Investigation of Ground Contamination near Pavillion, Wyoming

1. DRAFT EPA 600/R-00/000 | December 2011 | www.epa.gov/ord DRAFT Investigation of Ground Water Contamination near Pavillion, WyomingOffice of Research and DevelopmentNational Risk Management Research Laboratory, Ada, Oklahoma 74820

2. DRAFT

3. DRAFT Investigation of Ground Water Contamination near Pavillion, Wyoming Dominic C. DiGiulio Richard T. Wilkin Carlyle Miller U.S. Environmental Protection Agency Office of Research and Development National Risk Management Research Laboratory 919 Kerr Research Drive Ada, OK 74820 Gregory Oberley U.S. Environmental Protection Agency Region 8 1595 Wynkoop Street Denver, CO 80202Office of Research and DevelopmentNational Risk Management Research Laboratory, Ada, Oklahoma 74820

4. DRAFTNoticeThis report has been reviewed and approved by the U.S. Environmental Protection Agency’s Office of Researchand Development. Approval does not signify that the contents necessarily reflect the views and policies of theAgency, nor does mention of trade names or commercial products constitute endorsement or recommendationfor use.ii

5. DRAFT ForewordThe U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the Nation’s land, air,and water resources. Under a mandate of national environmental laws, the Agency strives to formulate andimplement actions leading to a compatible balance between human activities and the ability of natural systemsto support and nurture life. The scientific arm of EPA, the Office of Research and Development (ORD), conductsleading-edge research that helps provide the solid underpinning of science and technology for the Agency. Thework at ORD laboratories, research centers, and offices across the country helps improve the quality of air,water, soil, and the way we use resources. The research described in this report was designed and conducted byORDs National Risk Management Research Laboratory in Ada, Oklahoma, working in close collaboration withscientists from EPA Region 8 in Denver, Colorado. iii

6. DRAFTAcknowledgementsThe authors would to like to acknowledge valuable comments from 1 internal and 3 external reviewers used toimprove this manuscript. We would also like to acknowledge Dr. Randall Ross, Dr. Junqi Huang, Dr. Doug Beak,Mr. Steve Acree, Mr. Tony Lee, Mr. Ken Jewell, Mr. Tim Lankford, Mr. Russell Neil, and Ms. Kristie Hargrovefrom ORD/NRMRL/Ada and Mr. Christopher Ruybal and Ms. Alexandra Kirkpatrick (student contractors) for theirassistance in collecting ground water and gas samples. We would like to acknowledge Dr. Jennifer Gundersen ofEPAs Region 3 Laboratory for conducting analysis water samples for glycols and 2-butoxyethanol and Dr.William Batschelet of EPAs Region 8 Laboratory for conducting and arranging the analysis of water samples for anumber of classes of compounds. We also thank Mr. John Cox, Mr. Steve Markham, Ms. Tracy Pardue, Dr. FengLu, Mr. Joseph Karikari, Ms. Lisa Hudson, Dr. Sujith Kumar, Mr. Joe Tyburski, Mr. David Kessler, Mr. Jim Wilson(Shaw Environmental and Infrastructure), Mr. Mark White, Ms. Lynda Callaway, and Mr. Dennis Miller(ORD/NRMRL/Ada) for analytical support. We would like to thank Mr. Nathan Wiser, Mr. Robert Parker, and Ms.Johanna Miller of EPA Region 8 and Mr. Ron Mellor (SEE employee) for assistance in interpreting data andnumerous helpful comments. We would like to acknowledge Mr. Steve Vandegrift of ORD/NRMRL/Ada forproviding helpful comments in improving QA/QC aspects of this investigation and overseeing development ofthe QAPP and ADQs. We would like to acknowledge Dr. John Wilson for assistance in interpretation of data. Weare grateful to Ms. Ayn Schmit of EPA Region 8 and Dr. David Jewett of ORD/NRMRL/Ada for ongoing support intheir respective management roles and ability to effectively communicate technical details in this manuscript toothers. We would like to express our appreciation to Mr. Jeff Locker and Ms. ZoeAnn Randall for access to theirproperty for monitoring well installation and to Mr. Louis Meeks for access to his property for domestic wellsampling. We are grateful to Mr. John Fenton for access to his property for domestic well sampling andfacilitating contact with domestic well owners in the area. We are grateful to Ms. Kathy Tynsky of SRA forassistance in developing graphics in this document.iv

7. DRAFT ContentsNotice ......................................................................................................................................................................... iiForeword ................................................................................................................................................................... iiiAcknowledgements .................................................................................................................................................. ivFigures ....................................................................................................................................................................... viTables ........................................................................................................................................................................ ixExtended Abstract..................................................................................................................................................... xi1.0 Site Background ................................................................................................................................................... 12.0 Methods ............................................................................................................................................................... 5 Sampling Chronology..................................................................................................................................... 5 Deep Monitoring Well Installation ................................................................................................................ 5 Ground Water Sampling of Deep Monitoring Wells in Phase III and IV ...................................................... 11 Gas Sampling from Casing of Deep Monitoring Wells in Phase III and IV ................................................... 15 Domestic Well Sampling for Methane Using a Closed System in Phase IV ................................................. 15 Review of Borehole Geophysical Logs ........................................................................................................ 15 Review of Cement Bond/Variable Density Logs .......................................................................................... 163.0 Results and Discussion....................................................................................................................................... 17 Ground Water and Soil Sample Results near Three Pits ............................................................................ 17 Inorganic Geochemistry .............................................................................................................................. 17 Organic Geochemistry ................................................................................................................................. 23 Natural Gas Migration ................................................................................................................................. 27 Evaluation of Cement Bond/Variable Density Logs Along Transect............................................................ 29 Potential Migration Pathways ..................................................................................................................... 324.0 Conclusions ........................................................................................................................................................ 335.0 References.......................................................................................................................................................... 40Appendices .............................................................................................................................................................. A1 Appendix A – Summary of Analytical Results ............................................................................................ A1 Appendix B – Quality Assurance and Quality Control (QA/QC) for Analysis ..............................................B1 Appendix C – Photographic Log of Deep Monitoring Well Construction ...................................................C1 Appendix D – Photographic Log of Ground Water Sampling .................................................................... D1 Appendix E – Examples of Cement Bond/Variable Density Log Interpretation ......................................... E1 v

8. DRAFTFiguresFigure 1. (a) Location of Wind River Basin in Wyoming. (b) Location of Pavillion Gas Field in the Wind River Basin.Figure from Johnson et al. 2007 ................................................................................................................................. 1Figure 2. Chronology of production well completion at the Pavillion Gas Field ........................................................ 2Figure 3. Histograms summarizing depths of top of perforation interval of production wells, base of surfacecasing of production wells, and base of screened interval of domestic wells ........................................................... 2Figure 4. Generalized stratigraphic columns and correlations of Mississippian through Eocene strata in the WindRiver Basin, Wyoming. The Pavillion Gas Field is located in the Western Wind River Basin. Figure from Johnsonet al. 2007 .................................................................................................................................................................. 3Figure 5. Map illustrating location of oil and gas production wells, sampled PGDWxx series domestic wells (onlynumbers shown to conserve space), two deep monitoring wells, and three shallow monitoring wells near pits.PGDW07 and PGDW08 are municipal wells in the town of Pavillion......................................................................... 6Figure 6a. Schematic illustrating construction of MW01 .......................................................................................... 9Figure 6b. Schematic illustrating construction of MW02........................................................................................ 10Figure 7. Resistivity as a function of depth in MW01 and MW02. MW01 and MW02 were screened at 233 - 239m and 293 - 299 m bgs respectively corresponding to elevated resistivity and presence of coarse-grainedsandstone. FID readings in MW01 denote detections of methane during open air logging of mud. FIDmonitoring at MW02 was sporadic and is not illustrated here................................................................................ 11Figure 8. Variation of water level as a function of time in MW01 during Phase 4 well purging. The initialpumping rate was 24.2 L/min. After approximately 30 minutes of purging, the flow rate was decreased to 7.6L/min. This reduced flow rate caused partial recovery of the water level and confirmation that formation waterwas being accessed .................................................................................................................................................. 12Figure 9. Flow-cell readings as a function of time for specific conductance, dissolved oxygen, pH, and oxidation-reduction potential (well MW02, Phase IV sampling) ............................................................................................. 13Figure 10. Schematic of closed (no contact to atmosphere) sampling train for domestic wells. Water flow fromdomestic well and into sparge cell was approximately 5 and 1 L/min respectively. Excess water bled throughvalve used for sampling prior to sample collection. Gas flow into sparge cell and portable FID/PID sparge cellwas approximately 20 and 1 L/min. Excess air was bled through splitter above sparge cell ................................. 16Figure 11. Durov diagram showing ground water chemistry trends obtained in Phase I - IV sampling events andthe composition of irrigation water ......................................................................................................................... 17Figure 12. Depth trends of chloride, pH, sulfate, and potassium (filled black squares = domestic wells, filled redcircles = monitoring wells) ....................................................................................................................................... 18vi

9. DRAFTFigure 13. Saturation indices for (a) gypsum versus sulfate concentration and (b) calcite versus calciumconcentration. Saturation Index is equal to the logarithm of the ratio of the ion activity product to the mineralsolubility product. A Saturation Index of 0 corresponds to chemical equilibrium; values less than 0 and greaterthan 0 correspond to undersaturated and oversaturated conditions, respectively ................................................ 19Figure 14. Concentration trends versus specific conductivity. Note the monitoring wells show high pH and lowsulfate, calcium, and sodium relative to the general trend observed in the domestic wells (filled black squares =domestic wells, filled red circles = monitoring wells) .............................................................................................. 21Figure 15. (a) Results of KOH titration models plotted as pH versus grams of KOH added per kilogram ofsolution. Initial water compositions are from PGDW49, PGDW20, and PGDW32. Model accounts for reactionstaking place in solution as KOH is added and equilibrated. pH range in deep monitoring wells shown forreference; (b) Buffer Intensity plot or first derivative of titration plot, pH versus change in concentration of base(CB) per change in pH ................................................................................................................................................ 22Figure 16. Hydrogen and oxygen isotope values (permil, Vienna Standard Mean Ocean Water, VSMOW) forground water samples (black squares=domestic wells; red circles=deep monitoring wells) relative to the GlobalMeteoric Water Line from Craig (1961) .................................................................................................................. 22Figure 17. Organic compounds detected in deep monitoring wells MW01 and MW02 during Phase III and IVsampling events ........................................................................................................................................................ 25Figure 18. (a) Stable isotope ratios of carbon of methane versus ratio of methane (C1) to ethane (C2) andpropane (C3) in gas from production wells, monitoring wells, and domestic wells. Values of 100,000 are used todenote non detection of ethane and propane in samples. (b) Stable isotope ratios of carbon versus hydrogen ofmethane in gas from production wells (both literature and measured values), monitoring wells, and domesticwells. δD was not determined for PGDW32. Oxidation pathway (enrichment of 13C of remaining CH4 withbiodegradation) is illustrated. (c) Methane concentration in domestic and monitoring wells as a function ofproximity to production wells and depth. Values of 1.0 were used for non-detection (detection limit 5 µg/L) .... 28Figure 19. Map illustrating transect used to develop lithologic cross section and evaluation of CBL/VDLs .......... 30Figure 20. Lithologic cross-section along transect illustrating production wells (with evaluation of CBL/VDLs),domestic wells, and blowout location. Red arrows denote depths of hydraulic fracturing of unknown arealextent ....................................................................................................................................................................... 31Figure C1. Photograph of drilling rig on platform with shakers for mud recirculation at MW02 ...........................C2Figure C2. Photograph of blowout prevention (BOP) for annular space at base of drilling rig platform at MW02..................................................................................................................................................................................C2Figure C3. Photograph of blowout preventer for drillstem ....................................................................................C2Figure C4. Photograph of bit and drillstem with bit for mud rotary drilling at MW02 ...........................................C3Figure C5. Photograph of water truck used to transport water to mix mud ..........................................................C4Figure C6. Photograph of Quik-Gel bentonite (Halliburton) used to create mud for drilling .................................C4Figure C7. Photograph of mud additives EZ Mud Gold (Halliburton) and Dense Soda Ash ....................................C4 vii

10. DRAFTFigure C8. Photograph of mud additive Penetrol (Halliburton) .............................................................................C4Figure C9. Photograph of flow of mud and cuttings from borehole at MW02 .......................................................C5Figure C10. Photograph of monitoring of mud and cuttings using a Thermo Scientific TVA-1000B FID/PID atMW02 .......................................................................................................................................................................C5Figure C11. Photograph of pump used to transport mud and cuttings to shakers at MW02 ................................C6Figure C12. Photograph of flow of mud and cuttings to shakers at MW02............................................................C6Figure C13. Photograph of shakers separating mud from cuttings at MW02 ........................................................C7Figure C14. Photograph of cuttings transported to disposal bins at MW02...........................................................C8Figure C15. Photograph of pumping of mud back to borehole at MW02 ..............................................................C9Figure C16. Photograph of injection of mud to borehole at MW02 .....................................................................C10Figure C17. Photograph of collection of cuttings for lithologic characterization at MW02 .................................C11Figure C18. Photograph of removal of mud from cuttings at MW02 ...................................................................C11Figure C19. Photograph of white coarse-grained sand targeted by local well drillers and media in which screensare set in for both deep monitoring wells ..............................................................................................................C11Figure C20. Photograph of setting of stainless-steel pre-packed screen and sand basket into borehole at MW02................................................................................................................................................................................C12Figure C21. Photograph of securing sand basket and casing above screen..........................................................C13Figure C22. Photograph of placement of sand in sandbasket ..............................................................................C13Figure C23. Photograph of well development at MW02 ......................................................................................C14Figure D1. Photograph of flow from submersible pump through flowmeter at MW02........................................ D2Figure D2. Photograph of flow of water to purge water disposal tank at MW02 ................................................. D2Figure D3. Photograph (close-up) of flow of water into purge water disposal tank at MW02 ............................. D3Figure D4. Photograph of water (foaming) flowing into YSI flow cell at MW02 .................................................... D3Figure D5. Photograph of sampling at MW02. The sample train was split prior to entry into purge waterdisposal container ................................................................................................................................................... D4Figure D6. Photograph of field filtering samples for metals analysis at MW02 .................................................... D4Figure D7. Photograph of sample collection at PGDW14 ...................................................................................... D5Figure D8. Photograph of cooler packed with samples for shipment.................................................................... D5viii

11. DRAFTFigure E1. Example of CBL/VDL indicating "no cement" at Pavillion Fee 34-03B. The CBL/VDL indicates nocement 2750 feet below ground surface at the time of logging.............................................................................. E2Figure E2. Example of "sporadic bonding" at Pavillion Fee 41-10 from 1000 to 1640 ft bgs. Hydraulic fracturingoccurred at 1618 feet below ground surface. Arrow denotes interval of hydraulic fracturing .............................. E3Figure E3a. Example of "sporadic bonding" at Pavillion Fee 11-11B. Hydraulic fracturing occurred at 1516 feetbelow ground surface. Arrow denotes interval of hydraulic fracturing. Depths on CBL/VDL difficult to read andinserted on left margin ............................................................................................................................................. E4Figure E3b. Example of "sporadic bonding" Pavillion Fee 11-11B between 2350-3200 feet below ground suface.Hydraulic fracturing occurred at 3165 feet below ground surface. Arrow denotes interval of hydraulic fracturing.Depths on CBL/VDL difficult to read and inserted on left margin............................................................................ E5Figure E4. Example of "Sporadic Bonding" at Tribal Pavillion 24-02. Hydraulic fracturing occurred at 1538 feetbgs. Arrow denotes interval of hydraulic fracturing ............................................................................................... E6Figure E5. Example of "Good Bonding" (from surface casing at 645 ft bgs to 820 ft bgs) followed by "SporadicBonding" (from 820 ft bgs 1310 ft bgs) to "Good Bonding" at 1310 to target depth at Pavillion Fee 41-10B. ....... E7 TablesTable 1. Drilling additives, properties and product use recommendations.............................................................. 7Table 2. Analytical results of additives (compounds listed are those detected in ground water) ........................... 8Table 3. Geochemical impacts in deep ground water monitoring wells................................................................. 24Table 4. Association of inorganic and organic anomalies with compounds used for hydraulic fracturing ............ 26Table A1. Summary of subsurface sample locations, depth of sample collection, times (phases) of sampling,target analytes, laboratories utilized, and analytical methods ............................................................................... A2Table A2a. Geochemical results for Pavillion ground water .................................................................................. A8Table A2b. Charge balance calculations for deep monitoring wells ...................................................................... A9Table A3a. Summary of aqueous analysis of light hydrocarbons ........................................................................ A10Table A3b. Summary of gas and headspace analysis of light hydrocarbons ....................................................... A12Table A3c. Summary of isotopic data for dissolved, gas phase, and headspace analysis.................................... A14Table B1. Sample collection containers, preservation, and holding times for ground water samples ...................B2Table B2. Field QC samples for ground water analysis ...........................................................................................B3 ix

12. DRAFTTable B3. QA/QC requirements for analysis of metals and major ions ..................................................................B3Table B4. QA/QC requirements for analysis of dissolved gases, DIC/DOC, VOCs, low molecular weight acids andstable isotopes of water ...........................................................................................................................................B4Table B5. QA/QC requirements for analysis of semi-volatiles, GRO, and DRO.......................................................B5Table B6. QA/QC requirements for LC/MS/MS analysis of glycols ..........................................................................B6TableB7a. ICP-OES blank results for Phase III and Phase IV sampling ....................................................................B7Table B7b. ICP-OES blank results for Phase III and Phase IV sampling ...................................................................B7Table B7c. ICP-MS blank results for Phase III and Phase IV sampling .....................................................................B8Table B8. Blank results for Capillary Electrophoresis, Lachat Flow Injection Analysis, Dissolved Inorganic Carbon(DIC) and Dissolved Organic Carbon analyses for Phase III and Phase IV sampling .................................................B8Table B9. Blank results for Volatile Organic Compounds (μg/L) in Phase III and Phase IV sampling (Region 8laboratory, Golden, CO) ...........................................................................................................................................B9Table B10. Blank results for Volatile Organic Compounds (μg/L) in Phase IV sampling (ORD laboratory, Ada, OK)............................................................................................................................................................................... B11Table B11. Blank results for Semi-Volatile Organic Compounds (μg/L) in Phase III and Phase IV sampling (Region8 laboratory, Golden, CO) ..................................................................................................................................... B12Table B12. Blank results for GRO and DRO analyses for Phase III and Phase IV sampling (Region 8 laboratory,Golden, CO) and blank results for glycol ethers in Phase IV sampling (Region 3 laboratory, Fort Meade, MD) . B13Table B13. Duplicate data for selected major ions, DOC, and DIC in ground water samples collected duringPhase III and Phase IV sampling activities ............................................................................................................. B14Table B14. Duplicate data for methane and selected dissolved organic compounds in ground water samplescollected during Phase III and Phase IV sampling activities .................................................................................. B15Table B15. QA/QC requirements for analysis of δ13C of DIC ................................................................................ B16Table B16. QA/QC requirements for analysis for δ13C and δD of light hydrocarbons for aqueous and gas samples............................................................................................................................................................................... B16Table B17. QA/QC requirements for analysis of fixed gases and light hydrocarbons for aqueous and gas samples............................................................................................................................................................................... B16Table B18. Summary of quality control samples, purpose, method, and frequency to support gas analysis...... B17Table B19. Summary of analytes, instruments, calibration, and check standards for portable gas analyzers.... B17Table B20. QA/QC Requirements for portable gas analyzers .............................................................................. B17x

13. DRAFT Extended AbstractIn response to complaints by domestic well owners regarding objectionable taste and odor problems in wellwater, the U.S. Environmental Protection Agency initiated a ground water investigation near the town ofPavillion, Wyoming under authority of the Comprehensive Environmental Response, Compensation, and LiabilityAct. The Wind River Formation is the principal source of domestic, municipal, and stock (ranch, agricultural)water in the area of Pavillion and meets the Agencys definition of an Underground Source of Drinking Water.Domestic wells in the area of investigation overlie the Pavillion gas field which consists of 169 production wellswhich extract gas from the lower Wind River Formation and underlying Fort Union Formation. Hydraulicfracturing in gas production wells occurred as shallow as 372 meters below ground surface with associatedsurface casing as shallow as 110 meters below ground surface. Domestic and stock wells in the area arescreened as deep as 244 meters below ground surface. With the exception of two production wells, surfacecasing of gas production wells do not extend below the maximum depth of domestic wells in the area ofinvestigation. At least 33 surface pits previously used for the storage/disposal of drilling wastes and producedand flowback waters are present in the area. The objective of the Agencys investigation was to determine thepresence, not extent, of ground water contamination in the formation and if possible to differentiate shallowsource terms (pits, septic systems, agricultural and domestic practices) from deeper source terms (gasproduction wells). The Agency conducted four sampling events (Phase I - IV) beginning in March 2009 and ending in April, 2011.Ground water samples were collected from domestic wells and two municipal wells in the town of Pavillion inPhase I. Detection of methane and dissolved hydrocarbons in several domestic wells prompted collection of asecond round of samples in January, 2010 (Phase II). During this phase, EPA collected additional ground watersamples from domestic and stock wells and ground water samples from 3 shallow monitoring wells and soilsamples near the perimeter of three known pit locations. Detection of elevated levels of methane and dieselrange organics (DRO) in deep domestic wells prompted the Agency to install 2 deep monitoring wells screenedat 233 - 239 meters (MW01) and 293 - 299 meters (MW02) below ground surface, respectively, in June 2010 tobetter evaluate to deeper sources of contamination. The expense of drilling deep wells while utilizing blowoutprevention was the primary limiting factor in the number of monitoring wells installed. In September 2010(Phase III), EPA collected gas samples from well casing from MW01 and MW02. In October 2010, EPA collectedground water samples from MW01 and MW02 in addition to a number of domestic wells. In April 2011 (PhaseIV), EPA resampled the 2 deep monitoring wells to compare previous findings and to expand the analyte list toinclude glycols, alcohols, and low molecular weight acids.Detection of high concentrations of benzene, xylenes, gasoline range organics, diesel range organics, and totalpurgeable hydrocarbons in ground water samples from shallow monitoring wells near pits indicates that pits area source of shallow ground water contamination in the area of investigation. When considered separately, pitsrepresent potential source terms for localized ground water plumes of unknown extent. When considered aswhole they represent potential broader contamination of shallow ground water. A number of stock anddomestic wells in the area of investigation are fairly shallow (e.g., < 30 meters below ground surface)representing potential receptor pathways.Determination of the sources of inorganic and organic geochemical anomalies in deeper ground water wasconsiderably more complex than determination of sources in shallow media necessitating the use of mulitiple xi

14. DRAFTlines of reasoning approach common to complex scientific investigations. pH values in MW01 and MW01 arehighly alkaline (11.2-12.0) with up to 94% of the total alkalinity contributed by hydroxide suggesting addition ofa strong base as the causative factor. Reaction path modeling indicates that sodium-sulfate composition ofground water typical of deeper portions of the Wind River Formation provides little resistance to elevation of pHwith small addition of potassium hydroxide. Potassium hydroxide was used in a crosslinker and in a solvent atthis site.The inorganic geochemistry of ground water from the deep monitoring wells is distinctive from that in thedomestic wells and expected composition in the Wind River formation. Potassium concentration in MW02 (43.6milligrams per liter) and MW01 (54.9 milligrams per liter) is between 14.5 and 18.3 times values in domesticwells and expected values in the formation. Chloride concentration in monitoring well MW02 (466 milligramsper liter) is 18 times the mean chloride concentration (25.6 milligrams per liter) observed in ground water fromdomestic wells and expected in the formation. Chloride enrichment in this well is significant because regionalanion trends show decreasing chloride concentration with depth. In addition, the monitoring wells show lowcalcium, sodium, and sulfate concentrations compared to the general trend observed in domestic well waters.The formulation of fracture fluid provided for carbon dioxide foam hydraulic fracturing jobs typically consisted of6% potassium chloride. Potassium metaborate was used in crosslinkers. Potassium hydroxide was used in acrosslinker and in a solvent. Ammonium chloride was used in crosslinker.A number of synthetic organic compounds were detected in MW01 and MW02. Isopropanol was detected inMW01 and MW02 at 212 and 581 micrograms per liter, respectively. Diethylene glycol was detected in MW01and MW02 at 226 and 1570 micrograms per liter, respectively. Triethylene glycol was detected in MW01 andMW02 at 46 and 310 micrograms per liter, respectively. Another synthetic compound, tert-butyl alcohol, wasdetected in MW02 at a concentration of 4470 micrograms per liter. Isopropanol was used in a biocide, in asurfactant, in breakers, and in foaming agents. Diethylene glycol was used in a foaming agent and in a solvent.Triethylene glycol was used in a solvent. Tert-butyl alcohol is a known breakdown product of methyl tert-butylether (a fuel additive) and tert-butyl hydroperoxide (a gel breaker used in hydraulic fracturing). Material SafetyData Sheets do not indicate that fuel or tert-butyl hydroperoxide were used in the Pavillion gas field. However,Material Safety Data Sheets do not contain proprietary information and the chemical ingredients of manyadditives. The source of tert-butyl alcohol remains unresolved. However, tert-butyl alcohol is not expected tooccur naturally in ground water.Benzene, toluene, ethylbenzene, and xylenes (BTEX) were detected in MW02 at concentrations of 246, 617, 67,and 750 micrograms per liter, respectively. Trimethylbenzenes were detected in MW02 at 105 micrograms perliter. Gasoline range organics were detected in MW01 and MW02 at 592 and 3710 micrograms per liter. Dieselrange organics were detected in MW01 and MW02 at 924 and 4050 micrograms per liter, respectively.Aromatic solvent (typically BTEX mixture) was used in a breaker. Diesel oil (mixture of saturated and aromatichydrocarbons including naphthalenes and alkylbenzenes) was used in a guar polymer slurry/liquid gelconcentrate and in a solvent. Petroleum raffinates (mixture of paraffinic, cycloparaffinic, olefinic, and aromatichydrocarbons) were used in a breaker. Heavy aromatic petroleum naphtha (mixture of paraffinic, cycloparaffinicand aromatic hydrocarbons) was used in surfactants and in a solvent. Toluene and xylene were used in flowenhancers and a breaker.Detections of organic chemicals were more numerous and exhibited higher concentrations in the deeper of thetwo monitoring wells. Natural breakdown products of organic contaminants like BTEX and glycols includexii

15. DRAFTacetate and benzoic acid. These breakdown products are more enriched in the shallower of the two monitoringwells, suggesting upward/lateral migration with natural degradation and accumulation of daughter products.Hydraulic gradients are currently undefined in the area of investigation. However, there are flowing conditionsin a number of deep stock wells suggesting that upward gradients exist in the area of investigation.Alternative explanations were carefully considered to explain individual sets of data. However, when consideredtogether with other lines of evidence, the data indicates likely impact to ground water that can be explained byhydraulic fracturing. A review of well completion reports and cement bond/variable density logs in the areaaround MW01 and MW02 indicates instances of sporadic bonding outside production casing directly aboveintervals of hydraulic fracturing. Also, there is little lateral and vertical continuity of hydraulically fractured tightsandstones and no lithologic barrier (laterally continuous shale units) to stop upward vertical migration ofaqueous constituents of hydraulic fracturing in the event of excursion from fractures. In the event of excursionfrom sandstone units, vertical migration of fluids could also occur via nearby wellbores. For instance, at oneproduction well, the cement bond/variable density log indicates no cement until 671 m below ground surface.Hydraulic fracturing occurred above this depth at nearby production wells.A similar lines of reasoning approach was utilized to evaluate the presence of gas in monitoring and domesticwells. A comparison of gas composition and stable carbon isotope values indicate that gas in production andmonitoring wells is of similar thermogenic origin and has undergone little or no degradation. A similarevaluation in domestic wells suggests the presence of gas of thermogenic origin undergoing biodegradation.This observation is consistent with a pattern of dispersion and degradation with upward migration observed fororganic compounds.Elevated levels of dissolved methane in domestic wells generally increase in those wells in proximity to gasproduction wells. Near surface concentrations of methane appear highest in the area encompassing MW01.Ground water is saturated with methane at MW01 which is screened at a depth (239 meters below groundsurface) typical of deeper domestic wells in the area. A blowout occurred during drilling of a domestic well at adepth of only 159 meters below ground surface close to MW01. A mud-gas log conducted in 1980 (prior tointensive gas production well installation) located only 300 m from the location of the blowout does not indicatea gas show (distinctive peaks on a gas chromatograph) within 300 meters of the surface. Again, with theexception of two production wells, surface casing of gas production wells do not extend below the maximumdepth of domestic wells in the area of investigation. A number of production wells in the vicinity of MW01 havesporadic bonding or no cement over large vertical instances. Again, alternate explanations of data have beenconsidered. Although some natural migration of gas would be expected above a gas field such as Pavillion, datasuggest that enhanced migration of gas has occurred within ground water at depths used for domestic watersupply and to domestic wells. Further investigation would be needed to determine the extent of gas migrationand the fate and transport processes influencing migration to domestic wells. xiii

16. DRAFT Domestic wells in the area of investigation overlie the1.0 Pavillion gas field which is one of several gas fieldsSite Background within the Wind River Basin - a large, complex, structural, asymmetric, deep sedimentary basinIn early 2008, the U.S. Environmental Protection covering much of central Wyoming (Figure 1). Oil and gas exploration wells were drilled in the 1950s.Agency (EPA) received complaints from several Commercial natural gas extraction in the fielddomestic well owners near the town of Pavillion, commenced in 1960 (Single 1969) with gas productionWyoming regarding sustained objectionable taste and well installation activity intensifying in the late 1990sodor problems in well water following hydraulic through 2006 (Figure 2). The field currently consistsfracturing at nearby gas production wells. In responseto these complaints, EPA initiated a comprehensive of approximately 169 vertical production wells.ground water investigation in September 2008 under Ninety-seven production wells are designated as "Tribal Pavillion" and are regulated by the U.S. Bureauauthority of the Comprehensive Environmental of Land Management (BLM). The remaining wells areResponse, Compensation, and Liability Act. The areaof investigation is a sparsely populated rural area in designated as "Pavillion Fee" and are regulated by Wyoming Oil and Gas Conservation Commissionwest-central Wyoming directly east of the town ofPavillion. Land use by residents consists primarily of (WOGCC).ranching (horse and cattle) and alfalfa hay productionfor use by ranchers and commercial sale. Fields areperiodically flooded using water obtained from canalsand laterals. Figure 1. (a) Location of Wind River Basin in Wyoming. (b) Location of Pavillion gas field in the Wind River Basin. Figure from Johnson et al. 2007.1

17. DRAFT Gas extraction occurs from both the lower Eocene Wind River Formation and underlying Paleocene Fort Union Formation (Figure 4). The Wind River Formation consists of interbedded layers of sandstones and shale with coarse-grained meandering stream channel deposits (Osiensky et al. 1984) and extends from the surface to approximately 1036 m (3400 ft) bgs. The Fort Union Formation ranges in thickness from 762 to 914 m (2500 to 3000 ft) in the area (Flores and Keighin 1993). The Waltman Shale Member in the Fort Union Formation is absent below Figure 2. Chronology of production well completion the Pavillion Gas Field. The most productive zone of at the Pavillion gas field. gas extraction in the Wind River Formation occurs at its base and is often targeted for gas extraction (SingleA review of production well records obtained on line 1969). Gas trapping in the lower Wind River and Fortfrom WOGCC indicates that hydraulic fracturing in gas Union Formations occurs in localized stratigraphicproduction wells occurred as shallow as 372 m (1220 sandstone pinchouts on the crest and along flanks of aft) below ground surface (bgs) with associated surface broad dome (Mueller 1989, Keefer and Johnson 1993).casing in production wells as shallow as 110 m (361 ft)bgs. Information obtained from the Wyoming State There is substantial vertical and lateral stratigraphicEngineers Office and homeowners indicates that variation over short distances in both formationsdomestic wells (including stock wells) in the area of (Single 1969, Flores and Keighin 1993). Individualinvestigation are screened as deep as 244 m (800 ft) productive sandstones in the two formations generallybgs. With the exception of two production wells, vary in thickness from 1 to 21 m with permeabilitysurface casings of gas production wells do not extend varying from 0.1 to 300 millidarcies and porositybelow the maximum depth of domestic wells in the ranging from 4 to 28 percent (Single 1969). Gas fromarea of investigation (Figure 3). the Fort Union and lower Wind River Formations varies little in δ13C for methane, ethane, and propane with depth from the lower Eocene Wind River Formation to deeper mature and post-mature Upper Cretaceous source rocks (Figure 4) suggesting upward gas migration (Johnson and Rice 1993, Johnson and Keighin 1998) from deep source rocks. δ13C is defined as  δ13C ( ‰) =  ( 13 C 12 ) C sample  − 1 x1000   (13 C 12 C ) standard   where the standard is the Pee Dee Belemnite (PDB) Figure 3. Histograms summarizing depths of top of reference standard. Stable isotope ratios are reported perforation interval of production wells, base of as the relative difference in the ratio of the less surface casing of production wells, and base of screened interval of domestic wells. abundant heavier isotope to the more abundant lighter isotope of the sample with respect to a 2

18. DRAFT Figure 4. Generalized stratigraphic columns and correlations of Mississippian through Eocene strata in the Wind River Basin, Wyoming. The Pavillion Gas Field is located in the Western Wind River Basin. Figure from Johnson et al. 2007.3

19. DRAFTreference standard. Ratios are expressed in parts perthousand or permil (‰). A substantial amount ofadditional compositional and isotopic data is availableon the Wind River and Fort Union Formations but isclassified as Confidential Business Information by thegas field operator.Ground water from the upper Wind River Formation isthe principal source of domestic, municipal, and stock(ranching, agriculture) water in the Pavillion area (WYState Water Plan 2003). The Wind River Formationmeets the definition of an Underground Source ofDrinking Water (USDW) under the United States Codeof Federal Regulations, Title 40, Section 144.3. Wateryields from wells in the upper Wind River Formationrange up to 11,300 L/min with total dissolved-solids(TDS) concentrations ranging from 100 to 5,110 mg/L(WY State Water Plan 2003, Daddow 1996). The townof Pavillion has five municipal wells screened atdepths ranging from 122 to 158 m bgs with averagedaily use estimated at 60,000 L/day (WY State WaterPlan 2003). Fluids used for hydraulic fracturing wereinjected directly into the Wind River Formation. 4

20. DRAFT domestic wells. In April 2011 (Phase IV), EPA2.0 resampled the 2 deep monitoring wells to compareMethods previous findings and expand the analyte list to include glycols, alcohols, and low molecular weight acids. Eight previously sampled domestic wells andSampling Chronology three previously sampled stock/irrigation wells wereFour sampling events (Phase I - IV) were conducted also sampled at this time. Sampling chronology andcommencing in March 2009 and ending in April 2011. analytical methods for all sampling events areIn March 2009 (Phase I), EPA collected aqueous summarized in Table A1. The location of productionsamples from 35 domestic wells (including two wells, monitoring wells, and sampled domestic wells issamples from post reverse osmosis systems) in the illustrated in Figure 5.area of investigation and 2 municipal wells in the townof Pavillion. Detection of methane and dissolved Deep Monitoring Well Installationhydrocarbons in several domestic wells prompted EPA installed two deep monitoring wells (designatedcollection of a second round of samples in January as MW01 and MW02) using air (0 - 6 m bgs) and mud2010 (Phase II). During this phase, EPA collected: (1) rotary (6 m bgs to target depth). Mud rotary wasground water samples from 17 domestic wells (10 selected for installation of deep monitoring wellspreviously sampled), 4 stock wells, and 2 municipal because it allowed the use of blowout preventionwells; (2) a filter sample from a reverse osmosis (BOP). Use of mud rotary with BOP was necessarysystem; (3) surface-water and sediment samples from given that a blowout occurred during installation of a5 locations along Five-Mile Creek (a creek traversing domestic well at only 159 m (522 ft) bgs in Decemberthe area of investigation); (4) gas and produced 2005 in the vicinity of MW01. Both deep monitoringwater/condensate samples (organic compounds only) wells were located away from gas production wells,from 5 production wells; and (5) ground water known locations of pits, and areas of domestic wastesamples from 3 shallow monitoring wells and soil disposal (abandoned machinery). There were nosamples near the perimeter of three known pit incidents of fuel spillage used to power pumps andlocations. generators.Detection of elevated levels of methane and diesel Mud rotary required the use of drilling mud to removerange organics (DRO) in deep domestic wells cuttings and additives to avoid heaving of shale duringprompted EPA to install 2 deep monitoring wells in drilling and well placement. Jet Lube Well GuardJune 2010 to differentiate potential deep (e.g., gas hydrocarbon free lubricant was used for outsideproduction related) versus shallow (e.g., pits) sources threads for drillstem and submersible pipe connections. Mud composition consisted of formationof ground water contamination. Monitoring wells water, municipal drinking water from Riverton, WYMW01 and MW02 were screened at 233 - 239 m (765 (transported on site by water truck), Quik-Gel high– 785 ft) and 293 - 299 m (960 – 980 ft) bgs, yield bentonite and additives listed on Table 1.respectively. The expense of drilling deep wells while Municipal water was mixed with bentonite to createutilizing blowout prevention was the primary limiting drilling mud. The pH of mud during drilling variedfactor in the number of monitoring wells installed. In between pH 8 - 9. Aqua-Clear (Halliburton) was usedSeptember 2010 (Phase III), EPA collected gas samples during well development to facilitate removal of mud.from well casing from MW01 and MW02. In October Drilling additives were extracted in water (1:20 to2010, EPA collected ground water samples from 1:100 dilution) and analyzed for pH, inorganics,MW01 and MW02 in addition to a previously organics, glycols, and alcohols. Despite the highlyunsampled domestic well and two previously sampled concentrated nature of these solutions (not5

21. DRAFT Figure 5. Map illustrating location of oil and gas production wells, sampled PGDWxx series domestic wells (only numbers shown to conserve space), two deep monitoring wells, and three shallow monitoring wells near pits. PGDW07 and PGDW08 are municipal wells in the town of Pavillion.6

22. DRAFTrepresentative of significantly lower levels in drilling Composite samples of cuttings were collected andmud, see recommended product use mixture listed in sent to TestAmerica Laboratories in Denver, ColoradoTable 1), the pH of samples varied between 6.6 to for Toxic Characteristic Leaching Procedure (TCLP).11.2, potassium varied between 0.1 to 1.2 mg/L, Samples were analyzed for TCLP volatile organicchloride varied between not detected to 214 mg/L, compounds using gas chromatography-massethanol and isopropanol detections were less than 90 spectrometry (GC-MS) in accordance with EPA SW-846μg/L, and acetone, tert-butyl alcohol (TBA), benzene, Methods 1311/8260B, and for TCLP semivolatiletoluene, ethylbenzene, xylenes (BTEX), organic compounds (GC-MS) in accordance with EPAtrimethylbenzenes, and glycols were not detected SW-846 Methods 1311/8270C, for TCLP metals in(Table 2). Organics were not analyzed in the dense accordance with EPA SW-846 Methods 1311/ 6010B,soda ash and Quik-Gel because dissolved organic for TCLP mercury in accordance with EPA SW-846carbon concentrations were low and because of Methods 1311/7470A. Acetone, toluene, and m & p-difficulties in analyzing the viscous gel (Quik-Gel). xylene were detected in one sample at 6.9, 0.63, andSince inorganic and organic concentration patterns 1.0 μg/L, respectively. Cuttings were disposed offsitemeasured in the drilling additives do not match in a landfill.patterns observed in the deep monitoring wells andbecause large volumes of ground water were A photographic log of drilling, mud circulation, examination of cuttings, screen placement, and wellextracted from the wells during development and development is provided in Appendix C. Wellprior to sampling, it is unlikely that ground water construction schematics are provided for MW01 andchemistry was impacted by drilling additives. MW02 in Figures 6a and 6b, respectively. During7

23. DRAFTMW02 in Figures 6a and 6b, respectively. During and power washed (no detergents used) prior toinstallation of MW02, cuttings were allowed to settle (deployment. Locations of both MW01 and MW02at the cessation of drilling and form a 5 m (17 ft) base were in fields used for alfalfa hay production awayfor placement of the screen. Cuttings were never from production wells, pads, and pits.added to the borehole. Since a significant vertical Cuttings were continuously examined during drillingdistance existed between the depth of drilling and by manually washing drilling mud from rock fragmentsscreen placement at MW01, cement grout was utilized with observations recorded as a function of depth into form the base for screen placement. No lubricants borehole logs. At the cessation of drilling, open-holewere used to attach sections of casing or casing to geophysical logging (caliper, density, resistivity,screens. Well screens, sections of casing and tremie spontaneous potential, natural gamma) waspipe were mounted above ground (never touched soil) conducted by Colog Inc., prior to placement of welland power washed (no detergents used) prior to 8

24. DRAFT Figure 6a. Schematic illustrating construction of MW01.9

25. DRAFTFigure 6b. Schematic illustrating construction of MW02. 10

26. DRAFTconstruction materials. Examination of resistivity and Ground Water Sampling of Deep Monitoringcuttings indicated elevated resistivity at depths where Wells in Phase III and IVwhite coarse-grained sandstone was observed. Thisrelationship was utilized to place screens at both deep Ground water in deep monitoring wells was sampledmonitoring wells at the deepest observed interval of using dedicated explosion proof submersible pumpswhite coarse-grained sand (Figure 7). White coarse- (10-cm Franklin Electric 3HP). Wells were purged at agrained sandstones in the area of investigation flow rate of approximately 5 to 30 L/min. The rate ofcontain little or no shale and are targeted by local well pumping was measured using a Model TM0050 in-linedrillers for domestic well installation. During drilling, turbine flow meter with associated Model FM0208mud and cuttings were monitored in an open flow monitor manufactured by Turbines, Inc.atmosphere with a TVA-1000B Thermo-Scientific Drawdown during pumping was measured with aportable flame- and photo-ionization detector sonic water level sensor obtained from Eno Scientific,(FID/PID) for health and safety monitoring. Inc. (Model WS2010 PRO). The flow was split, withComparison of FID and PID readings (PID readings one portion going to waste and the other portionremained at background and are not sensitive to going to a flow-cell equipped with a YSI 5600methane) indicates the presence of methane at multiparameter probe to track stabilization of pHvarious intervals from ground surface in MW01 (<0.02 standard units per minute), oxidation-reduction(Figure 7). potential (<2 mV per minute), specific conductance (<1% per minute), dissolved oxygen (DO), and temperature. Purge volumes prior to sampling ranged from about 200 to 450 L (Phase III) and 1100 to 1250 L (Phase IV). Lower purge volumes in Phase III sampling were due to initial gas invasion into the screened intervals that caused cavitation and concern about prolonged pump operation. By the time of Phase IV sampling, disruptive gas invasion was no longer observed and extended purging was possible. Turbidity ranged from 1.7 to 29.7 Nephelometric Turbidity Units (NTUs) in domestic wells (Phase III and IV). Turbidity in MW01 was 7.5 NTUs in Phase III and 7.9 NTUs in Phase IV. Turbidity in MW02 was 28.8 NTUs in Phase III and 24.0 NTUs in Phase IV. Turbidity measurements in MW01 and MW02 could be impacted by gas exsolution. A photographic log of deep monitoring and domestic well sampling is provided in Appendix D. In April 2011, the static water level in MW01 prior to Figure 7. Resistivity as a function of depth in MW01 purging was 61.2 m (200.8 ft) below the top of the and MW02. MW01 and MW02 were screened at 233 - 239 m and 293 - 299 m bgs, respectively, casing (BTOC) measured using the Well Sounder 2010. corresponding to elevated resistivity and presence The initial pumping rate was approximately 27.6 of coarse-grained sandstone. FID readings in MW01 L/min. The pumping rate declined during purging to denote detections of methane during open air approximately 24.2 L/min as a result of the increasing logging of mud. FID monitoring at MW02 was sporadic and is not illustrated here. depth to water. At approximately 30 min after the11

27. DRAFTstart of purging, the pumping rate was reduced using An example of flow-cell readings through the purgingan in-line valve to 7.6 L/min. This resulted in of well MW02 is shown in Figure 9. The electrodeapproximately 18.2 m (60 ft) of rebound in the water readings show fairly rapid equilibration of pH andlevel within the well at the start of sampling (Figure 8). dissolved oxygen. Oxidation-reduction potentialGiven that the screen length is only 6.1 m (20 ft) and steadily decreased with the rate of change falling intothat the pump was set approximately 0.6 m (2 ft) the desired range (<2 mV per minute) by the end ofabove the screen, this indicates that ground water purging. Specific conductance readings were typicallyobtained during sampling was derived from the variable, likely due to continuous off-gassing andformation with no component of casing storage. The bubble formation within the conductivity sensor.total volume of water purged at the start of sampling After field measurements stabilized, ground waterwas approximately 1117 L. The static water level in was collected into sample bottles as summarized inMW02 prior to purging was 80.5 m (264.2 ft) BTOC Table B1. Samples were collected for a wide range ofmeasured using the Well Sounder 2010 (April 2011). inorganic, organic, and stable isotope analyses. A 500The initial pumping rate was approximately 18.9 mL sample was collected for field determinations ofL/min. The Eno Scientific well sounder was unable to alkalinity, turbidity, ferrous iron, and dissolved sulfide.measure the depth to water during most of the Alkalinity was determined onsite by incrementalpurging cycle perhaps due to a more rapid rate of titration of ground water with sulfuric acid. Turbiditydecline in the water level in the casing. Sampling was measurements were made with a portable meterinitiated after approximately 1249 L of water were (Hach 2100Q). Measurements were made forremoved. The pump cavitated after approximately dissolved sulfide and ferrous iron using the methylene1287 L were purged. The pump was subsequently blue and 1,10-phenanthroline colorimetric methods,stopped, allowed to cool, and restarted approximately respectively (APHA 1998a,b). Samples collected for10 min later to complete the sampling. dissolved gases, volatile organic compounds, semivolatile organic compounds, diesel-range organics, gasoline-range organics, glycols, low molecular weight acids, and δ13C/δD of methane were not filtered. δD is defined as (  2 H 1H sample δD ( ‰) =  2 1  ) − 1 x1000 (  H H standard   ) where the standard is the Vienna Standard Mean Ocean Water Standard (VSMOW). Samples collected for metals, anions, nutrients, dissolved organic carbon, dissolved inorganic carbon (DIC), δ13C of dissolved inorganic carbon, and δ18O/δD of water were filtered onsite using 0.45-micron pore-size, disposable-capsule filters. δ18O is defined as Figure 8. Variation of water level as a function of time in MW01 during Phase IV well purging. The initial pumping rate was 24.2 L/min. After approximately 30  δ18O ( ‰) =  ( 18 O 16 ) O sample  − 1 x1000 minutes of purging, the flow rate was decreased to 7.6   (18 O 16 O ) standard   L/min. This reduced flow rate caused partial recovery of the water level and confirmation that formation where the standard is the VSMOW. water was being accessed. 12

28. DRAFT Figure 9. Flow-cell readings as a function of time for specific conductance, dissolved oxygen, pH, and oxidation-reduction potential (well MW02, Phase IV sampling).13

29. DRAFTSample preservation and holding time criteria are Duplicate samples were collected in three locationslisted in Table B1. Field quality control (QC) samples during Phase III and Phase IV sampling activities.are summarized in Table B2. These included several Results for the duplicate analyses are presentedtypes of blanks, duplicate samples, and field matrix- Tables B13 and B14. Relative percent differencesspike samples. All of these QC sample types were (RPDs) were generally less than 10% for mostcollected, preserved, and analyzed using identical inorganic constituents indicating very good precision.methodologies as used for the water samples RPD is defined ascollected in the field (Table B1). Qualityassurance/quality control (QA/QC) requirements for  x1 − x2  RPD =   x1000analysis of metals and major ions are summarized in  ( x1 + x2) / 2   Table B3. QA/QC requirements for analysis ofdissolved gases, DIC/DOC, VOCs, low molecular weight where x1 = sample and x2 = sample duplicate. RPDsacids and stable isotopes of water are summarized in for methane, volatile organic compounds, and semi-Table B4. QA/QC requirements for analysis of volatile organic compounds were generally less thansemivolatile organic compounds (SVOCs), GRO, and 25% (Table B14). The lower reproducibility for theseDRO are summarized in Table B5. QA/QC compounds detetected in MW02 is likely due torequirements for analysis of glycols are summarized in difficulties in sampling and preserving water that isTable B6. Results of Phase III and Phase IV blank oversaturated in gas.samples are provided in Tables B7 to B12. Detections Major ions were quality checked by calculating ionobserved in the blank samples were generally very balances. The AqQA (v.1.1.1) software package waslow-level and generally much lower than used to evaluate cation/anion balance, which rangedconcentrations measured in the deep monitoring from <0.1 to 17.2% with 90% of the calculatedwells. Some blank samples showed detections of balances better than 5%.acetone (1 μg/L), m,p-xylene (up to 0.7 μg/L), toluene(up to 0.5 μg/L), benzoic acid (3 μg/L), and Geochemical equilibria in ground water weretetraethylene glycol (3 μg/L). Concentrations of these evaluated with the Geochemist’s Workbench packageanalytes in MW01 and MW02 in Phase III and Phase IV (version 8; Bethke 1996). Speciation and mineralsampling ranged from: 80 to 641 μg/L (acetone), non- equilibria calculations were made by entering thedetect to 750 μg/L (total xylenes), 0.6 to 617 μg/L concentrations of major cations (Na+, K+, Ca2+, Mg2+),(toluene), 209 to 457 μg/L (benzoic acid), and 7 to 27 anions (Cl-, SO42-, HCO3-), pH, and temperature. Forμg/L (tetraethylene glycol). Detected concentrations domestic well samples, bicarbonate concentrationsof toluene (Phase III), xylene (Phase IV), and were determined from alkalinity measurements. Fortetraethylene glycol (Phase IV) in MW01 are within the deep monitoring wells, because alkalinity includedabout 2 times the detected levels of these chemicals a significant contribution from hydroxide,in some of the applicable blank samples. concentrations of dissolved inorganic carbon wereConsequently, reported detections and concentrations used for bicarbonate/carbonate input. Activityof these chemicals in MW01 were used cautiously in corrections were made using the Debye-Hückelthis study. In one of the six blank samples collected equation. The LLNL (EQ3/6) thermodynamic databasefor DRO, an elevated concentration of 135 μg/L or 6 was selected for use in the calculations (Delany andtimes the reporting limit was observed (Table B12); all Lundeen 1990). Model simulations were alsoother DRO blank samples were non-detects (<20 conducted by tracing alkaline-addition titration paths.μg/L). Concentrations of DRO in the deep monitoring In order to do this, an additional entry was made towells ranged from 634 to 4050 μg/L. 14

30. DRAFTthe thermodynamic database describing the solubility TVA-1000B portable FID/PID and a 10 L Plexiglasof KOH (log K= 24.9; KOH(s) + H+ = K+(aq) + H2O(l)). sparge cell (Figure 10). Samples from domestic wells were routed through a closed (no contact with theAudits of Data Quality (ADQs) were conducted by a atmosphere to avoid offgassing) sample train andcontractor (independent of this investigation) or an collected in 0.5 L Cali-5 Bond gas sample bags.EPA QA Manager for all analyses conducted outside Ultrapure N2 gas was introduced into the bags andEPAs Contract Laboratory Program (CLP) with the placed on a rotary shaker for one hour prior toexception of data collected during Phase I, which is till headspace analysis on site using a portable GCin progress. This included data from EPAs Region VIII equipped with a thermal conductivity detector.laboratory in Golden, Colorado, EPA’s Region III Portable FID readings provided an immediatelaboratory in Fort Mead, MD, EPAs Office of Research indication of methane in well water prior to GCand Development Laboratory in Ada, Oklahoma, and analysis. Samples were also submitted to EPAs OfficeIsotech Laboratories in Champaign, Illinois. A of Research and Development (ORD) laboratory intechnical systems audit of Isotech Laboratories Ada, Oklahoma for analysis of dissolved gases.included an on-site visit by the independentcontractor and EPA QA Manager. Two on-site field Review of Borehole Geophysical Logstechnical system audits were also conducted by theindependent contractor and the EPA QA Manager to Borehole geophysical logs available on line fromensure compliance with the Category I (highest of four WOGCC were utilized to map lithology in the area oflevels in EPA) Quality Assurance Project Plan investigation. Depending upon the specific well,established for this site for ground water and gas various combinations of natural gamma, resistivity,sample collection. self-potential, density, and neutron porosity logs were utilized. Log resolution was sufficient to discernGas Sampling from Casing of Deep Monitoring distinct layers of shale 1 m or greater in thickness butWells in Phase III and IV not sufficient to differentiate coarse-, medium-, and fine-grained sandstones nor sandstones containingGas samples were collected from casing of deep various proportions of shale. Descriptions of cuttingsmonitoring wells by connecting a 12.7 mm NPT logged during installation of deep monitoring wellsstainless-steel Swagelok quick-connect body and a and domestic wells obtained from a local driller wereSwagelok single-end shutoff stem to a 12.7 mm brass used for near surface description. Neither grain sizeball valve. The stem was connected to 6.35 mm nor proportions of shale in sandstone wereinternal diameter Tygon Masterflex tubing and a 0.5 differentiated in near surface sandstones to maintainliter Cali-5 Bond gas sampling bag equipped with a consistency with descriptions from geophysical logs.Leur-Fit ValveTM and a Leur-taper Quick-MateTM Lithology in the area of investigation is highly variableconnector. A Masterflex E/S portable peristaltic pump and difficult to correlate from borehole to borehole,was used to extract gas at 1 L/min. Samples were even for boreholes in close proximity to one anothercollected after stabilization (± 1%) of O2, CO2, and CH4 consistent with other observations in the Wind Riverreadings on a GEM-2000 Plus CES-LANDTEC portable Formation (Osiensky 1984). Sandstone and shalegas analyzer. layers appeared thin and of limited lateral extent, again consistent with previous observations ofDomestic Well Sampling for Methane Using a lithology in the Wind River Formation (Single 1969,Closed System in Phase IV Flores and Keighin 1993).During the Phase IV sample event, water fromdomestic wells was screened using a Thermo-Scientific15

31. DRAFT Figure 10. Schematic of closed (no contact to atmosphere) sampling train for domestic wells. Water flow from domestic well and into sparge cell was approximately 5 and 1 L/min respectively. Excess water bled through valve used for sampling prior to sample collection. Gas flow into sparge cell and portable FID/PID sparge cell was approximately 20 and 1 L/min. Excess air was bled through splitter above sparge cell.Review of Cement Bond/Variable Density LogsCement bond/variable density (CBL/VDL) logs, CBL/VDLs provide an average volumetric assessmentavailable for less than half of production wells, were of the cement in the casing-to-formation annularobtained online from WOGCC to evaluate well space and are considered low resolution toolsintegrity. Sporadic bonding is defined as an interval compared to ultrasonic imaging tool logs whichhaving an amplitude (mV) greater than A80 (EPA 1994) provide a high-resolution 360° scan of the condition ofwhere the casing-to-cement bond (Bybee 2007). Acoustic A80 = 100.2log A0 + 0.8log A100 imaging tools do not directly measure cement seal. and A80, A0, and A100 = amplitude at 80%, 0%, and Communication of fluids between intervals has been100% bond respectively. A0 typically corresponds to observed to occur despite indication of "good toamplitude in free pipe whereas A100 corresponds to excellent" cement bond on acoustic logs (Boyd et al.the best-bonded interval on the CBL. Examples of "no 2006). All CBL/VDLs available from WOGCC reflectcement", "sporadic bonding", and "good bonding" are pre-hydraulic fracturing conditions.provided in Appendix E. 16

32. DRAFT organic compounds including benzene and m, p-3.0 xylene were detected at concentrations up to 390 andResults and Discussion 150 µg/L, respectively (EPA 2010), indicating pits as a source of shallow ground water contamination in the area of investigation. EPA’s maximum concentrationGround Water and Soil Sample Results level (MCL) for benzene is 5 μg/L.Near Three Pits Inorganic Geochemistry There are at least 33 pits previously used forstorage/disposal of drilling wastes, produced water, Inorganic geochemical results for ground water (alland flowback fluids in the area of investigation. phases) are summarized in Table A2a and Figure 11.Discussions are ongoing with stakeholders to Major ion chemistry of ground water in the Pavilliondetermine the location, delineate the boundaries, and area varies as a function of aquifer depth. Shallowextent (areal and vertical) of contamination associated ground waters (< 31 m bgs) collected from drinkingwith these pits. The operator has initiated water wells and stock wells are near-neutral (pH 7.7 ±remediation of selected pit areas. Concentrations of 0.4, n = 19) (Figure 12) and display calcium-DRO, gasoline range organics (GRO), and total bicarbonate composition. With increasing depth,purgeable hydrocarbons (TPH) detected in soil ground water becomes moderately alkaline (pH 9.0 ±samples adjacent to three pits investigated in Phase II 1.0, n = 55) (Figure 12), and with only one exceptionwere as high as 5010, 1760, and 6600 mg/kg, (MW02), is dominated by sodium and sulfate as therespectively (EPA 2010). Concentrations of GRO, DRO, major cation/anion pair (Figures 11 and 12, Tableand TPH in ground water samples from shallow (4.6 m A2a). This gradient in pH and water chemistry likelybgs) monitoring wells were as high as 2.4, 39, and 3.8 arises from the wide-scale surface application ofmg/L, respectively (EPA 2010). A wide variety of irrigation water from the Wind River to support Figure 11. Durov diagram showing ground water chemistry trends obtained in Phase I - IV sampling events and the composition of irrigation water.17

33. DRAFTFigure 12. Depth trends of chloride, pH, sulfate, and potassium (filled black squares = domestic wells, filled red circles =monitoring wells). 18

34. DRAFTcrop growth since irrigation water appears to bicarbonate. Some residents have described therepresent an endmember composition (Figure 11). development of particulates in ground water samplesThe chemical alteration from bicarbonate-type collected and stored in glass jars. Precipitates thatrecharge water to sulfate-type ground water involves formed from PGDW05 ground water were analyzed bymultiple water-rock interactions, including salt powder X-ray diffraction and found to be dominantlydissolution, carbonate mineralization, and exchange of calcite. Because calcite has retrograde solubility,divalent cations for sodium (Morris et al. 1959). Total precipitation of calcite is possibly triggered bydissolved solids concentrations are <6000 mg/L in all warming calcite-saturated ground water to ambientground water samples collected to depths up to 296 m conditions.(Figure 11). The geochemistry of ground water from the deepSaturation indices of gypsum (CaSO4·2H2O) and calcite monitoring wells is distinctive from that in the(CaCO3), plotted against sulfate and calcium domestic wells. Chloride enrichment in monitoringconcentrations, are shown in Figure 13. The trend for well MW02 is 18 times the mean chloridegypsum saturation suggests that sulfate concentration (25.6 mg/L) observed in ground waterconcentrations in the aquifer are limited by the from domestic wells. Chloride enrichment in this wellsolubility of gypsum. Ground water is also close to is significant because regional anion trends tend toequilibrium with calcite which likely is an important show decreasing Cl concentrations with depth. Thecontrol on pH and concentrations of calcium and mean potassium concentration in domestic wells Figure 13. Saturation indices for (a) gypsum versus sulfate concentration and (b) calcite versus calcium concentration. Saturation Index is equal to the logarithm of the ratio of the ion activity product to the mineral solubility product. A Saturation Index of 0 corresponds to chemical equilibrium; values less than 0 and greater than 0 correspond to undersaturated and oversaturated conditions, respectively.19

35. DRAFTscreened to 244 m bgs is 3 mg/L, with 99% of values Geochemical modeling was carried out by using<10 mg/L. Potassium enrichment in MW01 and ground water compositions for PGDW49, PGDW20,MW02 is between 8.2 and 18.3 times the mean value and PGDW32 (initial pH 7.3, 8.9, and 9.9, respectively).of domestic wells (Table A2a). pH values in MW01 Modeled titration results are shown in Figure 15a; pHand MW02 are highly alkaline (11.2-12.0), above the is plotted versus the mass of KOH added per kg ofpH range observed in domestic wells (6.9-10), and solution. Model titration results vary as a function ofabove the pH range previously reported for the Wind ground water composition. Samples PGDW20 andRiver Formation (Plafcan et al. 1995, Daddow 1996). PGDW32 have Na-SO4-type compositions typical ofIn the deep monitoring wells, up to 94% of the total deeper portions of the aquifer. In both of these cases,alkalinity is contributed by hydroxide. In addition, the attainment of pH values between 11.2 and 12.0monitoring wells show low calcium, sodium, and requires small quantities of KOH addition (<250 mgsulfate concentrations compared to the general trend KOH per kg of solution). Sample PGDW49 is elevatedobserved in domestic well waters (Figure 14). in Ca2+ and Mg2+, lower in pH, and typical of shallower ground water compositions. In this case, significantlyThe high pH measured in the deep monitoring wells more KOH addition is required to attain pH valueswas unusual and unexpected. Although ground water observed in the monitoring wells. The first derivativepH in these wells was >11, total alkalinity was not of the titration curve, or buffer intensity, is shown inparticularly high (<500 mg/kg), and as already noted Figure 15b. The buffer intensity indicates that groundup to 94% of the total alkalinity was present as water compositions like PGDW20 and PGDW32hydroxide (see charge balance calculations, Table inherently have little resistance to pH change up toA2b). Alkalinity contributed by carbonate/bicarbon- about pH 12, at which point increased KOH additionsate was less than the hydroxide component. In fact, are necessary to further increase pH. PGDW49 showsinorganic carbon concentrations were so low in MW02 a broad peak on the buffer intensity diagram (pH 10 toas to prevent the measurement of δ13C of dissolved 11) which reflects precipitation reactions to forminorganic carbon. Presence of hydroxide alkalinity calcium carbonate and magnesium hydroxide,suggests strong base addition as the causative factor reactions that consume hydroxide and therefore limitfor elevated pH in the deep monitoring wells. The pH increases, until divalent cations are completelypossibility of cement/grout intrusion into the screened consumed. The model results clearly show thatintervals was considered as a possibility for both ground water typical of the Pavillion aquifer belowmonitoring wells, although precautions were taken to 100 m depth (Na-SO4-type composition) is especiallyprevent downward migration of cement during well vulnerable to the addition of strong base, with smallconstruction. Cement intrusion typically leads to pH KOH additions driving significant upward pH changes.values between 10 and 11, lower than the pH values Paired values of δ18O and δ2H in ground water samplesmeasured in the deep monitoring wells (Gibb et al. plot below the Global Meteoric Water Line (Figure 16;1987). Prolonged purging did not show decreasing pH -16.6 to -12.4‰ δ18O and -129.2 to -97.4‰ δ2H).trends (e.g., Figure 9) and water chemistry results Shallow ground water samples generally tend to beindicate that ground water from the wells was highly depleted in 18O and 2H compared to deeper groundundersaturated with respect to cement phases (e.g., water samples and may be more reflective of localportlandite), suggesting that cement was not the recharge. Ground water isotope data from the deepcause of elevated pH. monitoring wells (red circles, Figure 16) follow alongIn order to gain additional insight, reaction path the same δ18O versus δ2H trajectory established by themodeling was conducted to evaluate pH response to domestic well data, suggesting similar recharge andaddition of strong base (potassium hydroxide, KOH). evolutional paths (e.g., Bartos et al. 2008). 20

36. DRAFT Figure 14. Concentration trends versus specific conductivity. Note the monitoring wells show high pH and low sulfate, calcium, and sodium relative to the general trend observed in the domestic wells (filled black squares = domestic wells, filled red circles = monitoring wells).21

37. DRAFT Figure 15. (a) Results of KOH titration models plotted as pH versus grams of KOH added per kilogram of solution. Initial water compositions are from PGDW49, PGDW20, and PGDW32. Model accounts for reactions taking place in solution as KOH is added and equilibrated. pH range in deep monitoring wells shown for reference; (b) Buffer Intensity plot or first derivative of titration plot, pH versus change in concentration of base (CB) per change in pH. Figure 16. Hydrogen and oxygen isotope values (permil, Vienna Standard Mean Ocean Water, VSMOW) for ground water samples (black squares=domestic wells; red circles=deep monitoring wells) relative to the Global Meteoric Water Line from Craig (1961). 22

38. DRAFTOrganic Geochemistry for slickwater and carbon dioxide foam fracture jobs. Some MSDSs list chemical formulation as proprietaryOrganic and inorganic geochemical impacts in deep (e.g., proprietary alcohols) or list a chemical familyground water monitoring wells (Phase III and IV) are (e.g., blend of organic surfactants) renderingsummarized in Table 3. The monitoring wells produce identification of constituents impossible. This reviewground water near-saturated in methane at ambient is summarized in Table 4. Inorganic additives arepressure, with concentrations up to 19.0 mg/L. Gas potential sources of elevated K, Cl, and OH in deepexsolution was observed while sampling at both monitoring wells.MW01 and MW02. A wide variety of organicchemicals was detected in the monitoring wells Detection of compounds associated with petroleum-including: GRO, DRO, BTEX, trimethylbenzenes, based additives in ground water samples usingphenols, naphthalenes, acetone, isopropanol, TBA, 2- analytical methods employed in this investigationbutoxyethanol, 2-butanone, diethylene glycol, would be manifested as GRO, DRO, BTEX,triethylene glycol, and tetraethylene glycol (Figure 17; naphthalenes, and trimethylbenzenes observed inTable 3). Concentrations of these chemicals range deep monitoring wells.from μg/L to mg/L levels. Concentrations of benzene TBA was detected in MW02 during Phase 4 samplingin MW02 exceed EPA’s MCL in drinking by a factor of at a concentration of 4470 μg/L. Two possible49 times. Detections of organic chemicals are more formation pathways for TBA are: 1) biodegradation ofnumerous and exhibit higher concentrations in the methyl tert-butyl ether (MTBE, synthetic chemicaldeeper of the two monitoring wells (Figure 17, Table used as a fuel additive) under methanogenic3). This observation, along with trends in methane, conditions (e.g., Mormile et al. 1994, Bradley et al.potassium, chloride, and pH, suggest a deep source 2001); and 2) breakdown of tert-butyl hydroperoxide(>299 m bgs) of contamination. Natural breakdown (a gel breaker used in hydraulic fracturing; e.g., Hiattproducts of organic contaminants like BTEX and et al. 1964). TBA biodegradation is generally slowglycols include acetate and benzoic acid; these compared to the degradation of MTBE; this suggestsbreakdown products are more enriched in the that TBA could be present and persist even aftershallower of the two monitoring wells, suggesting complete MTBE removal from ground water impactedupward/lateral migration with natural degradation by fuel releases (Wilson et al. 2005). MTBE was notand accumulation of daughter products (Corseuil et al. detected in either of the deep monitoring wells. A2011, Caldwell and Suflita 2000, Dwyer and Tiedje second pathway of TBA production is from the1983). Other trace-level detections of semi-volatile decomposition of the gel breaker tert-butylorganic compounds included: bis(2-ethylhexyl) hydroperoxide. Hiatt et al. (1964) found thatphthalate (MW01 and MW02, Phase III and IV), bis(2- decomposition of tert-butyl hydroperoxide yielded achloroethyl) ether, bis(2-ethylhexyl) adipate (MW01, 10-fold molar quantity of TBA, oxygen, di-tert-butylPhase IV), butyl benzyl phthalate, and 4-methyl-2- peroxide, and acetone. Acetone was detected inpentanone (MW02, Phase IV). MW02 during Phase 4 sampling at a concentration ofWell completion reports obtained online from WOGCC 641 μg/L. This breaker is used in hydraulic fracturingand Material Safety and Data Sheets (MSDSs) formulations; however, the MSDSs made available toobtained from the operator were reviewed to EPA do not indicate whether tert-butyl hydroperoxideexamine inorganic and organic compounds in was used in the Pavillion gas field for well stimulation.additives used for hydraulic fracturing and similarity Elevated concentrations of TBA are not expected inwith detected elements and compounds in ground unimpacted aquifers and its presence in MW02water. Well completion reports were limited to a remains unresolved. Additional insight about thesubset of production wells and included dates of occurrence of TBA (and other organic compounds)injection, injection depths, pressure, flow, and volume might be obtained by conducting compound-specific isotope analyses.23

39. DRAFT 24

40. DRAFT Figure 17. Organic compounds detected in deep monitoring wells MW01 and MW02 during Phase III and IV sampling events. Horizontal bars show method reporting limits for the individual analytes.25

41. DRAFT 26

42. DRAFTNatural gas condensates are composed primarily of reporting limit was over an order of magnitude lower,aliphatic hydrocarbons; however, condensates may suggesting that Method 8015 is prone to false positivecontain low quantities of aromatic compounds, such results (possibly due to interactions between theas BTEX. Gas from the Fort Union and lower Wind chromatographic column and organic compounds inRiver Formations is generally dry (C1/C1-C5 = 0.95 - sample water). This result points to the need for0.96 where methane = C1, ethane = C2, propane = C3, continued and future improvements of analyticalbutane = C4, pentane = C5) (Johnson and Rice 1993) methods to detect and quantitate low levels ofand unlikely to yield liquid condensates at ground organic chemicals that may be associated withwater pressure and temperature conditions. In hydraulic fracturing fluids. Although contaminationaddition, a condensate origin for BTEX compounds in was detected in some domestic wells proximal to theground water is doubtful because dissolved gas deep monitoring wells, underscoring potential futurecompositions and concentrations are similar between risk, the existing data at this time do not establish athe two deep monitoring wells and therefore would definitive link between deep and shallowyield similar liquid condensates, yet the compositions contamination of the aquifer. An increased number ofand concentrations of organic compounds detected in sampling points (monitoring wells) with verticalthese wells are quite different (Figure 17) further profiling in targeted locations are necessary to bettersuggesting a deep source of BTEX in MW02. The define transport and fate characteristics of organicpresence of synthetic compounds such as glycol and inorganic contaminants in the ground waterethers, along with enrichments in K, Cl, pH, and the system and impact on domestic wells.assortment of other organic components is explainedas the result of direct mixing of hydraulic fracturing Natural Gas Migrationfluids with ground water in the Pavillion gas field. A review of open-hole geophysical logs obtained fromAs noted previously, this investigation was prompted the WOGCC internet site indicates the presence ofby homeowner complaints over perceived changes in gas-filled porosity at three locations at 198, 208, andwater quality. Domestic well results showed: the 252 m bgs between the years 1965 - 1973 suggestingpresence of DRO and GRO (in 23 of 28 samples), and the presence of natural gas in ground water at depthstrace levels of exotic organic compounds in some used for domestic water supply prior to extensivedomestic wells including adamantanes, 2- commercial development. However, a review of 10butoxyethanol phosphate, phenols, naphthalene, and mud-gas logs recorded in the mid-1970s and earlytoluene (EPA 2009, EPA 2010). Methane was detected 1980s obtained on line from WOGCC, do not indicatein 10 of 28 samples at concentration levels below 0.8 gas shows within 300 m of the surface at any location.mg/L. Foul odors associated with some domestic wellscorrelate with detections of GRO and DRO. Aqueous analysis of light hydrocarbons, gas andAnomalous trends in inorganic constituents observed headspace analysis of light hydrocarbons, and isotopicin the deep monitoring wells (e.g., K, Cl, pH) were not data for dissolved, gas phase, and headspace analysisrevealed in domestic well waters. In several instances, are summarized in Tables A3a, A3b, and A3cglycols were detected in domestic wells using gas respectively (all investigative phases). Elevated levelschromatography with flame ionization detection (GC- of dissolved methane in domestic wells generallyFID; EPA Standard Method 8015). However, glycol increase in those wells in proximity to gas productionanalysis using liquid chromatography with tandem wells (Figure 18c). Methane was not detected inmass spectroscopy (GC/MS/MS) failed to replicate shallow domestic wells (e.g., < 50 m) regardless ofthese glycol detections, even though the method proximity to production wells (Figure 18c). With the exception of two domestic wells where methane was27

43. DRAFTFigure 18. (a) Stable isotope ratios of carbon of methane versus ratio of methane (C1) to ethane (C2) and propane (C3) ingas from production wells, monitoring wells, and domestic wells. Values of 100,000 are used to denote non detection ofethane and propane in samples. (b) Stable isotope ratios of carbon versus hydrogen of methane in gas from productionwells (both literature and measured values), monitoring wells, and domestic wells. δD was not determined for PGDW32. 13Oxidation pathway (enrichment of C of remaining CH4 with biodegradation) is illustrated. (c) Methane concentration indomestic (red circles and black squares) and monitoring wells (green squares) as a function of proximity to productionwells and AMSL. Values of 1.0 were used for non-detection (detection limit 5 µg/L). 28

44. DRAFTdetected at less than 22 μg/L, methane was not similar to produced gas and have undergone littledetected in domestic wells with 2 or less production oxidation or biodegradation. These observationswells within 600 m (Figure 18c). All domestic wells combined with radiocarbon analysis of CH4 (< 0.2%with the exception of PGDW25 with 2 or less percent modern carbon) obtained from gas in casingproduction wells within 600 m are located on the of both MW01 and MW02 indicate that methane inperiphery of the gas field (Figure 5). PGDW25 is deep monitoring wells is of thermogenic origin. Gaslocated within 1600 m of 15 gas production wells. from the Fort Union and lower Wind River Formations is isotopically heavy (δ13C-CH4 from to -40.24 to -Of particular interest is the area encompassing MW01, 38.04‰) and as previously stated, dry (Johnson andPGDW30, and PGDW05 (Figure 19). Ground water is Rice 1993, Johnson and Keighin 1998). Values of δ13C-saturated with methane at MW01 which is screened CH4 and δD -CH4 more negative than -64‰ and -at a depth (239 m bgs) typical of deeper domestic 175‰, respectively, are indicative of microbial originwells in the area. Methane was detected in PGDW30 (Schoell 1980). The absence of ethane and propane inat 808 μg/L at a depth of only 80 m, the highest level three of four domestic wells having sufficient methanein any domestic well. A blowout occurred during to allow isotopic analysis and a shift of δ13C-CH4 anddrilling at a depth of only 159 m bgs in December 2005 δD-CH4 values in a positive direction relative toadjacent to PGDW05. Natural gas exited the borehole produced gas suggests the presence of gas offor three days until the gas field operator was ordered thermogenic origin in domestic wells undergoingto plug the borehole with a dense mud. The owner of biodegradation and subsequent enrichment of δ13CPGDW05 was attempting at the time to replace this and δD. This observation is consistent with a patternwell due to taste, odor, and yield reduction he stated of dispersion and degradation with upward migrationoccurred after hydraulic fracturing at nearby observed for organic compounds. Values of δ13C-CH4production wells. A mud-gas log conducted on more positive than -64‰ and C1/(C2+C3) ratios above11/16/1980 at Tribal Pavillion 14-2 (illustrated on 1000 are often interpreted to indicate gas of mixedFigure 19 as 14-2) located only 300 m from the biogenic-thermogenic origin or gas of biogenic originlocation of the uncontrolled release does not indicate undergoing biodegradation (Whiticar 1999, Whiticara gas show (distinctive peaks on a gas chromatograph) and Faber 1986) since neither ethane nor propane arewithin 300 m of the surface. The owner of PGDW05 biogenically generated in significant amounts.complained that well yield decreased after hydraulic However, preferential loss of ethane and propanefracturing at nearby production wells. Records relative to methane in thermogenic gas produces aobtained from the Wyoming State Engineers office similar response (Valentine 2010, Kinnaman et al.dated January 1973 indicate a yield of 30 to 38 L/min 2007).with 1.2 meters of drawdown after 10 hours ofpumping. During a sampling event in April 2005, Evaluation of Cement Bond/Variable DensityPGDW05 became dry after pumping at a rate of 21.6 Logs Along TransectL/min for 14 minutes. The cause of reduced well yieldrequires further investigation. CBL/VDLs and lithology were examined along a transect (Figure 19) which included the deepSimilarity of δ13C values for methane, ethane, monitoring wells and three domestic wells wherepropane, isobutane, and butane between gas elevated levels of methane were detected. Atproduction and monitoring wells and plots of δ13C-CH4 Pavillion Fee 34-03B, a CBL/VDL conducted onversus δD -CH4 (Figure 18b) and δ13C-CH4 versus C1/(C2 10/22/2004 indicates no cement below surface casing+ C3) (Figure 18a) indicate that light hydrocarbons in until 802 m msl (Figure 20) and sporadic bonding tocasing and dissolved gas in deep monitoring wells are 604 m msl (not illustrated). The well completion29

45. DRAFT Figure 19. Map illustrating transect used to develop lithologic cross section and evaluation of CBL/VDLs.report for this production well indicates that hydraulic reports and CBL/VDLs indicates instances of sporadicfracturing was performed at 601 m msl on 11/9/2004. bonding directly above intervals of hydraulicA cement squeeze was subsequently performed at fracturing. This review also indicates instances where802 m msl on 4/1/2005 (no CBL/VDL after cement cement outside production casing is lacking over ansqueeze) with hydraulic fracturing at 689 m msl on extensive interval providing a potential conduit for4/19/2005. At Pavillion Fee 34-03R, the CBL/VDL fluid migration to within 300 m of the surface. Asindicates no cement below surface casing until 968 m graphically illustrated in Figure 20, production wellsmsl (Figure 20). At Tribal Pavillion 41-10 and 41-10B, having no or sporadic cement outside productionCBL/VDLs indicate sporadic bonding over extensive casing are located in proximity to deep monitoringintervals. A CBL/VDL conducted on 4/20/2005 at wells where aqueous constituents consistent withTribal Pavillion 24-02 after a squeeze perforation at hydraulic fracturing were detected and methanethe base of the surface casing indicates poor bonding exsolved from solution during sampling and locationsoutside production casing below surface casing to the of domestic wells where elevated levels of methanefirst perforation interval (Figure 20). At Tribal Pavillion were detected and where an uncontrolled release of11-11B, a CBL/VDL indicates poor or sporadic bonding natural gas occurred.to 991 m bgs and no cement or cement bridging from675 - 857 m msl. Thus, a review of well completion 30

46. 31 DRAFT Figure 20. Lithologic cross-section along transect illustrating production wells (with evaluation of CBL/VDLs), domestic wells, and blowout location. Red arrows denote depths of hydraulic fracturing of unknown areal extent. Sandstone units are undifferentiated between fine, medium and coarse-grained units.

47. DRAFTPotential Migration Pathways fracturing generates new fractures or enlarges existing ones above the target formation, increasing theFurther investigation is necessary to determine connectivity of the fracture system.mechanisms of aqueous and gas phase transport inthe area of investigation. However, at least three In all three transport pathways, a general correlationmechanisms can be postulated at this time. The first (spatial relationships ultimately determined by faultmechanism is aqueous and/or gas transport via and fracture systems in addition to lithology) wouldboreholes due to insufficient or inadequate cement exist between proximity to gas production wells andoutside production casing. Both aqueous (brine) and concentration of aqueous and gas phase constituentsgas phase migration vertically up compromised in ground water. For instance, Osborn et al. (2011)wellbores have been simulated (Nordbotten et al. observed a correlation between methane2004, 2005a, 2005b) and indicate decreasing mass flux concentration and proximity to hydraulically fracturedtoward the surface with increasing number of gas production wells at locations above the Marcelluspermeable formations encountered along the way. and Utica formations in Pennsylvania and New York.Thus, the severity of ground water contamination Isotopic data and other measurements for methane inincreases with depth. Migration of gas via wellbores is the drinking water were consistent with gas found inwell documented in the literature (e.g., Harrison 1983, deep reservoirs such as the Marcellus and Utica shalesHarrison 1985, Van Stempvoort et al. 2005, Taylor et at the active sites and matched gas geochemistry fromal. 2000). In Bainbridge, Ohio, an operator initiated shale-gas wells sampled nearby. Also, in all threehydraulic fracturing despite knowing that only 24 m of transport pathways, advective/dispersive transportcement was present above the perforation interval would be accompanied by degradation causing a(Bair et al. 2010, ODNR 2008). Hydraulic fracturing vertical chemical gradient as observed duringfluid flowed to the surface via surface-production sampling of MW01 and MW02. Reduced mass flux tocasing annulus which pressurized upon shut-in. Gas the near surface environment and subsequentsubsequently migrated through natural fractures to degradation along vertical and lateral transportdomestic wells eventually causing an explosion at one pathways would explain lack of detection in domestichome. In northeastern Pennsylvania, two operators wells of compounds observed in MW02.were fined for enhanced gas migration into domesticwells attributed to incomplete or inadequate cementoutside production casing in wells used for hydraulicfracturing (PADEP 2009a, 2009b, 2010).The second mechanism is fracture fluid excursion fromthin discontinuous tight sandstone units intosandstone units of greater permeability. This wouldbe accompanied by physical displacement of gas-richsolutions in both tight and more permeable sandstoneformations. As illustrated in Figure 20, there is littlelateral and vertical continuity to hydraulicallyfractured tight sandstones and no lithologic barrier(laterally continuous shale units) to upward verticalmigration of aqueous constituents of hydraulicfracturing in the event of excursion from fractures. Athird mechanism is that the process of hydraulic 32

48. DRAFT complex scientific investigations. In this approach,4.0 individual data sets and observations are integrated toConclusions formulate an explanation consistent with each data set and observation. While each individual data set or observation represents an important line of reasoning,The objective of this investigation was to determine taken as a whole, consistent data sets andthe presence of ground water contamination in the observations provide compelling evidence to supportWind River Formation above the Pavillion gas field and an explanation of data. Using this approach, theto the extent possible, identify the source of explanation best fitting the data for the deepcontamination. The combined use of shallow and monitoring wells is that constituents associated withdeep monitoring wells allowed differentiation hydraulic fracturing have been released into the Windbetween shallow sources of contamination (pits) and River drinking water aquifer at depths above thedeep sources of contamination (production wells). current production zone.Additional investigation is necessary to determine the Lines of reasoning to support this explanation consistareal and vertical extent of shallow and deep ground of the following.water contamination. 1. High pH valuesDetection of high concentrations of benzene, xylenes,gasoline range organics, diesel range organics, and pH values in MW01 and MW02 are highlytotal purgeable hydrocarbons in ground water alkaline (11.2-12.0), above the pH rangesamples from shallow monitoring wells near pits observed in domestic wells (6.9-10), andindicates that pits are a source of shallow ground above the pH range previously reported forwater contamination in the area of investigation. Pits the Wind River Formation with up to 94% ofwere used for disposal of drilling cuttings, flowback, the total alkalinity contributed by hydroxide.and produced water. There are at least 33 pits in the The presence of hydroxide alkalinity suggestsarea of investigation. When considered separately, addition of base as the causative factor forpits represent potential source terms for localized elevated pH in the deep monitoring wells.ground water plumes of unknown extent. When Reaction path modeling indicates that sodium-considered as whole they represent potential broader sulfate composition ground water typical ofcontamination of shallow ground water. A number of deeper portions of the Wind River Formationstock and domestic wells in the area of investigation provides little resistance to elevation of pHare fairly shallow (e.g., < 30 m) representing potential with small addition of potassium hydroxide.receptor pathways. EPA is a member of a stakeholdergroup working with the operator to determine the With the exception of soda ash, the pH ofareal and vertical extent of shallow ground water drilling additives in concentrated aqueouscontamination caused by these pits. The operator of solution was well below that observed in thethe site is currently engaged in investigating and deep monitoring wells. Dense soda ash wasremediating several pit areas. added to the drilling mud which variedDetection of contaminants in ground water from deep between pH 8 - 9.sources of contamination (production wells, hydraulicfracturing) was considerably more complex than The possibility of cement/grout intrusion intodetection of contaminants from pits necessitating a the screened intervals was considered as amultiple lines of reasoning approach common to possibility for elevated pH in both monitoring33

49. DRAFT wells. However, cement intrusion typically Potassium metaborate was used in leads to pH values between 10 and 11 – below crosslinkers (5-10%, 30-60%). Potassium that observed in deep monitoring wells. hydroxide was used in a crosslinker (<5%) and Prolonged purging did not show decreasing in a solvent. Ammonium chloride was used in pH trends. Water chemistry results indicate crosslinker (1-27%). that ground water from the wells was highly undersaturated with respect to cement Alternative explanations for inorganic geochemical anomalies observed in deep phases (e.g., portlandite). monitoring wells have been provided and Material Safety Data Sheets indicate that considered. These alternate explanations include contamination from drilling fluids and potassium hydroxide was used in a crosslinker additives, well completion materials, and (<5%) and in a solvent. surface soil, with contamination from all these2. Elevated potassium and chloride sources exacerbated by poor well development. Contamination by drilling fluids The inorganic geochemistry of ground water and additives is inconsistent with analysis of from the deep monitoring wells is distinctive concentrated solutions of bentonite and additives. Well construction materials (screen from that in the domestic wells and expected composition in the Wind River formation. and sections of casing) consisted of stainless Potassium concentration in MW02 (43.6 steel and were power-washed on site with detergent-free water prior to use. Sections of mg/L) and MW01 (54.9 mg/L) is between 14.5 and 18.3 times the mean value of levels tremie pipe used to inject cement above observed in domestic wells (3 mg/L, 99% of screened intervals were also power washed with detergent-free water prior to use. values < 10 mg/L). Chloride enrichment in monitoring well MW02 (466 mg/L) is 18 times Stainless-steel screens and sections of casing the mean chloride concentration (25.6 mg/L) and tremie pipe remained above ground level (did not touch soil) prior to use. Both deep observed in ground water from domestic wells. Chloride concentration in this well is monitoring wells were purposefully located significant because regional anion trends away from the immediate vicinity of gas production wells, known locations of pits, and show decreasing chloride concentrations with depth. In addition, the monitoring wells show areas of domestic waste disposal (abandoned low calcium, sodium, and sulfate machinery) to minimize the potential of surface soil contamination. Conductor pipe concentrations compared to the general trend installed over the first 30.5 m (100 ft) of observed in domestic well waters. drilling at both deep monitoring wells Potassium levels in concentrated solutions of eliminated the possibility of surface soil entry into the borehole. Turbidity measurements in drilling additives were all less than 2 mg/L. One additive (Aqua Clear used during well MW01 during sampling ranged from 7.5 and development) contained 230 mg/L chloride in 7.9 Nephelometric Turbidity Units (NTUs). Turbidity measurements in MW02 during a concentrated solution. Information from well completion reports and Material Safety sampling ranged from 24.0 to 28.0 NTUs, Data Sheets indicate that the formulation of slightly above the stated goal of 10.0 NTUs but nevertheless was clear water typical of fracture fluid provided for foam jobs typically consisted of 6% potassium chloride. domestic wells during sampling. A low 34

50. DRAFT recharge rate in MW02 necessitated a a surfactant (30-60%), in breakers (<1%, 10- prolonged period of well development which 30%), and in foaming agents (<3%, 1-5%, 10- was likely due in part to gas flow (reduced 30%). Diethylene glycol was used in a foaming relative permeability to water) into the well agent (5-10%) and in a solvent (0.1-5%). during development. Triethylene glycol was used in a solvent (95- 100%). Material Safety Data Sheets do not 3. Detection of synthetic organic compounds indicate that tert-butyl hydroperoxide was used in the Pavillion gas field. The source of Isopropanol was detected in MW01 and this compound remains unresolved. However, MW02 at 212 and 581 μg/L, respectively. tert-butyl alcohol is not expected to occur Diethylene glycol was detected in MW01 and naturally in ground water. Material Safety MW02 at 226 and 1570 μg/L, respectively. Data Sheets do not contain proprietary Triethylene glycol was detected in MW01 and information and the chemical ingredients of MW02 at 46 and 310 μg/L, respectively. many additives. Another synthetic compound, tert-butyl alcohol, was detected in MW02 at a Alternative explanations provided to date and concentration of 4470 μg/L. Tert-butyl alcohol considered by EPA for detection of synthetic is a known breakdown product of methyl tert- organic compounds in deep monitoring wells butyl ether (a fuel additive) and tert-butyl include arguments previously listed and hydroperoxide (a gel breaker used in hydraulic addressed. fracturing). EPA methods were utilized for analysis when applicable for compounds or 4. Detection of petroleum hydrocarbons classes of compounds. Detection of synthetic organic compounds in MW01 and MW02 was Benzene, toluene, ethylbenzene, and xylenes made in part through the use of non- (BTEX) were detected in MW02 at concentrations of 246, 617, 67, and 750 μg/L commercially available modified EPA analytical methods. For instance, high respectively. Trimethylbenzenes were performance liquid chromatography/mass detected in MW02 at 105 μg/L. Gasoline range organics were detected in MW01 and spectrometry/mass spectrometry was utilized for analysis of diethylene, triethylene and MW02 at 592 and 3710 μg/L, respectively. tetraethylene glycols. Ethylene glycol, which Diesel range organics were detected in MW01 and MW02 at 924 and 4050 μg/L respectively. was widely used for well stimulation, required additional method modification and was not Naphthalene was detected in MW02 at 6 analyzed during this investigation. μg/L. EPA methods were utilized for analysis. BTEX and trimethylbenzenes were not Isopropanol was detected in concentrated detected in concentrated solutions of drilling solutions of drilling additives at a maximum additives. concentration of 87 μg/L, well below that detected in deep monitoring wells. Glycols Material Safety Data Sheets indicate that were not detected in concentrated solutions aromatic solvent (typically BTEX mixture) was of drilling additives. used in a breaker (<75%). Diesel oil (mixture of saturated and aromatic hydrocarbons Material Safety Data Sheets indicate that including naphthalenes and alkylbenzenes) isopropanol was used in a biocide (20-40%), in was used in a guar polymer slurry/liquid gel35

51. DRAFTconcentrate (30-60%) and in a solvent (60- pumps would have to contain high levels of100%). Petroleum raffinates (a mixture of water extractable petroleum compounds andparaffinic, cycloparaffinic, olefinic, and consist of a matrix allowing rapid massaromatic hydrocarbons) were used in a transfer, neither of which is plausible.breaker (<30-60%). Heavy aromaticpetroleum naphtha (mixture of paraffinic, Another alternate explanation is thatcycloparaffinic and aromatic hydrocarbons) detection of petroleum hydrocarbons in ground water is expected above a natural gaswas used in surfactants (5-10%, 10-30%, 30-60%) and in a solvent (10-50%). Toluene was field. Gas from Fort Union and Wind Riverused in a flow enhancer (3-7%). Xylenes were Formations is dry and unlikely to yield liquid condensates at ground water pressure andused in a flow enhancer (40-70%) and a temperature conditions. In addition, abreaker (confidential percentage). Gasolinerange organics correspond to a hydrocarbon condensate origin for petroleumrange of C6 – C10. It includes a variety of hydrocarbons in ground water is doubtful because dissolved hydrocarbon gasorganic compounds ketones, ethers, mineralspirits, stoddard solvents, and naphthas. compositions and concentrations are similarDetection of gasoline range organics does not between the two deep monitoring wells and therefore would yield similar liquidinfer the use of gasoline for hydraulic condensates, yet the compositions andfracturing. concentrations of organic compoundsAlternative explanations provided to date and detected in these wells are quite different.considered by EPA for detection of petroleumcompounds in deep monitoring wells include 5. Breakdown products of organic compoundsarguments previously listed and addressed. An Detections of organic chemicals were moreadditional alternate explanation for detection numerous and exhibited higherof petroleum compounds includes use of concentrations in the deeper of the twolubricants on the drillstem and well casing, use monitoring wells. Natural breakdown productsof electrical tape on submersible pumps, and of organic contaminants like BTEX and glycolscomponents of submersible pumps. Jet Lube include acetate and benzoic acid. TheseWell Guard hydrocarbon free lubricant breakdown products are more enriched in thespecifically designed for monitoring well shallower of the two monitoring wells,installation was used for drillstem suggesting upward/lateral migration withconnections. No lubricants were used to natural degradation and accumulation ofattach sections of casing or sections of tremie daughter products.pipe during cementation. Clamps, notelectrical tape, were used to bind electrical Hydraulic gradients are currently undefined inwires for submersible pumps. Water collected the area of investigation. However, there arefor samples during recharge at MW01 and flowing stock wells (e.g., PGDW44 - one of theMW02 would have a short contact time with deepest domestic wells in the area ofcomponents of submersible pumps. For investigation at 229 m below ground surface)components of submersible pumps to be a suggesting that upward gradients exist in thecausative factor of high concentrations of area of investigation. In the Agencys reportpetroleum hydrocarbons observed in MW01 on evaluation of impacts to USDWs byand MW02, components of submersible hydraulic fracturing of coalbed methane 36

52. DRAFT reservoirs (EPA, 2004), hypothetical on Tribal Pavillion 24-02 after a squeeze job at conceptual models were presented on the base of the surface casing indicates contaminant migration in a USDW during sporadic bonding outside production casing injection of fracturing fluids into a USDW. In below surface casing to the interval of these conceptual models, highly concentrated hydraulic fracturing at 469 m (1538 ft) below contaminant plumes exist within the zone of ground surface. At Tribal Pavillion 11-11B, a injection with dispersed lower concentration cement bond/variable density log indicates areas vertically and laterally distant from sporadic bonding between 305 to 503 m injection points. Data from deep monitoring (1000 to 1650 ft) below ground surface with wells suggests that this conceptual model may hydraulic fracturing occurring at 463 m (1516 be appropriate at this site. ft) below ground surface.6. Sporadic bonding outside production casing 7. Hydraulic fracturing into thin discontinuous directly above intervals of hydraulic fracturing sandstone units It is possible that wellbore design and integrity There is little lateral and vertical continuity to issues were one causative factor in deep hydraulically fractured tight sandstones and ground water contamination at this site no lithologic barrier (laterally continuous shale (surface casing of production wells not units) to stop upward vertical migration of extending below deepest domestic wells, little aqueous constituents of hydraulic fracturing in vertical separation between fractured zones the event of excursion from fractures. and domestic wells, no cement or sporadic Sandstone units are of variable grain size and bonding outside production casing). permeability indicating a potentially tortuous path for upward migration. A review of well completion reports and cement bond/variable density logs in the area In the event of excursion from sandstone around MW01 and MW02 indicates instances units, vertical migration of fluids could also of sporadic bonding outside production casing occur via nearby wellbores. For instance, at directly above intervals of hydraulic fracturing. Pavillion Fee 34-03R, the cement For instance, at Pavillion Fee 34-03B, a cement bond/variable density log indicates no cement bond/variable density log conducted on until 671 m (2200 ft) below ground surface. 10/22/2004 indicated no cement until 838 m Hydraulic fracturing occurred above this depth (2750 ft) and sporadic bonding to 1036 m at nearby production wells. (3400 ft) below ground surface. The well completion report for this production well Although some natural migration of gas would be indicates that hydraulic fracturing was expected above a gas field such as Pavillion, data performed at 1039 m (3409 ft) below ground suggest that enhanced migration of gas has occurred surface on 11/9/2004 prior to cement squeeze to ground water at depths used for domestic water jobs at 823 m (2700 ft) and 256 m (840 ft) supply and to domestic wells. Lines of reasoning to below ground surface in April 2005. At Tribal support this explanation consist of following. Pavillion 41-10 a cement bond/variable density log indicates sporadic bonding directly 1. Hydrocarbon and isotopic composition of gas above the interval of hydraulic fracturing at The similarity of δ13C values for methane, 493 m (1618 ft) below ground surface. A ethane, propane, isobutane, and butane cement bond/variable density log conducted37

53. DRAFT between gas production and monitoring wells An alternative explanation of high methane and plots of δ13C-CH4 versus δD -CH4 and δ13C- concentrations in this area is that it is close to CH4 versus methane/(ethane + propane) the top of the dome comprising the Pavillion indicate that light hydrocarbons in casing and gas field which may facilitate natural gas dissolved gas in deep monitoring wells are migration toward the surface. However, this similar to produced gas and have undergone geologic feature would also facilitate little oxidation or biodegradation indicative of enhanced gas migration. Also, a mud-gas log advective transport. The absence of ethane conducted on 11/16/1980 (prior to intensive and propane in three of four domestic wells gas production well installation) at Tribal having sufficient methane to allow isotopic Pavillion 14-2 located only 300 m from the analysis and a shift of δ13C-CH4 and δD-CH4 location of the uncontrolled release does not values in a positive direction relative to indicate a gas show (distinctive peaks on a gas produced gas suggests the presence of gas of chromatograph) within 300 m of the surface. thermogenic origin in domestic wells 4. Shallow surface casing and lack of cement or undergoing biodegradation. This observation is consistent with a pattern of dispersion and sporadic bonding outside production casing degradation with upward migration observed With the exception of two production wells, for organic compounds. surface casing of gas production wells do not extend below the maximum depth of2. Elevation of dissolved methane concentrations in proximity to production domestic wells in the area of investigation. Shallow surface casing combined with lack of wells cement or sporadic bonding of cement Levels of dissolved methane in domestic wells outside production casing would facilitate generally increase in those wells in proximity migration of gas toward domestic wells. to gas production wells. With the exception of 2 domestic wells where methane was The discussion on migration of fluids associated with hydraulic fracturing is relevant detected at less than 22 μg/L, methane was not detected in domestic wells with 2 or less for gas migration and is not repeated here for brevity. Of particular concern are wellbores production wells within 600 m. having no or little cement over large vertical3. Spatial anomaly near PGDW05 instances. For instance, at Pavillion Fee 34- 03R, the cement bond/variable density log Methane concentrations in ground water indicates no cement until 671 m (2200 ft) appear highest in the area encompassing below ground surface. At Pavillion Fee 34-03B, MW01, PGDW30, and PGDW05. Ground a cement bond/variable density log conducted water is saturated with methane at MW01 on 10/22/2004 indicated no cement until 838 which is screened at a depth (239 m bgs) m (2750 ft) below ground surface. Migration typical of deeper domestic wells in the area. of gas via wellbores having no cement or poor Methane was detected in PGDW30 at 808 cement bonding outside production casing is μg/L at a depth of only 80 m, the highest level well documented in the literature. in any domestic well. A blowout occurred during drilling at a depth of only 159 m bgs in An alternative explanation of wellbore gas December 2005 adjacent to PGDW05. migration provided to EPA and considered is that domestic wells are poorly sealed and thus 38

54. DRAFT constitute a potential gas migration pathway. Hydraulic fracturing for coal-bed methane recovery is However, lack of cement and sporadic often shallow and occurs directly into USDWs (EPA bonding outside casing in production 2004). TDS less than 10,000 mg/L in produced water constitutes a major potential gas migration is common throughout the Rocky Mountain portion of pathway to the depth of deep monitoring and the United States (USGS 2011; Dahm et al. 2011). domestic wells. It is possible that domestic Ground water contamination with constituents such wells could subsequently facilitate gas as those found at Pavillion is typically infeasible or too migration toward the surface. expensive to remediate or restore (GAO 1989). Collection of baseline data prior to hydraulic fracturing5. Citizens complaints is necessary to reduce investigative costs and to verify Finally, citizens complaints of taste and odor or refute impacts to ground water. problems concurrent or after hydraulic Finally, this investigation supports recommendations fracturing are internally consistent. Citizens made by the U.S. Department of Energy Panel (DOE complaints often serve as the first indication 2011a, b) on the need for collection of baseline data, of subsurface contamination and cannot be greater transparency on chemical composition of dismissed without further detailed evaluation, hydraulic fracturing fluids, and greater emphasis on particularly in the absence of routine ground well construction and integrity requirements and water monitoring prior to and during gas testing. As stated by the panel, implementation of production. these recommendations would decrease the An alternate explanation provided and likelihood of impact to ground water and increase considered by EPA is that other residents in public confidence in the technology. the Pavillion area have always had gas in their wells. Unfortunately, no baseline data exists to verify past levels of gas flux to the surface or domestic wells.A lines of reasoning approach utilized at this site bestsupports an explanation that inorganic and organicconstituents associated with hydraulic fracturing havecontaminated ground water at and below the depthused for domestic water supply. However, furtherinvestigation would be needed to determine if organiccompounds associated with hydraulic fracturing havemigrated to domestic wells in the area ofinvestigation. A lines of evidence approach alsoindicates that gas production activities have likelyenhanced gas migration at and below depths used fordomestic water supply and to domestic wells in thearea of investigation.Hydraulic fracturing in the Pavillion gas field occurredinto zones of producible gas located within anUnderground Source of Drinking Water (USDW).39

55. DRAFT Bybee, K. (2007). Cement-bond-log interpretation5.0 reliability. Journal of Petroleum Technology, Feb. 2007, p. 64-66.References Caldwell, M.E. and Suflita, J.M. (2000). Detection of phenol and benzoate as intermediates of anaerobic benzene biodegradation under different terminalAmerican Public Health Association; American Water electron-accepting conditions. Environmental ScienceWorks Association; Water Environment Federation and Technology, v. 34, p. 1216-1220.(1998a). Method 3500-Fe B. Phenanthroline Method.In: Standard Methods for the Examination of Water Corseuil, H.X., Monier, A.L., Fernandes, M., Schneider,and Waste Water, 20th Edition. Editors Clesceri, L.S., M.R., Nunes, C.C. Rosario, M., and Alvarez, P.J.J.Greenberg, A.E., and Eaton, A.D. Washington D.C. (2011). BTEX plume dynamics following an ethanol blend release: Geochemical footprint andAmerican Public Health Association; American Water thermodynamic constraints on natural attenuation.Works Association; Water Environment Federation Environmental Science and Technology, v. 45, p. 3422-(1998b). Method 4500-S2- D. Methylene Blue Method. 3429.In: Standard Methods for the Examination of Waterand Waste Water, 20th Edition. Editors Clesceri, L.S., Craig, H. (1961). Isotopic variations in meteoricGreenberg, A.E., and Eaton, A.D. Washington D.C. waters. Science, v. 133, p. 1702-1703.Bair, S.E., Freeman, D.C., and Senko, J.M. (2010). Daddow, R.L. (1996). Water resources of the WindExpert Panel Technical Report Subsurface Gas Invasion River Indian Reservation, Wyoming. U.S. GeologicalBainbridge Township, Geauga County, Ohio, Survey Water-Resources Investigation Report 95-submitted to Ohio Department of Natural Resources 4223, 121 p.Division of Mineral Resources Management, June2010. Dahm, K.G., Guerra, K.L., Xu, P., and Drewes, J.E. (2011). Composite geochemical database for coalbedBartos, T.T., Quinn, T.L., Hallberg, L.L., and Eddy- methane produces water quality in the RockyMiller, C.A. (2008). Quality of shallow ground water in Mountain Region. Environmental Science andthree areas of unsewered low-density development in Technology, v. 45, p. 7655-7663.Wyoming and Montana, 2001. U. S. Geological SurveyScientific Investigations Report 2008-5012, 118 p. Delany J.M. and Lundeen S.R. (1990). The LLNL thermochemical database. Lawrence LivermoreBethke, C. M. (1996). Geochemical Reaction Modeling. National Laboratory Report, UCRL-21658. LawrenceOxford University Press, New York. Livermore National Laboratory. Dwyer, D.F. and Tiedje, J.M. (1983). Degradation ofBoyd, D., Al-Kubti, S., Khedr, O., N. Khan, and K. Al- ethylene glycol and polyethylene glycols byNayadi, prepared for the 2006 SPE Abu Dhabi methanogenic consortia. Applied and EnvironmentalInternational Petroleum Exhibition and Conference, Microbiology, v. 46, p. 185-190.Abu Dhabi, UAE, 5–8 November. Flores, R.M. and Keighin, C.W. (1993). ReservoirBradley, P.M., Chapelle, F.H., and Landmeyer, J.E. anisotropy and facies stratigraphic framework in the(2001). Effect of redox conditions on MTBE Paleocene Front Union Formation, western Wind Riverbiodegradation in surface water sediments. Basin, Wyoming, in W.R. Keefer, W.J. Metzger and L.H.Environmental Science and Technology, v. 35, p. 4643- Godwin, eds., Oil and Gas and Other Resources of the4647. Wind River Basin, Wyoming: Wyoming Geological Association Special Symposium, 1993, p. 121–141. 40

56. DRAFTGibb, J.P. and Jennings, K.V.B. (1987). Forum: How Kinnaman, F.S., Valentine, D.L., and Tyler, S.C. (2007).drilling fluids and grouting materials affect the Carbon and hydrogen isotope fractionation associatedintegrity of ground water samples from monitoring with the aerobic microbial oxidation of methane,wells. Ground Water Monitoring and Remediation, v. ethane, propane and butane. Geochimica et7, p. 33-42. Cosmochimica Acta, v. 71, p. 271-283.Harrison, S.S. (1983). Evaluating system for ground Mormille, M.R., Liu, S., and Suflita, J.M. (1994).water contamination hazards due to gas-well drilling Anaerobic biodegradation of gasoline oxygenates:on the glaciated Appalachian Plateau. Ground Water, Extrapolation of information to multiple sites andv. 21, 689-700. redox conditions. Environmental Science and Technology, v. 28, p. 1727-1732.Harrison, S.S. (1985). Contamination of aquifers byoverpressuring the annulus of oil and gas wells. Morris, D.A., Hackett, O.M., Vanlier, K.E., Moulder,Ground Water, v. 23, 317-324. E.A., and Durum, W.H. (1959). Ground water resources of Riverton irrigation project area,Hiatt, R., Clipsham, J., and Visser, T. (1964). The Wyoming, Geological Survey Water-Supply Paperinduced decomposition of tert-butyl hydroperoxide. 1375, 205 p.Canadian Journal of Chemistry, v. 42, p. 2754-2757. Mueller, C. (1989). Pavillion; Wyoming GeologicalJohnson, R.C. and Rice, D.D. (1993). Variations in Association, Wyoming Oil and Gas Fields Symposium,composition and origins of gases from coal bed and Bighorn and Wind River Basins, p. 356-358.conventional reservoirs, Wind River Basin, Wyoming,in W.R. Keefer, W.J. Metzger and L.H. Godwin, eds., Nordbotten, J.M., Celia, M.A., and Bachu, S. (2004).Oil and Gas and Other Resources of the Wind River Analytical solutions for leakage rates throughBasin, Wyoming: Wyoming Geological Association abandoned wells. Water Resources Research, v. 40,Special Symposium, 1993, p. 319-335. W04204.Johnson, R.C. and Keighin, W.C. (1998). Origins of Nordbotten, J.M., Celia, M.A. and Bachu, S. (2005a).natural gases from upper Cretaceous reservoirs, Injection and storage of CO2 in deep saline aquifers:Bighorn Basin, Wyoming and Montana, and Analytical solution for CO2 plume evolution duringcomparison with gases from the Wind River Basin, injection. Transport Porous Media, v. 58, p. 339-360.Wyoming, In Forty-Ninth Guidebook, WyomingGeological Association, p. 223- 249. Nordbotten, J.M., Celia, M.A., Bachu, S., and Dahle, H.K. (2005b). Semianalytical solution for CO2Keefer, W.R. and Johnson R.C. (1993). Stratigraphy leakage through an abandoned well. Environmentaland oil and gas resources in uppermost Cretaceous Science and Technology, v. 39, p. 602-611.and Paleocene rocks, Wind River Reservation,Wyoming, in W.R. Keefer, W.J. Metzger and L.H. Ohio Department of Natural Resources, Division ofGodwin, eds., Oil and Gas and Other Resources of the Mineral Resources Management, Report on theWind River Basin, Wyoming: Wyoming Geological Investigation of the Natural Gas Invasion of Aquifers inAssociation Special Symposium, 1993, p. 71-86. Bainbridge Township of Geauga County, Ohio September 1, (2008).Johnson, R.C., Finn, T.M., Kirschbaum, M.A., Roberts,S.B., Roberts, L.N.R., Cook, T., and Taylor, D.J. (2007). Osborn S.G., Vengosh, A., Warner, N.R., and Jackson,The Cretaceous-Lower Tertiary Composite Total R.B. (2011). Methane contamination of drinking waterPetroleum System, Wind River Basin, Wyoming, accompanying gas-well drilling and hydraulicChapter 4 of Petroleum Systems and Geologic fracturing. Proceedings of the National Academy ofAssessment of Oil and Gas in the Wind River Basin Sciences, v. 108, p. 8172-8176.Province, Wyoming. U.S. Geological Survey DigitalData Series DDS–69–J.41

57. DRAFTOsiensky, J.L., Winter, G.V., and Williams, R.E. (1984). http://www.epa.gov/osw/nonhaz/industrial/special/oiMonitoring and mathematical modeling of l/rtc1987.pdfcontaminated ground water plumes in fluvialenvironments. Ground Water, v. 22, p. 298-306. U.S. Environmental Protection Agency - Region VIII, Ground Water Section Guidance No. 34 - March 31,Pennsylvania Department of Environmental Protection 1994.(PADEP) Consent Order and Agreement Cabot Oil andGas Corporation Dimock and Springville Townships U.S. Environmental Protection Agency (2004).Susquehanna County (2009). Evaluation of Impacts to Underground Sources of Drinking Water by Hydraulic Fracturing of CoalbedPennsylvania Department of Environmental Protection Methane Reservoirs, Office of Water Office of Ground(PADEP) Modification to Consent Order and Water and Drinking Water (4606M), EPA 816-R-04-Agreement dated November 4, 2009 - April 15, 2010. 003, June 2004.Pennsylvania Department of Environmental Protection U.S. Environmental Protection Agency (2009). Site(PADEP) Modification to Consent Order and Inspection - Analytical Results Report, Pavillion AreaAgreement dated November 4, 2009 - April 15, 2010. Ground Water Investigation Site, August 8, 2009. available atPlafcan, M., Eddy-Miller, C.A., Ritz, G.F., and Holland, http://www.epa.gov/region8/superfund/wy/pavillion/J.P.R. (1995). Water resources of Fremont County,Wyoming. U.S. Geological Survey, Water-Resources U.S. Environmental Protection Agency (2010).Investigations Report 95-4095, 133 p. Expanded Site Inspection - Analytical Results Report, Pavillion Area Ground Water Investigation Site, AugustSchoell, M. (1980). The hydrogen and carbon isotopic 30, 2010. available atcomposition of methane from natural gases of various http://www.epa.gov/region8/superfund/wy/pavillion/origins. Geochimica et Cosmochimica Acta, v. 44, p.649-661. U.S. General Accounting Office (1989). Safeguards are not Preventing Contamination from Injected Oil andU.S. Department of Energy. Secretary of Energy Gas Wastes, GAO/RCED-89-97, July 1989.Advisory Board. Shale Gas Production Subcommittee.Ninety-Day Report – August 11, 2011. U.S. Geological Survey Produced Waters Database, Chemistry of Produced Water in the United States,U.S. Department of Energy. Secretary of Energy http://energy.cr.usgs.gov/prov/prodwat/tds.htm. SiteAdvisory Board. Shale Gas Production Subcommittee. accessed November, 2011.Second Ninety-Day Report – November 18, 2011. Valentine, D.L. et al. (2010). Propane respiration jump-Single, E.L. (1969). in Wyoming Geological Association starts microbial response to a deep oil spill. Science, v.21st Field Conference Guidebook, p. 101-103. 330, p. 208-211.Taylor, S.W., Sherwood Lollar, B., and Wassenaar, L.I. Van Stempvoort, D., Maathuis, H., Jaworski, E., Mayer,(2000). Bacteriogenic ethane in near-surface aquifers: B., and Rich, K. (2005). Oxidation of fugitive methaneImplications for leaking hydrocarbon well bores. in ground water linked to bacterial sulfate reduction,Environmental Science and Technology, v. 34, p. 4727- Ground Water, v. 43, p. 187-199.4732. Whiticar, M.J. (1999). Carbon and hydrogen isotopeU.S. Environmental Protection Agency (1987). systematic of bacterial formation and oxidation ofManagement of Wastes from Exploration, methane. Chemical Geology, v. 161, p. 291–314.Development, and Production of Crude Oil, NaturalGas, and Geothermal energy, EPA/530-SW-88-003,December, 1987. Available at 42

58. DRAFTWhiticar, M.J. and Faber, E. (1986). Methane oxidationin sediment and water column environments-isotopicevidence. Organic Geochemistry, v. 10, p. 759-768.Wilson, J.T., Kaiser, P.M., and Adair, C. (2005).Monitored natural attenuation of MTBE as a riskmanagement option at leaking underground storagetank sites. EPA/600/R-04/1790.Wyoming State Water Plan, Wyoming WaterDevelopment Office, Wind/Bighorn River Basin Plan,Jan. 14, 2003.http://waterplan.state.wy.us/plan/bighorn/techmemos/grnddet.html43

59. DRAFT Appendix ASummary of Analytical Results A1

60. DRAFT A2

61. DRAFT A3

62. DRAFT A4

63. DRAFTLaboratories, Analytes, and MethodsA - ALS Laboratory Group, Salt Lake City, UT. VOCs, SVOCs, pesticides, TCBs, TICs determined using methods specified under the CLP.A4 - A4 Scientific, The Woodlands, TX. TAL metals determined using methods specified under the CLP.E1 - Energy Laboratories Inc., Billings, MT. Heterotrophic plate counts, iron reducing bacteria, sulfur reducing bacteria.E2 - Energy Laboratories Inc., Billings, MT. GRO, DRO, THE, and TPH.I1 - Isotech Laboratories, Champaign, IL under contract by EnCana. Fixed gases and light hydrocarbons determined using ASTM D1945-03 in gas samplesand headspace of aqueous samples. δ13C and δD for C1 determined using gas stripping and IRMS in aqueous samples. δ13C and δD for C1-C4 determinedusing IRMS for gas samples.I2 - Isotech Laboratories, Champaign, IL. Fixed gases and light hydrocarbons determined using ASTM D1945-03 in headspace of aqueous samples. δ13C andδD for C1 and δ13C for C2 and C3 determined using gas stripping and IRMS in aqueous samples. δ13C DIC using gas stripping and IRMS.I3 - Isotech Laboratories, Champaign, IL. Fixed gases and light hydrocarbons determined using ASTM D1945-03 in headspace of aqueous samples. δ13C andδD for C1 , δ13C for C2 - C5, and δ13C for DIC gas stripping and IRMS in aqueous samples.I4 - Isotech Laboratories, Champaign, IL. Fixed gases and light hydrocarbons determined using ASTM D1945-03 in gas samples. δ13C and δD for C1 - C3 usingIRMS in gas samples.I5 - Isotech Laboratories, Champaign, IL. Fixed gases and light hydrocarbons determined using ASTM D1945-03 in gas samples. δ13C and δD for C1 - C3 usingIRMS in gas samples. 14C using AMS in gas samples.K - KAP Laboratories, Vancouver, WA. TAL metals determined under the CLP.L - Liberty Analytical, Salt Lake City, UT. VOCs, SVOCs, PCBs, and TICs determined under the CLP.O1 - EPA, ORD, Ada, OK. SO4, Cl, F, and Br determined using RSKSOP 276v3 and EPA Method 6500. NO3 + NO2 and NH4 determined using RSKSOP 214v5and EPA Method 350.1 and 353.2O2 - EPA, ORD, Ada, OK. DIC and DOC determined using RSKSOP-330v0 and EPA Method 9060A.O3 - EPA, ORD, Ada, OK. C1 determined using RSKSOP 175v5 and Cali-5 gas sampling bags.R3 - U.S. EPA Region 3 Laboratory, Fort Mead, MD. Diethylene glycol, triethylene glycol, tetraethylene glycol, and 2-butoxyethanol analysis by LC/MS/MS.This method is under development with no finalized SOP. EPA Methods 8000C and 8321 were followed for method development and QA/QC limits whereapplicable.R81 - U.S. EPA Region 8 Laboratory, Golden, CO (fluoride, chloride, nitrite-N, nitrate-N, orthophosphate-P, and sulfate determined using EPA Method 300.0and EPA Region SOP 310. Alkalinity determined using EPA Method 310.0).R82 - U.S. EPA Region 8 Laboratory, Golden, CO. VOCs determined using EPA Method 8260B.R83 - U.S. EPA Region 8 Laboratory, Golden, CO. SVOCs determined using ORGM-515 r1.1 and EPA Method 8270D.R84 - U.S. EPA Region 8 Laboratory, Golden, CO. GRO determined using ORGM-506 r1.0 and EPA Method 8015D. DRO determined using ORGM-508 r1.0and EPA Method 8015D.R85 - U.S. EPA Region 8 Laboratory, Golden, CO. Dissolved C1 in Phase I and dissolved C1-C3 in Phase II using EPA Method 524.2.S1 - Shaw Inc, Ada, OK in Phases III and IV. Metals and metals speciation determined using RSKSOP 213v4 and 257v2, or 332V0 and EPA Methods 200.7 and6020.S2 - Shaw Inc, Ada, OK in Phases III and IV. Aromatics and chlorinated hydrocarbons determined using method RSKSOP-259v1 and EPA Method 5021A plus8260C.S3 - Shaw Inc, Ada, OK . Alcohols, aromatics, and chlorinated hydrocarbons determined using method RSKSOP-259v1.S4 - Shaw Inc, Ada, OK. Low molecular weight acids determined using RSKSOP-112v6.S5 - Shaw Inc, Ada, OK. Dissolved gases C1-C4 determined using RSKSOP 194v4 and 175v5.S6 - Shaw Inc, Ada, OK. Hydrogen and oxygen isotope ratios of water determined using RSKSOP-296v0. A5

64. DRAFTAbbreviationsI () - Phase I(laboratory/method). Samples collected March, 2009 VOCs - volatile organic compoundsII() - Phase II(laboratory/method). Samples collected January, 2010 SVOCs - semivolatile organic compoundsIII() - Phase III(laboratory/method). Samples collected September and October 2010 PCBs - polychlorinated biphenylsIV() - Phase IV(laboratory/method). Samples collected April 2011. TICs - tentatively identified compoundsPG - gas production well DRO - diesel range organicsMW - deep monitoring wells GRO - gasoline range organicsPGM - shallow monitoring wells near pits TEH - total extractable hydrocarbonsPGS - soil samples near pits TPH - total purgeable hydrocarbonsDW - domestic wells DIC - dissolved inorganic carbonPGP - municipal wells in the Town of Pavillion TAL - target analyte listIRMS - isotope-ratio mass spectrometry CLP - U.S. EPA Contract Laboratory ProgramAMS - accelerated mass spectrometryC1 (methane), C2 (ethane), C3 (propane), iC4 (isobutane), nC4 (normal butane), iC5 (isopentane), nC5 (normal pentane), C6+ (hexanes + other lighthydrocarbons)Analytical MethodsORGM-506 r1.0 - Region 8 Standard Operating Procedure.ORGM-508 r1.0 - Region 8 Standard Operating Procedure.ORGM-515 r1.1 - Region 8 Standard Operating Procedure.RSKSOP-112v6 – Standard Operating Procedure for Quantitative Analysis of Low Molecular Weight Acids in Aqueous Samples by HPLC, 22 p.RSKSOP-175v5 - Sample Preparation and Calculations for Dissolved Gas Analysis in Water Samples Using a GC Headspace Equilibration Technique, 16 p.RSKSOP-194v4 - Gas Analysis by Micro Gas Chromatographs (Agilent MIcro 3000), 13 p.RSKSOP-213v4 - Standard operating procedure for operation of Perkin Elmer Optima 3300 DV ICP-OES, 21 p.RSKSOP-214v5 - Quality control procedures for general parameters analysis using Lachat Flow Injection analysis (FIA), 10 p.RSKSOP-259v1 - Determination of volatile organic compounds (fuel oxygenates, aromatic and chlorinated hydrocarbons) in water using automatedheadspace gas chromatography/mass spectrometry TEKMAR 7000 HS-Varian 2100T GC/MS system-ION trap detector, 28 p.RSKSOP-257v2 - Standard operating procedure for elemental analysis by ICP-MS, 16 p.RSKSOP-299v1 – Determination of Volatile Organic Compounds (Fuel Oxygenates, Aromatic and Chlorinated Hydrocarbons) in Water Using AutomatedHeadspace Gas Chromatography/Mass Spectrometry (Agilent 6890/5973 Quadruple GC/MS System), 25 p.RSKSOP-276v3 - Determination of major anions in aqueous samples using capillary ion electrophoresis with indirect UV detection and Empower 2software, 11 p.RSKSOP-296v0 - Determination of hydrogen and oxygen isotope ratios in water samples using high temperature conversion elemental analyzer (TC/EA), acontinuous flow unit, and an isotope ratio mass spectrometer (IRMS), 8 p.RSKSOP-297v1 – Metals Speciation Determination by LC/ICP-MS, 21 p.RSKSOP-298v1 - Arsenic Speciation Determination by LC/ICP-MS with Anion Suppression and NaOH Mobile Phase, 21 p.RSKSOP-313v1 - Determination of R-123 using the H25-IR Infrared Refrigerant Gas Leak Detector, 12 p.RSKSOP-314v1 - Determination of Fixed Gases using the GEM2000 and GEM2000 Plus Gas Analyzers & Extraction Monitors, 13 p.RSKSOP-320v1 - Determination of Organic and Inorganic Vapors Using the TVA-1000B Toxic Vapor Analyzer, 18 p.RSKSOP-330v0 – Determination of Various Fractions of Carbon in Aqueous Samples Using the Shimadzu TOC-VCPH Analyzer, 16 p.U.S. EPA Method 200.7 - Determination of Metals and Trace Elements in Water and Wastes by Inductively Coupled Plasma-Atomic Spectrometry, Rev. 5,Jan 2001.U.S. EPA Method 300.0 - Determination of Inorganic Anions by Ion Chromatography, Rev. 2.1, Aug. 1993..U.S. EPA method 310.1 - Alkalinity (Titrimetric, pH 4.5), Rev. 1978.U.S. EPA Method 350.1 - Determination of Ammonia Nitrogen by Semi-Automated Colorimetry, Rev. 2, Aug. 1993.A6

65. DRAFTU.S. EPA Method 5021A - Volatile Organic Compounds in Various Sample Matrices Using Equilibrium Headspace Analysis, Rev. 1, June 2003.U.S. EPA Method 6020 - Inductively Coupled Plasma-Mass Spectrometry, Rev. 1, Feb. 2007.U.S. EPA Method 6500 - Dissolved Inorganic Anions in Aqueous Matrices by Capillary Electrophoresis, Rev. 0, Feb. 2007.U.S. EPA Method 8260C - Volatile Organic Compounds by Gas Chromatography/Mass Spectrometry (GC/MS), Rev. 3, Aug. 2006.U.S. EPA Method 8015B - Determination of Nonhalogenated Organics Using GC/FID, Rev. 2, Dec. 1996.U.S. EPA Method 8015D - Nonhalogenated Organics Using GC/FID, Rev. 4, May 2003.U.S. EPA Method 8270D - Determination of Semivolatile Organic Compounds by Gas Chromatography/Mass Spectrometry (GC/MS), Rev. 4, Feb. 2007.U.S. EPA Method 8000C - Determinative Chromatographic Separations, Rev. 3, Mar. 2003.U.S. EPA Method 8260C - Volatile Organic Compounds by Gas Chromatography/Mass Spectrometry (GC/MS), Rev. 3, Aug. 2006.U.S. EPA Method 8270D - Semivolatile Organic Compounds by Gas Chromatography/Mass Spectrometry (GC/MS), Rev. 4, Feb. 2007.U.S. EPA Method 9060A - Total Organic Carbon, Rev. 1, Nov. 2004. A7

66. DRAFTA8

67. DRAFT A9

68. DRAFTA10

69. DRAFT A11

70. DRAFTA12

71. DRAFT A13

72. DRAFT A14

73. DRAFT A15

74. DRAFT

75. DRAFT Appendix BQuality Assurance and Quality Control (QA/QC) for Analysis B1

76. DRAFTB2

77. DRAFT B3

78. DRAFTB4

79. DRAFT B5

80. DRAFTB6

81. DRAFT B7

82. DRAFT B8

83. DRAFT B9

84. DRAFTB10

85. DRAFT B11

86. DRAFTB12

87. DRAFT B13

88. DRAFT B14

89. DRAFT B15

90. DRAFTB16

91. DRAFT B17

92. DRAFT

93. DRAFT Appendix CPhotographic Log of Deep Monitoring Well Construction C1

94. DRAFT Figure C1. Photograph of drilling rig on platform with shakers for mud recirculation at MW02. Figure C2. Photograph of blowout prevention (BOP) for annular space at base of drilling rig platform at MW02. Figure C3. Photograph of blowout preventer for drillstem.C2

95. DRAFTFigure C4. Photograph of bit and drillstem with bit for mud rotary drilling at MW02. C3

96. DRAFT Figure C5. Photograph of water truck used to transport water to mix mud. Figure C7. Photograph of mud additives EZ Mud Gold (Halliburton) and Dense Soda Ash. Figure C8. Photograph of mud additive Penetrol (Halliburton). Figure C6. Photograph of Quik-Gel bentonite (Halliburton) used to create mud for drilling.C4

97. DRAFT Figure C9. Photograph of flow of mud and cuttings from borehole at MW02.Figure C10. Photograph of monitoring of mud and cuttings using a Thermo ScientificTVA-1000B FID/PID at MW02. C5

98. DRAFT Figure C11. Photograph of pump used to transport mud and cuttings to shakers at MW02. Figure C12. Photograph of flow of mud and cuttings to shakers at MW02.C6

99. DRAFTFigure C13. Photograph of shakers separating mud from cuttings at MW02. C7

100. DRAFT Figure C14. Photograph of cuttings transported to disposal bins at MW02.C8

101. DRAFTFigure C15. Photograph of pumping of mud back to borehole at MW02. C9

102. DRAFT Figure C16. Photograph of injection of mud to borehole at MW02.C10

103. DRAFT Figure C17. Photograph of collection of cuttings for lithologic characterization at MW02. Figure C18. Photograph of removal of mud from Figure C19. Photograph of white coarse‐grained sand cuttings at MW02. targeted by local well drillers and media in which screens are set in for both deep monitoring wells. C11

104. DRAFT Figure C20. Photograph of setting of stainless- steel pre-packed screen and sand basket into borehole at MW02.C12

105. DRAFT Figure C21. Photograph of securing sand basket and casing above screen. Figure C22. Photograph of placement of sand in sandbasket. C13

106. DRAFT Figure C23. Photograph of well development at MW02.C14

107. DRAFT Appendix DPhotographic Log of Ground Water Sampling D1

108. DRAFT Figure D1. Photograph of flow from submersible pump through flowmeter at MW02. Figure D2. Photograph of flow of water to purge water disposal tank at MW02.D2

109. DRAFTFigure D3. Photograph (close-up) of flow of water into purge water disposal tank at MW02. Figure D4. Photograph of water (foaming) flowing into YSI flow cell at MW02. D3

110. DRAFT Figure D5. Photograph of sampling at MW02. The sample train was split prior to entry into purge water disposal container. Figure D6. Photograph of field filtering samples for metals analysis at MW02.D4

111. DRAFT Figure D7. Photograph of sample collection at PGDW14.Figure D8. Photograph of cooler packed with samples for shipment. D5

112. DRAFT

113. DRAFT Appendix EExamples of Cement Bond/Variable Density Log Interpretation E1

114. DRAFT Figure E1. Example of CBL/VDL indicating "no cement" at Pavillion Fee 34-03B. The CBL/VDL indicates no cement 2750 feet below ground surface at the time of logging.E2

115. DRAFTFigure E2. Example of "sporadic bonding" at Pavillion Fee 41-10 from 1000 to 1640 ft bgs. Hydraulic fracturingoccurred at 1618 feet below ground surface. Arrow denotes interval of hydraulic fracturing. E3

116. DRAFT 1100’ 1200’ 1300’ 1400’ 1500’ Figure E3a. Example of "sporadic bonding" at Pavillion Fee 11-11B. Hydraulic fracturing occurred at 1516 feet below ground surface. Arrow denotes interval of hydraulic fracturing. Depths on CBL/VDL difficult to read and inserted on left margin.E4

117. DRAFT 2450’ 3200’Figure E3b. Example of "sporadic bonding" Pavillion Fee 11-11B between 2350-3200 feet below groundsuface. Hydraulic fracturing occurred at 3165 feet below ground surface. Arrow denotes interval of hydraulicfracturing. Depths on CBL/VDL difficult to read and inserted on left margin. E5

118. DRAFT Figure E4. Example of "Sporadic Bonding" at Tribal Pavillion 24-02. Hydraulic fracturing occurred at 1538 feet bgs. Arrow denotes interval of hydraulic fracturing.E6

119. DRAFTFigure E5. Example of "Good Bonding" (from surface casing at 645 ft bgs to 820 ft bgs) followed by "SporadicBonding" (from 820 ft bgs 1310 ft bgs) to "Good Bonding" at 1310 to target depth at Pavillion Fee 41-10B. E7

120. DRAFT

121. DRAFT PRESORTED STANDARD POSTAGE & FEES PAID EPA PERMIT NO. G-35Office of Research and Development (8101R)Washington, DC 20460Official BusinessPenalty for Private Use$300

SlideShare for iOS

Investigation of Ground Contamination near Pavillion, Wyoming

by Marcellus Drilling News

on Dec 09, 2011

Statistics

Views

Actions

1 Embed 1,068

Accessibility

Categories

Upload Details

Usage Rights

Report content

Investigation of Ground Contamination near Pavillion, Wyoming Document Transcript