson, TX: Society of Petroleum Engineers, http: //dx.doi.org/10.2118/110029-
MS
Waldron. P. (2014). In search of greener fracking for natural gas.
http://www.insidescience.org/content/search-greener-fracking-natural-gas/1791
Walsh. FR. Ill: Zoback. MP. (2015). Oklahoma's recent earthquakes and saltwater disposal. Science
Advances 1: el500195. http ://dx. doi. org/10.1126/sciadv. 1500195
Walsh. TM. (2013). Water management for hydraulic fracturing in unconventional resources: Part 1.
Oil and Gas Facilities 2. https: //www.spe.org/en/ogf/ogf-article-detail/?art=422
Walter. GR: Benke. RR: Pickett. PA. (2012). Effect of biogas generation on radon emissions from
landfills receiving radium-bearing waste from shale gas development J Air Waste Manag
Assoc 62: 1040-1049. http://dx.doi.org/10.1080/10962247.2012.696084
11-67
-------
Chapter 11 - References
Wang. T: Rahman. SS. (2015). An investigation of fluid leak-off due to osmotic and capillary effects
and its impact on micro-fracture generation during hydraulic fracturing stimulation of gas
shale. In EUROPEC 2015,1-4 June, 2015, Madrid, Spain: Society of Petroleum Engineers.
http://dx.d0i.0rg/l 0.2118/174392-MS
Wang. W: Dahi Taleghani. A. (2014). Cement sheath integrity during hydraulic fracturing; an
integrated modeling approach. In 2014 SPE Hydraulic Fracturing Technology Conference, 4-
6 February, 2014, The Woodlands, TX: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/168642-MS
Wang. Z: Krupnick. A. (2013). A retrospective review of shale gas development in the United States.
What led to the boom? (RFF DP 13-12). Washington, D.C.: Resources for the Future.
http://www.rff.org/RFF/documents/RFF-DP-13-12.pdf
Warner. NR: Christie. CA: Tackson. RB: Vengosh. A. (2013a). Impacts of shale gas wastewater
disposal on water quality in western Pennsylvania. Environ Sci Technol 47: 11849-11857.
http://dx.doi.org/10.1021/es402165b
Warner. NR: Kresse. TM: Hays. PD: Down. A: Karr. ID: Tackson. RB: Vengosh. A. (2013b).
Geochemical and isotopic variations in shallow groundwater in areas of the Fayetteville
Shale development, north-central Arkansas. Appl Geochem 35: 207-220.
Warner. NR: Darrah. TH: Tackson. RB: Millot. R: Kloppmann. W: Vengosh. A. (2014a). New tracers
identify hydraulic fracturing fluids and accidental releases from oil and gas operations.
Environ Sci Technol 48: 12552-12560. http://dx.doi.org/10.1021/es5Q32135
Warner. NR: Darrah. TH: Tackson. RB: Millot. R: Kloppmann. W: Vengosh. A. (2014b). New tracers
identify hydraulic fracturing fluids and accidental releases from oil and gas operations:
Supplemental materials [Supplemental Data], Environ Sci Technol 48: 12552-12560.
Warner. NR: Tackson. RB: Darrah. TH: Osborn. SG: Down. A: Zhao. K: White. A: Vengosh. A. (2012).
Geochemical evidence for possible natural migration of Marcellus Formation brine to
shallow aquifers in Pennsylvania. Proc Natl Acad Sci USA 109: 11961-11966.
http://dx.doi.org/10.1073/pnas.1121181109
Warpinski. N. (2009). Microseismic monitoring: Inside and out. J PetTech 61: 80-85.
http://dx.d0i.0rg/l 0.2118/118537-TPT
Water Research Foundation. (2010). Assessment of inorganics accumulation in drinking water
system scales and sediments. Denver, CO.
http://www.waterrf.org/PublicReportLibrarv/3118.pdf
Water Research Foundation. (2014). Water quality impacts of extreme weather-related events.
Denver, CO. http://www.waterrf.org/Pages/Proiects.aspx?PID=4324
Watson. TL: Bachu. S. (2009). Evaluation of the potential for gas and C02 leakage along wellbores.
SPE Drilling & Completion 24: 115-126. http://dx.doi.org/10.2118/106817-PA
Wattenberg. EV: Bielicki. TM: Suchomel. AE: Sweet. IT: Void. EM: Ramachandran. G. (2015).
Assessment of the acute and chronic health hazards of hydraulic fracturing fluids. J Occup
Environ Hyg 12: 611-624. http://dx.doi.org/10.1080/15459624.2015.1029612
Watts. KR. (2006). A preliminary evaluation of vertical separation between production intervals of
coalbed-methane wells and water-supply wells in the Raton Basin, Huerfano and Las
Animas Counties, Colorado, 1999-2004. 15.
11-68
-------
Chapter 11 - References
WAWSA (Western Area Water Supply Authority). (2011). Project progress report: Western area
water supply project: Appendix N. Williston, ND. http: / /www.legis.nd.gov/assemblv/6 2-
2 011 /docs/pd f/wrl 01011 appe nd ixn. pd f
WDEO (Wyoming Department of Environmental Quality). (2015). Pavillion, Wyoming area
domestic water wells draft final report and palatability study.
http://deq.wyoming.gov/media/attachments/Water%200uality/Pavillion%20lnvestigatio
n / D ra ft% 2 0 Re po rt / 01 Pavillion%20WY%20Area%20Domestic%20Water%20Wells%20D
ra ft% 2 0 Fi na 1% 2 0 Re po rt. pd f
WDEO (Wyoming Department of Environmental Quality). (2016a). Factsheet#13: Groundwater
cleanup levels. WYDEQ Voluntary Remediation Program.
http://deq.wyoming.gov/media/attachments/Solid%20%26%20Hazardous%20Waste/Vol
untary%20Remediation%20Program/Fact%20Sheets/FS 13 whqqBPe.pdf
WDEO (Wyoming Department of Environmental Quality). (2016b). Permitted commercial oil
disposal facilities.
http://deq. wvoming. gov/media/attachments/Water%20Oualitv/Permitting/Produced%2
0Water%20Disposal%20%26%20Treatment/Active-CQWDF-Listpdf
Weaver. 1W: Xu. 1: Mravik. SC. (2016). Scenario analysis of the impact on drinking water intakes
from bromide in the discharge of treated oil and gas wastewater. J Environ Eng 142:
04015050. http://dx.doi.Org/10.1061/fASCElEE.1943-7870.0000968
Weiiermars. R. (2014). US shale gas production outlook based on well roll-out rate scenarios. Appl
Energ 124: 283-297. http://dx.doi.Org/10.1016/j.apenergy.2014.02.058
Weingarten. M: Ge. S: Godt. 1W: Bekins. BA: Rubinstein. 1L. (2015). Induced seismicity. High-rate
injection is associated with the increase in U.S. mid-continent seismicity. Science 348: 1336-
1340. http://dx.doi.org/10.ri26/science.aab 1345
Weinhold. B. (2012). The future of fracking: New rules target air emissions for cleaner natural gas
production. Environ Health Perspect 120: a272-a279. http://dx.doi.org/10.1289/ehp.120-
a272
Weinstein. 1: Phillippi. M: Walters. HG. (2009). Dry-polymer blending eliminates need for
hydrocarbon carrier fluids. In 2009 SPE/EPA/DOE E&P Environmental & Safety Conference,
23-25 March, 2009, San Antonio, TX: Society of Petroleum Engineers.
http://dx.doi.org/10.2118/121002-MS
Wen. T: Castro. MC: Nicot. IP: Hall. CM: Larson. T: Mickler. P: Darvari. R. (2016). Methane sources
and migration mechanisms in shallow groundwaters in Parker and Hood counties, Texas-A
heavy noble gas analysis. Environ Sci Technol 50: 12012-12021.
http://dx.doi.org/10.1021/acs.est.6b01494
Wendel. K. (2011). Wastewater technologies critical for continued growth of Marcellus.
http://www.ogfi.com/articles/print/volume-8/issiie-ri/featiires/wastewater-
technologies-critical-for.html (accessed March 9, 2015).
Weng. X: Kresse. 0: Cohen. C: Wu. R: Gu. H. (2011). Modeling of hydraulic fracture network
propagation in a naturally fractured formation. Paper presented at SPE Hydraulic
Fracturing Technology Conference, January 24-26, 2011, The Woodlands, TX.
http://www.onepetro.org/mslib/app/Preview.do?paperNumber=SPE-140253-
PA&societyCode=SPE
Wertz. I. (2014). Fracking site operator faces contempt complaint after acid spill.
http://stateimpactnpr.org/oklahoma/2014/08/14/fracking-site-operator-faces-contempt-
complaint-after-acid-spill/
Wess. I: Ahlers. H: Dobson. S. (1998). Concise international chemical assessment document 10: 2-
butoxyethanol. (Document 10). Geneva, Switzerland: World Health Organization.
http://www.who.int/ipcs/publications/cicad/cicad 10 revised.pdf
11-69
-------
Chapter 11 - References
West Virginia PEP (West Virginia Department of Environmental Protection). (2011). Memorandum
of agreement from the Division of Water and Waste Management to the Division of
Highway: WVDOH/WVDEP salt brine from gas wells agreement.
http://www.dep.wv.gOv/WWE/Documents/WVD0HWVDEP%20Salt%20Brine%20Agreem
entpdf
West Virginia PEP (West Virginia Department of Environmental Protection). (2013). West Virginia
water resources management plan. (Article 22-26). Charleston, WV.
http://www.dep.wv.gOv/WWE/wateruse/WVWaterPlan/Documents/WV WRMP.pdf
West Virginia PEP (West Virginia Department of Environmental Protection). (2014). Personal
communication: email from Jason Harmon, West Virginia PEP to Megan Fleming, U.S. EPA
with attachment of WV PEP fracturing water database.
Westerhoff. P: Chao. P: Mash. H. (2004). Reactivity of natural organic matter with aqueous chlorine
and bromine. Water Res 38: 1502-1513. http://dx.doi.Org/10.1016/j.watres.2003.12.014
Whittemore. PP. (2007). Fate and identification of oil-brine contamination in different
hydrogeologic settings. Appl Geochem 22: 2099-2114.
http://dx.doi.Org/10.1016/j.apgeochem.2007.04.002
WHO (World Health Organization). (2015). Concise international chemical assessment documents.
http://www.who.int/ipcs/publications/cicad/en/
Williamson. P. (2013). Prilling fluid basics [Magazine], Oilfield Rev, 25, 63-64.
Wilson. B. (2014). Geologic and baseline groundwater evidence for naturally occurring, shallowly
sourced, thermogenic gas in northeastern Pennsylvania. AAPG Bulletin 98: 373-394.
http://dx.d0i.0rg/l 0.1306/08061312218
Wilson. TM: Van Briesen. TM. (2013). Source water changes and energy extraction activities in the
Monongahela River, 2009-2012. Environ Sci Technol 47: 1257512582.
http://dx.doi.org/10.1021/es402437n
Wilson. TM: Vanbriesen. TM. (2012). Oil and gas produced water management and surface drinking
water sources in Pennsylvania. Environmental Practice 14: 288-300.
Winter. TC: Harvey. TW: Franke. PL: Alley. WM. (1998). Ground water and surface water: A single
resource. (USGS Circular 1139). Penver, CO: U.S. Pepartment of the Interior, U.S. Geological
Survey, http://pubs.usgs.gov/circ/circll39/pdf/circll39.pdf
Wirfs-Brock. 1. (2015). Wastewater spills in North Pakota: Whatthe data tell us.
http://insideenergy.org/2015/01/3Q/wastewater-spills-in-north-dakota-what-the-data-
tell-us/
Woitanowicz. AK. (2008). Environmental control of well integrity. In ST Orszulik (Ed.),
Environmental technology in the oil industry (pp. 53-75). Houten, Netherlands: Springer
Netherlands.
Wolfe. P: Graham. G. (2002). Water rights and beneficial use of produced water in Colorado.
Penver, CO: American Water Resources Association.
http://www.gwpc.org/sites/default/files/event-sessions/Pick Wolfe PWC02 O.pdf
Wong. I: Nemser. E: Bott. I: Pober. M. (2015). Induced seismicity and traffic light systems as related
to hydraulic fracturing in Ohio. Columbus, OH: Ohio Oil and Gas Association.
http://c.ymcdn.com/sites/www.ooga.org/resource/resmgr/Files/OOGA IS TLS White Pap
er fina.pdf
11-70
-------
Chapter 11 - References
Wright. PR: McMahon. PB: Mueller. DK: Clark. ML. (2012). Groundwater-quality and quality-control
data for two monitoring wells near Pavillion, Wyoming, April and May 2012. (USGS Data
Series 718). Reston, Virginia: U.S. Geological Survey.
http://pubs.usgs.gov/ds/718/DS718 508.pdf
Wuchter. C: Banning. E: Mincer. TI: Drenzek. NI: Coolen. Mi. (2013). Microbial diversity and
methanogenic activity of Antrim Shale formation waters from recently fractured wells.
FMICB 4: 1-14. http://dx.doi.org/10.3389/fmich.2013.00367
Wunz. EC. (2015). Summary of shale gas wastewater treatment and disposal in Pennsylvania 2014.
State College, PA: Ben Franklin Shale Gas Innovation & Commercialization Center.
http://www.sgicc.Org/iiploads/8/4/3/l/843ri64/marcelliis shale wastewater mgmt rep
ort 2015 update.pdf
WYOGCC (Wyoming Oil and Gas Conservation Commission). (2014a). Pavillion Field Pit Review.
Casper, Wyoming.
http://wogcc.state.wv.us/pavillionworkinggrp/Draft%20Pavillion%20Field%20Pits%20Re
view 11242014.pdf
WYOGCC (Wyoming Oil and Gas Conservation Commission). (2014b). Pavillion Field Well Integrity
Review. Casper, Wyoming.
http://wogcc.state.wy.us/pavillionworkinggrp/PAVILLION REPORT 1082014 Final Repor
tpdf
WYOGCC (Wyoming Oil and Gas Conservation Commission). (2015). Environmental rules, including
underground injection control program rules for enhanced recovery and disposal projects.
Casper, WY.
Wyoming State Engineer's Office. (2014). Groundwater control areas and advisory boards.
http://seo.wvo.gov/ground-water/groundwater-control-areas-advisorv-boards
Xu. B: Hill. AD: Zhu. D: Wang. L. (2011). Experimental evaluation of guar fracture fluid filter cake
behavior. Paper presented at SPE Hydraulic Fracturing Technology Conference, January 24-
26, 2011, The Woodlands, TX.
http://www.onepetro.org/mslib/app/Preview.do?paperNumber=SPE-140686-
MS&societvCode=SPE
Xu. P: Drewes. IE: Heil. D. (2008). Beneficial use of co-produced water through membrane
treatment: Technical-economic assessment Desalination 225: 139-155.
http://dx.doi.Org/10.1016/i.desal.2007.04.093
Yang. L: Grossmann. IE: Manno. I. (2014). Optimization models for shale gas water management
AlChE j 60: 34903501. http: //dx.doi.org/10.1002/aic.l 4526
Yang. X: Shang. C. (2004). Chlorination byproduct formation in the presence of humic acid, model
nitrogenous organic compounds, ammonia, and bromide. Environ Sci Technol 38: 4995-
5001. http://dx.doi.org/10.1021/es049580g
Yang. Y: Robart. CI: Ruegamer. M. (2013). Analysis of U.S. hydraulic fracturing fluid system trends.
In 2013 SPE Hydraulic Fracturing Technology Conference, 4-6 February, 2013, The
Woodlands, TX: Society of Petroleum Engineers, https: //www.onepetro.org/conference-
paper/SPE-163875-MS
Yeager. RR: Bailey. DE. (2013). Diesel-based gel concentrate improves Rocky Mountain region
fracture treatments. Richardson, TX: Society of Petroleum Engineers,
http://dx.doi.org/10.2118/17535-MS
Yergin. D. (2011). The quest: Energy, security and the remaking of the modern world. New York,
NY: Penquin Press.
Ying. L: O'Connor. F: Stolz. IF. (2015). Scintillation gamma spectrometer for analysis of hydraulic
fracturing waste products. J Environ Sci Health A Tox Hazard Subst Environ Eng 50: 511-
515. http://dx.doi.Org/10.1080/10934529.2015.992682
11-71
-------
Chapter 11 - References
Yost. EE: Stanek. T: Burgoon. LP. (In Press) A decision analysis framework for estimating the
potential hazards for drinking water resources of chemicals used in hydraulic fracturing
fluids. Sci Total Environ, http://dx.doi.org/10.1016/i.scitotenv.2016.08.167
Yost. EE: Stanek. 1: Dewoskin. RS: Burgoon. LP. (2016a). Estimating the potential toxicity of
chemicals associated with hydraulic fracturing operations using quantitative structure-
activity relationship modeling. Environ SciTechnol 50: 7732-7742.
http://dx.doi.org/10.1021/acs.est5b05327
Yost. EE: Stanek. 1: Dewoskin. RS: Burgoon. LP. (2016b). Overview of chronic oral toxicity values for
chemicals present in hydraulic fracturing fluids, flowback, and produced waters. Environ Sci
Technol 50: 4788-4797. http://dx.doi.org/10.1021/acs.est5b04645
Younos. T: Tulou. KE. (2005). Overview of desalination techniques. Journal of Contemporary Water
Research & Education 132: 3-10. http://dx.doi.org/10.ll 11/j. 1936-
704X.2005.mpl 32001002.x
Zhang. L: Anderson. N: Dilmore. R: Soeder. DT: Bromhal. G. (2014a). Leakage detection of Marcellus
Shale natural gas at an Upper Devonian gas monitoring well: A 3-d numerical modeling
approach. Environ Sci Technol 48: 10795-10803. http://dx.doi.org/10.1021/es501997p
Zhang. T: Gregory. K: Hammack. RW: Vidic. RD. (2014b). Co-precipitation of radium with barium
and strontium sulfate and its impact on the fate of radium during treatment of produced
water from unconventional gas extraction. Environ Sci Technol 48: 4596-4603.
http://dx.doi.org/10.1021/es405168b
Zhang. T: Bain. D: Hammack. R: Vidic. RD. (2015a). Analysis of radium-226 in high salinity
wastewater from unconventional gas extraction by inductively coupled plasma-mass
spectrometry. Environ Sci Technol 49: 2969-2976. http://dx.doi.org/10.1021/es504656q
Zhang. T: Hammack. RW: Vidic. RD. (2015b). Fate of radium in Marcellus shale flowback water
impoundments and assessment of associated health risks. Environ Sci Technol 49: 9347-
9354. http://dx.doi.org/10.1021/acs.est5b01393
Zhang. L: Soeder. DT. (2016). Modeling of methane migration in shallow aquifers from shale gas well
drilling. Ground Water 54: 345-353. http://dx.doi.org/10.llll/gwatl2361
Zhou. Z. (2016) The impact of capillary imbibition and osmosis during hydraulic fracturing of shale
formations. (Doctoral Dissertation). Colorado School of Mines, Golden, CO.
https://dspace.library.colostate.edu/bitstream/handle/11124/166681/ZH0U mines 0052
E 10883.pdf?sequence=l&isAllowed=y
Ziemkiewicz. P: Donovan. 1: Hause. 1: Gutta. B: Fillhart. 1: Mack. B: O'Neal. M. (2013). Water quality
literature review and field monitoring of active shale gas wells: Phase II for assessing
environmental impacts of horizontal gas well drilling operations. Charleston, WV: West
Virginia Department of Environmental Protection, http: //www.dep.wv.gov/oil-and-
gas/Documents/Phase%20II%20report%20FINAL%20Qct%2 07.%2 02 013.pdf
Ziemkiewicz. P: Ouaranta. ID: Mccawlev. M. (2014). Practical measures for reducing the risk of
environmental contamination in shale energy production. Environ Sci Process Impacts 16:
1692-1699. http://dx.doi.org/10.1039/c3em00510k
Zirogannis. N: Alcorn. I: Rupp. 1: Carlev. S: Graham. ID. (2016). State regulation of unconventional
gas development in the U.S.: An empirical evaluation. ERSS 11: 142-154.
http://dx.doi.Org/10.1016/i.erss.2015.09.009
Zoback. MP. (2010). Reservoir geomechanics. Cambridge, UK: Cambridge University Press.
11-72
-------
Chapter 11 - References
Zolfaghari. A: Dehghanpour. H: Noel. M: Bearinger. D. (2016). Laboratory and field analysis of
flowback water from gas shales. J Unconven Oil Gas Resour 14: 113-127.
http://dx.doi.Org/10.1016/i.iuogr.2016.03.004
11-73
-------
Front cover (top): Illustrations of activities in the hydraulic fracturing water cycle.
From left to right: Water Acquisition, Chemical Mixing, Well Injection, Produced Water Handling,
and Wastewater Disposal and Reuse.
Front cover (bottom): Aerial photographs of hydraulic fracturing activities.
Left: Near Williston, North Dakota. Image ©J Henry Fair / Flights provided by LightHawk.
Right: Springville Township, Pennsylvania. Image ©J Henry Fair / Flights provided by LightHawk.
Back cover: Top left: DOE/NETL. All other images courtesy of the U.S. EPA.
-------
-------
vvEPA
United States
Environmental Protection
Agency
Office of Research and Development (8101R)
U.S. Environmental Protection Agency
Washington, DC 20460
Official Business
Penalty for Private Use
$300
W
Recycled/Recyclable
Printed with vegetable-based ink on paper that
contains a minimum of 50% post-consumer
fiber and is processed chlorine free.
------- |