Scholarly article on topic 'Carbon isotope characterization of powder river basin coal bed waters: Key to minimizing unnecessary water production and implications for exploration and production of biogenic gas'

Carbon isotope characterization of powder river basin coal bed waters: Key to minimizing unnecessary water production and implications for exploration and production of biogenic gas Academic research paper on "Earth and related environmental sciences"

CC BY
0
0
Share paper
Keywords
{"Coalbed methane" / "Carbon isotope" / "Powder River Basin" / "Biogenic gas" / "Produced water" / "Reservoir characterization"}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Scott A. Quillinan, Carol D. Frost

Abstract Compared to other natural waters, water associated with biogenic natural gas is enriched in 13-carbon. Shallow coal seams regularly contain abundant resources of biogenic gas; as such water associated with biogenic gas in these coal beds is isotopically distinct from other waters. The production of gas from coal beds requires the removal of large volumes of produced water. Thus a method of discerning coalbed reservoir water from other natural waters (surface and groundwater) is important to both the coalbed natural gas (CBNG) industry and associated environmental and regulatory agencies. Although isotopic tracers have been employed to identify coalbed natural gas produced waters, the isotopic variability within the reservoir has not been documented and explained. In this study, we present the isotopic compositions of dissolved inorganic carbon, oxygen and hydrogen for water produced from 197 CBNG wells in the Powder River Basin of Wyoming and Montana. This extensive database allows us to distinguish variations in isotopic compositions that may occur by multiple processes. These include variations that identify efficient dewatering of coal beds, variations characterizing incomplete hydraulic isolation of coal beds from adjacent strata and the subsequent mixing of groundwaters, variations related to well completion design, and variations associated with geochemical and biogenic processes that occur along groundwater flow paths. These data suggest that little change in δ13CDIC occurs within the reservoir as a result of water and gas production; thus, the carbon isotopic composition informs other processes within the reservoir unrelated to coalbed natural gas recovery. The δ13CDIC and δD of groundwater vary along flow-path across the basin, reflecting different methanogenic pathways that are associated with different isotopic fractionations, and the pathways that dominate in different areas within the basin. In areas where several producing coal seams are present, the δ13CDIC and δD of produced waters from each seam are distinct. Therefore on a local scale, the isotopic composition of produced water can identify the particular coal seam from which water and gas are withdrawn. The methods and results presented in this case study provide examples that illustrate how water quality and isotopic data can be used to determine the hydraulic connectivity between coal and non-coal strata, identify and quantify water from individual coal horizons, as well as predict and understand the isotopic variability of the reservoir.

Academic research paper on topic "Carbon isotope characterization of powder river basin coal bed waters: Key to minimizing unnecessary water production and implications for exploration and production of biogenic gas"

ELSEVIER

Contents lists available at ScienceDirect

International Journal of Coal Geology

journal homepage: www.elsevier.com/locate/ijcoalgeo

Carbon isotope characterization of powder river basin coal bed waters: Key to minimizing unnecessary water production and implications for exploration and production of biogenic gas^

Scott A. Quillinan a,b,*I Carol D. Frosta

a Department of Geology and Geophysics, University of Wyoming Dept. 3006,1000 E. University Ave., Laramie, WY 82071, USA b Carbon Management Institute, School of Energy Resources, University of Wyoming, 2020 Grand Avenue Suit 500, Laramie, WY 82071, USA

ARTICLE INFO

ABSTRACT

Article history: Received 1 April 2013

Received in revised form 27 September 2013 Accepted 1 October 2013 Available online 17 October 2013

Keywords: Coalbed methane Carbon isotope Powder River Basin Biogenic gas Produced water Reservoir characterization

Compared to other natural waters, water associated with biogenic natural gas is enriched in 13-carbon. Shallow coal seams regularly contain abundant resources of biogenic gas; as such water associated with biogenic gas in these coal beds is isotopically distinct from other waters. The production of gas from coal beds requires the removal of large volumes of produced water. Thus a method of discerning coalbed reservoir water from other natural waters (surface and groundwater) is important to both the coalbed natural gas (CBNG) industry and associated environmental and regulatory agencies.

Although isotopic tracers have been employed to identify coalbed natural gas produced waters, the isotopic variability within the reservoir has not been documented and explained. In this study, we present the isotopic compositions of dissolved inorganic carbon, oxygen and hydrogen for water produced from 197 CBNG wells in the Powder River Basin of Wyoming and Montana. This extensive database allows us to distinguish variations in isotopic compositions that may occur by multiple processes. These include variations that identify efficient dewatering of coal beds, variations characterizing incomplete hydraulic isolation of coal beds from adjacent strata and the subsequent mixing of groundwaters, variations related to well completion design, and variations associated with geochemical and biogenic processes that occur along groundwater flow paths. These data suggest that little change in S13CDIC occurs within the reservoir as a result of water and gas production; thus, the carbon isotopic composition informs other processes within the reservoir unrelated to coalbed natural gas recovery. The S13CDIC and 8D of groundwater vary along flow-path across the basin, reflecting different methanogenic pathways that are associated with different isotopic fractionations, and the pathways that dominate in different areas within the basin. In areas where several producing coal seams are present, the S13CDIC and 8D of produced waters from each seam are distinct. Therefore on a local scale, the isotopic composition of produced water can identify the particular coal seam from which water and gas are withdrawn. The methods and results presented in this case study provide examples that illustrate how water quality and isotopic data can be used to determine the hydraulic connectivity between coal and non-coal strata, identify and quantify water from individual coal horizons, as well as predict and understand the isotopic variability of the reservoir.

© 2013 The Authors. Published by Elsevier B.V. All rights reserved. This is an open access article under the CC BY

license (http://creativecommons.org/licenses/by/3.0/).

1. Introduction

Strongly positive S13CDIC (> 10%») have been documented in produced water from several biogenic gas-producing fields, including the Antrim Shale, New Albany Shale, Forest City Basin shales and coals, Atlantic Rim coals, and the Powder River Basin coals (Martini et al., 1998,

☆ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

* Corresponding author at: Carbon Management Institute, School of Energy Resources, University of Wyoming, 2020 Grand Avenue Suit 500, Laramie, WY 82071, USA. Tel.: +1 307 766 6697.

E-mail address: scottyq@uwyo.edu (S.A Quillinan).

2003; Mcintosh et al., 2008; McLaughlin et al., 2011; Quillinan et al., 2012; Sharma and Frost, 2008). Studies have used this diagnostic characteristic of water associated with biogenic gas generation to trace produced water on the surface (Sharma and Frost, 2008) and in the subsurface (Martini et al., 1998; McLaughlin et al., 2011; Sharma and Frost, 2008). The basis for the tracer lies in the fact that most natural waters have negative ô13CDIC (—12 to — 7%»; Mook and Tan, 1991) while waters associated with biogenic gas are positive ( + 10 to + 30%»; Grossman et al., 1989). Enriched ô13CDIC are recorded in reducing environments where the production of biogenic methane, either by acetate fermentation or by CO2 reduction, removes 12C, leaving 13C-enriched CO2 in the system, which accounts for the positive ô13CDiC in formation waters (Grossman et al., 1989). intermediate

http://dx.doi.org/10.1016/j.coal.2013.10.006

0166-5162/© 2013 The Authors. Published by Elsevier B.V. All rights reserved. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/).

values may reflect mixing of these two types of natural waters (Martini et al., 1998; McLaughlin et al., 2011; Quillinan et al., 2012; Sharma and Frost, 2008). This study seeks to understand both the natural isotopic variations of water within a biogenic natural gas reservoir along with any that are induced as a result of gas production. We have selected the Powder River Basin (PRB) of Montana and Wyoming for our study, a mature and well-studied biogenic coalbed natural gas (CBNG) field, which contain thousands of producing CBNG wells (Fig. 1).

The PRB in northeastern Wyoming and southeastern Montana is an asymmetric syncline formed during the Laramide orogeny in Late Cretaceous to Early Eocene times. The coals in the PRB are found within the Tongue River member of the Paleocene Fort Union Formation and the overlying Eocene Wasatch Formation (Love and Christiansen, 1985). The Tongue River Member of the Fort Union Formation hosts the thick, laterally continuous coal beds that are mined on the eastern side of the basin and contain CBNG across the basin (WOGCC, 19872010). Jones (2008) has identified 10 distinct coal zones that contain 26 mappable coal beds within the Tongue River Member and in the Wasatch Formation (Fig. 2). The natural gas resource from these coal beds is estimated to hold more than 30 trillion cubic feet (TCF, 850 billionm3) of recoverable methane (Wyoming State Geological Survey, 2010) of which <5 TCF (141 billion m3) has been recovered (WOGCC, 1987-2010).

2. Background

PRB coals are primarily subbituminous. Because of their low rank, PRB coals have not generated large quantities of thermogenic gas (Montgomery, 1999). CBNG from the Fort Union Formation has C1 /C2 + C3 ratios from 1000 to 4000 with a few lower values in the deepest part of the basin (Flores et al., 2008), indicating that it is almost exclusively of biogenic origin.

Biogenic gas is commonly associated with shallow (less than 600 m), organic-rich rocks that are thermally immature (Shurr and Ridgley, 2002). Two metabolic pathways have been identified to produce biogenic gas: acetate fermentation and CO2 reduction (Jenden and Kaplan, 1986; Schoell, 1980; Whiticar et al., 1986; Woltemate et al, 1984). Gasses originating from the fermentation pathway are generally more enriched with respect to carbon-13 and more depleted in deuterium than gas generated via CO2 reduction. In CO2 reduction the carbon isotope ratio of the methane is controlled by the initial carbon isotope composition of the CO2 (Shurr and Ridgley, 2002).

CH3COOH^CH4 + CO2(Acetate fermentation) (1)

CO2 + 4H2^CH4 + 2H2O(CO2 reduction) (2)

CO2 is readily available in most coal beds as a result of carbonate dissolution and bacterial processes (Freeze and Cherry, 1979). Most bacteria generate methane through CO2 reduction (Rice, 1993), although several species utilize the fermentation pathway most commonly acting on an acetate substrate. The metabolic pathway also varies with time: biogenic gas that is generated early in the coal bed burial history likely results from the carbonate reduction pathway (Rice, 1992 and Rice and Claypool, 1981). Biogenic gas formed long after deposition can be a result of either acetate fermentation or by carbonate reduction (Barker and Fritz, 1981; Carothers and Kharaka, 1980; Coleman et al., 1988; Grossman et al., 1989; and Rice and Threlkheld, 1982) and is found in coal beds of all ranks (Rice, 1993).

Montgomery (1999) noted that Fort Union coal beds contain significant quantities of biogenic gas. On the basis of this observation Shurr and Ridgley (2002) suggest that the PRB coals have remained at depths suitable for biogenic gas production for extensive periods, allowing for both ancient and recent biogenic gas generation. Recent

bacterial gas generation has been documented on the basin margins, whereas early-stage (older) gas generation is theorized to have occurred across the basin (Flores et al., 2008; Shurr and Ridgley, 2002). Similar to the San Juan (U.S.), Surat (Aus.), and Sydney (Aus.) basins, recent biogenic gas on the margins of the PRB has been found to form by the acetate fermentation and CO2 reduction pathways (Faiz et al., 2003; Flores et al., 2008; Gorody, 1999; Moore, 2012; Rice, 1993 and Whiticar et al., 1986).

The shallow nature of PRB coals also leads to low sorbed gas content. Gas content from these coals ranges from 22 to 74 cubic feet per ton (0.6 to 2 m3/t). The chemical composition of the gas in Fort Union coal beds of the Powder River Basin is, on average, 86% methane, 6% carbon dioxide, and 4% nitrogen (Flores et al., 2008).

The methane produced from the Powder River Basin has a 813CCH4 of - 83.4%» to - 53.4%» and a SD CH4 of - 333%» to - 283.4%» (Flores et al., 2008; Gorody, 1999; Rice, 1993). Flores et al. (2008) observed depleted CH4 with respect to carbon on the basin margins and enriched gas in the center of the basin. They suggested that this is a result of methanogenesis by mixed acetate fermentation and CO2 reduction pathways on the basin margins and dominantly by CO2 reduction in the center of the basin.

Isotopes of hydrogen of the formation water also are fractionated during methanogenesis, creating isotopically light methane. During acetate fermentation, 25% of the hydrogen is supplied from the formation water as compared to 100% for CO2 reduction (Whiticar et al., 1986). The SD of formation water may be enriched by methanogenesis (Clark and Fritz, 1997; Quillinan, 2011; Siegel et al., 1990) to a greater degree by CO2 reduction than by acetate fermentation. However SD is not considered a reliable indicator of methanogenic pathways (Waldron etal., 1999).

The SD and S18O of precipitation lie along the global meteoric water line (SD = 8S18O + 10%»; from Craig, 1961). Produced water that originates as meteoric water should also lie approximately along this line. The exact position is a function of temperature (Craig, 1961).

Meteoric recharge for Fort Union coal beds in the PRB takes place on the basin margins (Bartos and Ogle, 2002 and references therein; Quillinan and Frost, 2012). These models suggest that residence times of water in coal bed aquifers increase with increasing depth and distance from the outcrop. On the basis of limited tritium data and simple ground water flow models, Pearson (2002) suggested that the residence time for water to flow from eastern recharge to the central part of the basin was between 7000 and 70,000 years. 14C dating by Frost and Brinck (2005) further refined this estimate to less than 20,000 years.

Water derived from coal beds in the Powder River Basin have S18O that ranges from - 21.5% to - 16.15% and SD ranging from -158.4% to - 121.3% (Bartos and Ogle, 2002; Flores et al., 2008; Gorody, 1999). The S13CDIC of waters measured from the Upper and Lower Wyodak coal zone range from 12% to 22% (Sharma and Frost, 2008). These waters are strongly sodium bicarbonate-type, with total dissolved solid (TDS) concentrations that range from 500 to 5600 mg/L (Bartos and Ogle, 2002; Campbell et al., 2008; Frost et al., 2002; Pearson, 2002; Quillinan, 2011; Rice et al., 2000). The TDS of these waters is found to increase along approximate groundwater flow paths from the southern and eastern margins toward the basin center. This geospatial trend is observed in all coal zones (Quillinan and Frost, 2012).

3. Methods

Samples were collected from the major exploited coal zones for CBNG development (Quillinan and Frost, 2012): the Wyodak Rider, Upper Wyodak, Lower Wyodak, Cook and Wall coal zones (Fig. 2). We analyzed 181 produced water samples from wells in Wyoming and 16 from wells in Montana for stable isotopic compositions of

Fig. 1. Sample location map of the Powder River Basin study area. Color of the dots marking the sample locations identify the coal zone in which the well is completed. Some well pads contain more than one well; symbols for these wells are enclosed by boxes.

Formation Coal zone Coal bed

Upper Wasatch U!m Lake DeSmet U Cross

Wasatch Felix Felix Rider Upper Felix Felix

Lower Wasatch Arvada Unnamed

Roland Upper Roland Roland of Baker Roland of Tafi

Wyodak Rider Smith Rider Smith/Big George Lower Smith

East West

Fort Union Upper Wyodak Anderson Rider (Anderson) Anderson (Wyodak) Lower Anderson Dietz 1 Dietz 2

Tongue River East West

member Lower Wyodak Canyon Rider Canyon Dietz 3

Cook Cook (Werner) Lower Cook (Gates)

Wall Wall Lower Wall Pawnee

Basal Tongue River Moyer

Fig. 2. Powder River Basin coal stratigraphy. From Jones (2008; per. comm., 2010).

carbon, oxygen, and hydrogen. Major ion compositions for these water samples are presented in Quillinan (2011).

The water samples were taken from the valve at the wellhead, and all wells were pumping at the time of sampling, obviating the need to remove several casing volumes of water from the wells prior to sample collection. Water samples were collected into 250 mL acid washed bottles. From this sample, 4 aliquots were prepared all of which were filtered using a Cameo 0.045 |jm nylon pre-filter that was attached to the tip of a syringe. All the samples were filled to the point of overspill to eliminate headspace. One 30 mL sample was prepared for carbon isotopic analysis as follows: water was drawn into a 60 cc pre-rinsed Luer-lock syringe and dispensed into a 30 mL Wheaton vial. Two to three drops of benzalkonium chloride was added to halt microbial activity. The Wheaton vial was capped with Teflon® septa and sealed with an aluminum cap using a crimper. One 10 mL sample was captured in a glass vial for deuterium and oxygen isotope analysis. Two samples were prepared for water chemistry analysis into 30 mL acid washed plastic bottles, one of which was acidified using 0.5 mL of HNO3 for cation analysis.

Carbon isotope ratios of dissolved inorganic carbon were analyzed on a GasBench-II device coupled to a Finnigan DELTA plus mass spectrometer in the Stable Isotope Facility at the University of Wyoming (UWSIF). The accuracy was monitored by replicate analysis of the samples and internal lab standards, and was found to be accurate to ±0.1%. The 613C values are reported in per mil relative to V-PDB standard (Fig. 3). The oxygen and hydrogen isotope analysis of the water samples were analyzed on a Liquid Water Isotope Analyzer (LWIA) developed by Los Gatos Research (model DLT-100) installed at UWSIF. The system uses off-axis integrated-cavity output spectroscopy (off-axis ICOS), which measures isotope ratios based on laser absorption. Ten replicate injections of ~ 1 |jL

were used to calculate the 618O and 62H composition of each water sample. Measurement precision was assessed by replicate analysis of samples and internal UWSIF lab standards and was better than ±0.4%% for S18O and ± 0.6% S2H. The S18O and S2H values are reported in per mil (%) relative to V-SMOW (Fig. 4). Cation concentrations were determined using ICP MS, anions by ion chromatography, and the alkalinity by potentiometric titration. Geochemical data are reported in Quillinan (2011).

4. Results

4.1. General characteristics ofPRB CBNG produced waters

The 618O of produced water from PRB coal beds ranges from — 26.3% to —16.9%. In general, the 618O and 6D are depleted with respect to PRB surface waters measured by Carter (2008). This suggests that CBNG waters were likely recharged during a colder climate than the present day (Fig. 4). On average the most depleted 618O are found in waters produced from the stratigraphically-lowest coal beds (Wall coal zone) in the northeast of the study area, while the most enriched values are found in water produced from coals near the southeast margin closest to outcrop. The most depleted 6D values, like the most depleted 618O values, are found in waters from the stratigraphically lowest coals in the northeast portion of the study. The large range measured in the northeast of the study area indicates that these coals are recharged under differing meteoric conditions. The fact that these samples are found within close proximity of each other (Fig. 1) may indicate the coal beds are compartmentalized in that area. The samples with the most enriched 6D ( —131.0%) are measured in the northern edge of the Big George coal bed (Wyodak Rider coal zone) near the center of the basin. Samples from this location do not have the most enriched 618O.

The S13CDIC of CBNG produced water is positive (typically +10 to + 25%o) where collected from single-completion wells irrespective of coal zone (Fig. 3). In deeper coal beds positive S13CDIC values decrease along groundwater flow paths. Variable, including negative, S13CDIC are measured in wells that are shallow (generally < 180 m), are completed in multiple coal beds, or have open-hole completions that extend beyond a single coal bed (Fig. 3; Table 1).

Coals from two geographic areas with relatively shallow wells appear to be characterized by depleted S13CDIC and 6D: coals from an area of Montana producing from the Dietz 1, 2 and 3, and coals in an area in the southern PRB where one shallow well is completed in the Anderson-Canyon (Fig. 3). It is possible that these samples have been affected by methane oxidation, which creates a positive shift in 6D and 613C of methane and leaves DIC depleted in the heavy isotope of carbon (Clark and Fritz, 1997). Isotopic analysis of methane would be required to further evaluate this hypothesis; since only 6 of 197 samples appear to be affected, we do not consider this process further in this paper.

The salinity of CBNG water is mainly contributed by dissolved sodium and bicarbonate ions. Calcium and magnesium concentrations increase in the shallow wells at low alkalinity until they reach saturation. Sulfate contents are typically very low and in only a few of the shallow wells are sulfate contents above detection limits.

42. General characteristics of surface water and waters from non-coal PRB aquifers

Three samples were measured that were not associated with coal beds, these were used to confirm that water not associated with methanogenesis is characterized by negative S13CDIC. These included one surface water sample from Crazy Woman Creek to supplement the 30-sample S13CDIC water data set from the Powder River collected by Carter (2008), and two water samples from wells completed in the Cretaceous Fox Hills and Fox Hills/Lyons Formations. The S13CDIC of these samples ranged from — 7.8 to — 9.1 (Fig. 4). The surface water

Wyodak Rider Upper Wyodak Lower Wyodak Cook Wall Non Coal bed

Big George Big George -, Lower Smith Anderson Anderson-Canyon1 Dietz 1,2,&3' Wyodak Big George-Canyon 1 Lower Canyon Werner/ Cook Wall-Lower Canyonä Wall Foibills Foxhills Lyons Crazy Wu nui il Creek

b g i ■ s b m d D Ol ÖD Di ODO d I DD D DU 0 1 0 000 ID I D DDnD h d : S B -

Fig. 3. 813Cdic are shown by coal bed and coal zone. Non-coal bed waters include a surface water sample and two samples from the Cretaceous Fox hills and Lyons formations. Note the consistent enrichment in all coal zones and the comparably low values in the multi-zone of the Wyodak split and Dietz 1,2, and 3.

sample is Ca-SO4 type. The Fox Hills and Lyons samples are Na-HCO3-type.

5. Discussion

The goals of this study include understanding both the natural stable isotopic variations within produced waters from coal beds along with those that are induced as a result of gas production. In order to evaluate

whether stable isotope compositions, particularly carbon isotopic ratios, can fingerprint specific coal beds, or to identify the extent to which the carbon isotopic composition of produced waters varies spatially, we first must determine whether there is any time-dependence of carbon isotopic compositions. Accordingly, we first evaluate whether there is any correlation between the length of time that a well has been producing water and gas and the ô13CDIC of the produced water.

5.1. Relationship of 8UC to years of gas production

— -160 a

-220 -28

• Surface water (Carter, 2008) 0 Wyodak Rider ■ Upper Wyodak • Lower Wyodak • Cook o Wall

oZ ■ /

-23 5isO (%„)

Fig. 4. 8D as a function of 818O PRB produced waters are depleted with respect to PRB surface waters from Carter (2008). This indicates that coal bed aquifers were recharged during cooler climatic conditions than present day.

Prior to this study, the number of S13CDIC analyses of coal bed produced water were too limited to evaluate the possibility that the carbon isotopic composition of water produced from a given well might change over time. However, there are processes by which a change might occur. For example, it might be the case that as water is withdrawn from a coal seam and the seam is depressurized, water from adjacent strata is drawn into the coal. If the non-coal strata above or below the coal have distinct S13CDIC, then the water withdrawn from the coalbed well may change with time. Non-coal aquifers typically have water with negative S13CDIC, so we might predict a decrease in S13CDIC of produced water with time. Two variables are considered to investigate temporal changes in S13CDIC, the first consideration is the length of time the well has been producing water and the second is the volume of water that has been produced from the well.

The samples analyzed in this study were collected from wells that were completed between 1994 and 2008. We plot the S13CDIC of produced water from 147 wells with single completions as a function of the year that the well was completed by coal seam (Fig. 5 a, b, and c). The S13CDIC of produced water from the Wyodak coal zone varies from 13.2 to 20.1%, with no time-dependent variation as a function of age of the well (Fig. 5a). This is despite a very large difference in the total

Table 1

Statistical summary of isotopic compositions measured in groundwater of Powder River Basin coal beds.

Coal zone

Coal bed sample subset

No. of samples Total depth Spud date (yr) range (ft)

Isotopic ratios

ö13C%o

Ö O%o

Min Mean Max Min Mean Max Min Mean Max

Wyodak Rider Big George 112 1000-2620 2000-2008 -8.7 13.5 24.5 -152.5 -137.8 -131.0 -20.9 -18.2 -16.9

Big George—Lower Smith 3 1256-1313 2004-2005 13.2 13.2 13.6 - 134.1 - 135.8 -137.0 - 18.4 - 18.1 -17.6

Upper Wyodak Anderson 5 685-900 2002-2003 12.9 21.0 24.1 - 145.5 - 142.2 - 134.2 - 19.2 - 18.8 -17.9

Wyodak 15 620-1222 1994,1997-2002 13.2 17.4 21.0 - 147.8 - 143.3 - 140.7 - 19.1 - 18.5 -17.4

Anderson-Canyona 15 408-652 2005-2006 - 24.7 0.7 8.4 - 144.3 -137.1 - 134.6 - 18.7 -17.7 -17.3

Deitz 1 13 191-552 - -23.4 5.2 14.5 - 164.1 - 142.5 - 134.2 - 20.4 - 18.1 -17.0

Deitz 1,2, and 3a 3 346-508 - - 24.1 -15.2 -8.7 -169.5 - 169.5 -162.8 - 21.8 - 20.0 - 20.4

Lower Wyodak Big George-Canyona 1 1166 2003 - 22.2 - - - 150.7 -150.7 - - 19.0 -

Lower Canyon 1 782 2001 - 21.8 - - -151.0 -151.0 - - 18.5 -

Lower Canyon-Wall 2 10,341,085 2003 21.3 21.3 21.3 -186.9 - 186.8 -186.7 - 23.2 -23.2 -23.1

Cook Cook (Werner) 5 965-1845 2002, 2005-2006 13.2 16.8 20.9 - 145.4 - 140.7 -136.4 - 19.7 - 19.0 - 18.2

Wall Wall 13 960-2290 2001-2005 10.5 16.8 22.3 -208.5 - 162.1 - 141.7 - 26.3 - 20.7 -17.3

Identifies wells with multiple horizon production.

amount of water withdrawn from the oldest wells compared to the younger ones: The oldest well, drilled in 1994, has produced nearly 800,000 barrels (1.2 x 108 L) of water compared to the wells drilled in 2002 that have produced on average 275,000 barrels (4.3 x 107 L) of water per well (Fig. 5b). We note that the 1994-Wyodak well has not produced gas since 2002 (WOGCC, 1987-2010). The fact that the 613CDIC of the produced water is still comparable to that of water from other Wyodak wells suggests that even though the gas from the seam

has been exhausted, the water is still originating from the coal aquifer. This suggests that the Wyodak coal bed is hydraulically isolated from other strata.

Although the produced water samples from the Big George exhibit a larger variation in S13CDIC (13.0 to 24.5%) than do the samples from the Wyodak coal zone, they likewise exhibit no temporal variation in carbon isotopic composition (Fig. 5b). The oldest wells, completed in 2000 and which have produced an average of nearly 2.3 million gal (3.6x107L) of

Fig. 5. The S13CDic for the Big George, Wyodak and Wall coal beds sorted by the first year of reported production. Even wells that have been producing for several years have comparable 8,3Cdic to new wells.

water are isotopically indistinguishable from the wells completed in 2008 from which only 8700 gal (3.2 x 104 L) were withdrawn before sampling (WOGCC, 1987-2010). The isotopic variations are most likely related to the large differences in well depth in the Big George sample set, and are discussed below. Alternatively, variable S13CDIC could suggest local lack of confinement within the reservoir.

Cumulative water production from the Wall coal zone wells analyzed in this study varies from 400,000 barrels (6.4 x 108 L) for the oldest wells to 44,000 (7.0 x 106 L) for the youngest ones. Wells completed between 2001 and 2004 yielded produced water samples that showed no trend in S13CDIC with time; these ranged from 10.5% to 22.3%. The youngest well, completed in 2005, has slightly more enriched S13CDIC, with S13CDIC of 22.3% (Fig. 5c). This could indicate a lesser degree of hydraulic isolation within the Wall coal zone. However this well is located at some distance from the others and its more enriched S13CDIC could be a result of vertical or lateral variation.

Ideally, the time-dependence of S13CDIC of produced water would be evaluated by repeated sampling of a single well over a period of many years. Our approach of sampling multiple wells that were drilled over a period of years is a less ideal strategy, but nevertheless the data reveal no time-dependence of S13CDIC of produced water. Therefore, we can address other factors that account for the observed variations in subsequent sections.

5.2. 813CDic as a unique identifier of coal beds

Much time and effort has been invested to accurately correlate and map the subsurface extent of coals and coal zones in the Powder River Basin (e.g. Ayers and Kaiser, 1984; Flores, 1999; Flores and Bader, 1999; Flores et al., 1999; Goolsby and Finley, 2000; Jones, 2008, 2010; McClurg, 1988; Pocknall and Flores, 1987; Seeland, 1993; Sholes and Cole, 1981). Resource assessments for both CBNG and minable coal are based on these subsurface correlations. Although PRB coal zones are thick and laterally continuous, accurate correlations can be problematic because PRB coal beds within these zones can pinch-out, split and merge between wells (Flores, 1999). Campbell et al. (2008) identified regional coal bed-specific variations in strontium isotopes and Sharma and Frost (2008) suggest that S13CDIC can be used to fingerprint individual coal beds. Our data set confirms that locally, individual coal beds have distinct S13CDIC, which can be used to identify and correlate specific coal beds.

5.2.1. Vertically stacked coal beds

The isotopic compositions of waters produced from different coal beds were investigated by sampling twin and triplet wells; that is, multiple wells that share the same well pad, but are producing from different coal bed intervals. Waters from these wells were measured to determine isotopic variations between coal beds from the same location. Data were obtained from three sample sites in the PRB. Fig. 6 shows the isotopic variation of the stacked coal beds from these three different locations. In each case, the S13CDIC of waters from different coal beds from the same well pad are isotopically distinct. On log I, for produced water samples collected from coals at increasing depth, the S13CDIC value of the Anderson coal is 22.2%, the Werner is 20.9%, and the Wall is 17.2%. Log II and III also show similar trends of distinct S13CDIC for each coal in a vertical section, and a decrease in S13CDIC with increasing depth (Fig. 6).

5.2.2. Field scale S13CDiC variability

Basin wide, S13CDIC of produced water from single coal beds varies by as much as 17% (Table 1). Table 2 gives the average S13CDIC and calculated standard deviation within the Big George coal bed sorted by location, which makes it possible to look at S13CDIC variations in the Big George from gas field to gas field. CBNG fields in the PRB generally are 40 to 100 km2 in size. Within a given field, the average S13CDIC is

relatively consistent: all of the standard deviations are less than 2.5, and generally are less than 1.5 (Table 2).

Deviations from the characteristic isotopic compositions of a coal bed within a given field can be instructive. As an illustration, we consider five Anderson coal bed wells located in close proximity to each other. Four of the five wells have a S13CDIC that are comparable, ranging only from 22.2 to 24.1%. The fifth well has a lower S13CDIC of 14.0%. The well with the lower S13CDIC also yields a TDS value that is approximately 600mg/L. This is 30% lower than the TDS concentrations of the surrounding wells. The lower TDS and S13CDIC both are consistent with the dilution of coal bed water with non-coal bed water, suggesting that the coal tapped by this well is not as well-confined as coals tapped by the four other wells. These results conform to the observations of McLaughlin et al. (2011), who showed in their study of the Atlantic Rim CBNG production area that lower S13CDIC of produced water may identify an unconfined reservoir. McLaughlin et al. (2011) also showed that the degree of hydraulic confinement of a coal bed correlates with water to gas ratios produced from a well, where unconfined coal beds require greater production of water per unit of gas.

Overall, these data show that stacked coals in the PRB are locally isotopically distinct, and that there is little isotopic variation amongst single coal beds on a field scale. Thus, S13CDIC of a reservoir can be used locally to identify the coal bed(s) from which the produced water is withdrawn. Moreover, that S13CDIC can be used to identify wells that may be unconfined.

5.3. 813CDiC identification of water sources in wells with multiple-seam completions

PRB CBNG wells often target more than one coal seam from a single well (so-called multiple-seam or commingled completions). To produce gas effectively, multi-seam wells must dewater multiple coal beds. In multi-seam wells, it is difficult to ascertain from which coal bed(s) the water is being withdrawn, and therefore whether all targeted coal beds are being dewatered. Since, as demonstrated above, different coal beds in a given location may produce water with distinctive S13CDIC, binary mixing models can be used to identify the origin of the produced water in multi-seam completions. In the examples below, two types of multi-seam well completion techniques are considered; open-hole and perforated casing.

Open-hole completions refer to a well bore that does not have cemented casing between producing horizons but instead a steel or PVC liner is slotted over the desired coal bed interval. This was the preferred method for the early and shallow wells in the Powder River Basin. As production moved basin-ward targeting deeper coals a more conventional method was used, in which the length of the well bore is lined with steel casing cemented and then perforated at the targeted coal bed intervals.

5.3.1. Open hole-completions

In the southern portion of the PRB study area the Wyodak seam splits into two main benches; the Anderson and Canyon coals. Open-hole wells that are producing from the Anderson-Canyon have anomalously low and variable S13CDIC, whereas single completion wells that are producing solely from the Wyodak down-dip of the split have uniformly enriched S13CDIC.

Open-hole commingled wells were correlated (Fig. 7). The lithology of strata immediately above, below and in the interburden between the coal seams is shown in cross-section B-B' (Fig. 7). The western side of the cross-section has a higher percentage of sand and larger interburdens between producing horizons, whereas the east side has less sand and smaller interburden intervals. Lower S13CDIC of produced water corresponds to wells that have sandy zones within the interburden of the two coals. The S13CDIC is higher in wells that are associated with less sand and thinner interburden intervals. In the case of these PRB open-hole wells, we suggest that the absence

Fig. 6. Sampled results of the twin and triplet wells. Note the depleted 813Cdic relative to the stratigraphically higher coal beds for the same geographical area.

of cement in open-hole wells allows water from isotopically depleted non-coal bed strata to infiltrate the well bore. Most likely this water is mixed with water from the adjacent sands within the interburden, effectively producing the lower S13CDIC recorded in open-hole commingled wells.

Binary mixing models are applied to examine relative percentages of water from different aquifers (Faure, 1991). If in fact the open-hole commingled wells allow non-coal bed water to infiltrate the well bore, then a binary mixing model can be used to develop a first order mixing

approximation of the of proportions of water from the coal and non-coal strata. We apply the following equation:

(813C)M = (813C)AfA + (^B^A-1)' (3)

where A and B represent the S13CDIC of the two end member components and M represents the measured S13CDIC of the measured well. For these analyses we use estimated end members of S13CDIC =

Table 2

Average 813Cdic and standard deviations within wells completed in the Big George coal bed sorted by natural gas fields.

Big George Field No. of samples Average 813Cdic Standard deviation

Bear Draw 7 12.6 1

Big Cat 15 11.9 2.1

Big George 14 13.4 1.4

Big Mike 13 13.1 1.6

Coal Gulch 2 12.3 1.6

Falxa 3 12.8 2.4

Highland 2 12.4 0.8

Kinney Divide 5 14.9 0.9

Powder Valley 16 13.9 1.1

River 6 13.9 0.6

Savageton 2 22.1 0.9

Tear Drop 3 13.2 0.3

Welles 3 13.4 0.5

Whiskey Draw 8 12.5 2

Williams Draw 5 17.2 0.7

16% (A) to represent a conservative isotopic composition of the water produced from the Wyodak coal bed and 613CDIC = —10% and (B) to represent water that has not been affected by methanogenesis (non-coal bed water). This model assumes that the coal bed and non-coal bed waters have comparable DIC concentrations, which appears to be a reasonable assumption for Powder River Basin aquifers (Sharma and Frost, 2008).

The results derived from this mixing equation are shown on Fig. 8. Because all of the open-hole commingled wells in this area have water

with ô13CDIC that fall between the two end members (A and B), we conclude that all include contributions of water from non-coal bed aquifers. The proportion of coal bed water in the mixtures varies from 26 to 71%. Mixing models show that W-07 and W-13 are producing the largest percentages of non-coal bed water, while W-11, W-16 and W-19 are producing the largest percentage of coal bed water (Fig. 8)

5.3.2. Perforated casing

In perforated casing-type completions, cemented casing secures the well bore from water infiltration over non-coal bearing intervals. Using 613CDIC of produced water from single-completion wells in the Lower Canyon (21.8%) and Wall (16.8%) coal zones as end member compositions, we can calculate the proportion of water contributed by each seam in a multi-seam completion. Carbon isotopic values of the comingled well (21.3%), fall between the values of the water from single completion wells in the two seams, suggesting that the water is a combination of the two horizons. However, it is strongly weighted towards the isotopic value of the stratigraphically higher coal bed, the Lower Canyon. Thus, it appears that most of the water being produced from these commingled wells is sourced from the Lower Canyon (approximately 80%, assuming comparable DIC contents in water from each coal bed) and the deeper Wall coal bed is not effectively dewatered by either of these wells.

These data indicate that once baseline conditions are established, 613CDIC can be used as a diagnostic measurement to identify the source of water in multi-zone wells and to determine the proportion of water being produced from each seam. Our analysis shows that open-hole

Meters Above Mean Sea Level

"0.3 XX

- 0.7 %o W-6

0.5 X,

Identifying Unconfined Coalbeds S13CDIC Variability

Meters Above Mean Sea Level

Fig. 7. Cross-section B-B' depicts lithology in around the producing coals. The north-west side of the cross-section has a sandier lithology and lower 813Cdic than the southeast side of the cross-section.

100% 90%+-80%-70%-60%-50% 40% 30% 20% 10% 0%

W-6 W-7 W-8 W-9 W-10 W-11 W-12 W-13 W-14 W-15 W-16 W-18 W-19 W-20

□ Ambient groundwater 59% 77% 64% 63% 39% 30% 56% 74% 58% 43% 29% 56% 33% 42%

■ CBNG water 41% 23% 36% 37% 61% 70% 44% 26% 42% 57% 71% 44% 67% 58%

Fig. 8. First order mixing approximations of the sources of water from open-hole type multiple producing zone wells. Note that for some wells over 70% of the produced water may be coming from non-coalbed strata.

type completions are not as efficient at dewatering the reservoir as wells that utilize fully cemented, perforated casing.

5.4. Isotopic variation along flow paths

On a regional scale, isotopic compositions of CBNG water in the PRB vary along groundwater flow paths. To illustrate this we show the S13CDIC and 6D measured across a 50-mile section of the Big George coal bed (Figs. 9 and 10). This cross-section is constructed parallel to groundwater flow direction (Bartos and Ogle, 2002 and references therein; Quillinan and Frost, 2012), and thus can be used to investigate isotopic changes along groundwater flow paths. S13CDIC values of groundwater are 23.3% on the up-gradient side nearer the basin margin and decrease to 13.7% towards the basin axis. In contrast 6D values increase across the same section from —151% to —135% from the basin margin to the axis. A number of processes may result in isotopic fractionations along flow paths. These may include groundwater mixing, hyperfiltration, and methanogenesis (Phillips et al., 1986; Coplen and Hanshaw, 1973; and Clark and Fritz, 1997).

5.4.1. Groundwater mixing

A decrease in S13CDIC and increase in 6D observed along flow paths could indicate a loss of hydraulic confinement within the reservoir (McLaughlin et al., 2011). However, incorporation of groundwater from other aquifers would likely be accompanied by distinct changes in water quality, which is not observed. Moreover, TDS exhibits a steady increase, which is not expected if coals are confined in some places but not in others. In addition, an increase in 6D inversely correlates with decreasing carbon isotopic compositions, which would not be expected of mixing of groundwater from two or more aquifers. Groundwater near the basin axis is around 20,000 years old and thus was likely recharged in a cooler climate (Frost and Brinck, 2005; Gillespie et al., 2004), thus compared to the present-day, these formation waters should be depleted with respect to 6D.

5.4.2. Hyperfiltration

Hyperfiltration can affect the isotopic composition of groundwater and may explain isotopic fractionations in low permeability formations

(e.g. Gref et al., 1965; and Hitchon and Friedman, 1969). Groundwater moving by advection along a steep gradient through a fine-grained matrix can act like a filtration membrane that retards larger molecules (Clark and Fritz, 1997). Experimental work with smectite and montmorillonite clays has recorded depletions in 6D (2.5 to 5%, Coplen and Hanshaw, 1973; Coplen, 1970) and enrichment for S13CDIC (1.5%; Fritz et al., 1987). This is the opposite of what is observed in the PRB. In addition, permeability measurements of PRB coals are high, ranging from tens of millidarcies to several darcys (Mavor et al., 2003), and is not conducive to hyperfiltration fractionations. The observed deuterium enrichment of 16% in the Big George coal bed does not suggest hyperfiltration.

5.4.3. Methanogenesis

Biogenic methanogenesis has been observed across the PRB (Flores et al., 2008). Progressive S13CDIC enrichment from methanogenesis along a gradient has been documented in dolomite aquifers and coal beds of the Atlantic Rim (Clark and Fritz, 1997; Quillinan, 2011). Therefore it seems reasonable that if methanogenic processes remain constant, S13CDIC will be continually enriched as groundwater flows farther into the PRB. Indeed this is what is observed, S13CDIC increase rapidly as groundwater enters the basin (— 5% to greater than 20%, Fig. 11). However, the pattern of S13CDIC in deeper coal beds is more complex. After the initial enrichment, S13CDIC gradually decrease as groundwater flows deeper into the basin (Fig. 11). Fig. 9 shows this gradual decrease in greater detail along flow gradient in the Big George coal seam.

The initial enrichment of S13CDIC of the groundwater is interpreted to reflect fractionations during methanogenesis, which Flores et al. (2008) showed occur by both acetate fermentation and the CO2 reduction pathways near the basin margins. Deeper in the basin CO2 reduction dominates (Flores et al., 2008). During CO2 reduction, the carbon isotope composition of the methane is a direct function of the initial CO2 (Shurr and Ridgley, 2002). Proximal to outcrop the CO2 reduction pathway uses a S13CDIC characteristic of the meteoric water recharging the coal bed, which has S13CDIC of approximately —12 to — 7%. As methanogens preferentially take up light carbon, S13CDIC becomes progressively more enriched, and the CO2-reducing methanogens are forced to utilize a more enriched CO2. The more enriched CO2 leads to more enriched S13CCH4, leaving more depleted S13CDIC (Fig. 9). Thus a

Meters Above Mean Sea Level

1435 „T t 1400 1340 1280 1220 1160 1100 1035 975 915 855 790 730 670

Big George Cross-Section 513C Variability

Explanation

— Gamma Log

Undifferentiated — Surface Elevation Lithology

Vertical exaggeration ~ 80x Approximate length 80 Km

Fig. 9. Cross-section A-A' highlights the decreasing 813Cdic of deeper coal beds along groundwater flow path.

gradually decreasing S13CDIC is expected if CO2 reduction is the main pathway that generates gas in the center of the basin.

This interpretation is consistent with the observed enrichment in the 6D along the same section. The 6D values enrich from —151% near the margin to —135.3% towards the basin axis (Fig. 10). Because the hydrogen isotope ratio of the methane is related to the hydrogen isotopic composition of the water, methanogenesis by the CO2 reduction pathway should enrich the SDH2O along flow path in the deeper portions of the basin. Deuterium enrichment in the groundwater as a result of methanogenesis is also observed in Atlantic Rim coal beds, and landfill leachates where large quantities of methane have been generated (Clark and Fritz, 1997; Quillinan, 2011; and Siegel et al., 1990).

These interpretations are supported by the isotopic compositions of the methane measured by Flores et al. (2008), who report a slightly more enriched S13CCH4 and a slightly more depleted SDCH4 towards the basin axis, interpreted as a result of a shift towards the CO2 reduction pathway. 613C enrichment and depletion in 6D in gas is associated with depletion of 613C and enrichment of 6D in the water, just as we observed.

The restriction of the acetate fermentation pathway to the basin margins could be a result of two conditions that differ between basin margin and center. First is the supply of acetate: fermentation bacteria consume acetate and any other fermentative substrates and may exhaust it as groundwater flows towards the basin center. The second may relate to increases in salinity along flow path. Acetate fermentation dominates in freshwater sediments, whereas CO2 reduction dominates in marine sediments (Whiticar et al., 1986). Whereas the waters of PRB coal beds are far less saline than

seawater, TDS concentrations exceed 5000 mg/L near the basin axis (Quillinan and Frost, 2012). In any case, it appears that a decrease of up to 10% in S13CDIC can result from methane production by CO2 reduction across the basin.

5.5. Volume of gas generation

Positive S13CDIC is a powerful indicator of methanogenesis because even small amounts ofmethanogenesis can enrich the relatively small pool of DIC. In contrast to the DIC pool, the 6D of formation water is not as easily enriched. Clark and Fritz (1997, pg. 157) suggest that "a measurable positive shift in 6D requires a large quantity of methane generation". Therefore waters that have both S13CDIC and 6D enrichment must be associated with relatively large quantities of methane generation. A positive shift of 16% is recorded in the Big George coal seam suggesting a significant amount of methanogenesis. This qualitative evidence of large amounts of methane generation is supported by the volume of biogenic methane that has been recovered from the Big George coal bed: the more than 2000 BCF (5.6 x 107 m3) produced from that coal exceeds the combined total production from all other coal beds in the PRB (Quillinan and Frost, 2012). We therefore suggest that the combination of enriched S13CDIC and enriched 6D is a robust indicator of environments that have generated large quantities of biogenic methane.

6. Conclusions

Our stable isotopic investigation of produced waters associated with coalbed natural gas in the Powder River Basin ofWyoming and Montana

Meters Above Mean Sea Level

1525 1435 1400 1340 1280 1220 1160 1100 1035 975 915 855 790 730 670

iT Not Sampled Inferred

Big George Cross-Section 513DH20Yariability

Explanation

Coal — Gamma Log

Undifferentiated — Surface Elevation Lithology

Vertical exaggeration ~ 80x Approximate length 80 Km

Big George Coal Seam

5D„2o

n High -135.0 9

Not Sampled Inferred

Not Sampled Inferred

Not Sampled Inferred

-152.5 9 !

A B-104 n Iii B-102 B-113 Sampled \B-35 W B-103 -B-10,

B-23 B-2

B-17 T

Not Sampled ' . Not Sampled

No t Sampled

0 2.5 5 10 Miles W-33 CumpbtU W-44 ^A'

Fig. 10. Cross-section A-A' highlights the increasing 8DH20 of deeper coal beds along groundwater flow path.

Low 152.5 9

has shown that S13CDIC compositions of +10 to +25% identify produced water that is associated with methanogenesis. This is the range of isotopic compositions of produced water from most PRB CBNG reservoirs. Our analyses of water collected from wells that have been pumping for between 1 and 15 years show no correlation between the amount of water that has been withdrawn from the well and the S13CDIC of the water now being produced. Most importantly, coal bed produced water is isotopically distinct from non-coal bed waters. Regardless of how long the well has been in production,

S13CDICvs. Depth

30 25 20 15 10 5 0 -5 -10

♦ Shallow wells o Deep Wells

400 600

Depth (m)

Fig. 11. 813CDIC as a function of depth. Shallow wells (<300 m) have a rapid 813Cdic enrichment trend, while deeper wells illustrate a gradual decreasing trend in 813CDIC.

coal bed waters have positive 813CDIC, whereas non-coal bed waters are negative.

Although the carbon isotopic compositions of produced waters from individual PRB coal beds vary on a regional scale, they are uniform on a local field scale, and different coal beds have distinct 813CDIC. Stacked coal beds are easily distinguished from each other: the deeper beds are less enriched with respect to carbon and more enriched with respect to deuterium. This attribute can be used to help identify the source of water in single and multi-production zone environments, and help determine the efficiency of dewatering from different coal beds.

813CDIC isotope compositions of produced water that is lower than +10% should be interpreted to indicate the possibility that coal beds are in communication with waters that have not hosted methanogenesis, usually from another aquifer. Binary mixing models can be used to estimate contributions from other aquifers or from multiple producing zones. Extreme negative values 813CDIC within a CBNG reservoir (<—15%) occurs in water from some shallow wells (< 500 ft.) and indicate that SO4 reducing bacteria may have been oxidizing methane.

The particular methanogenic pathways operating at different places along flow paths contribute to the variability that is observed regionally. Enriched 813CDIC values decrease and 8D is enriched with increasing residence time in the deeper coal beds where CO2 reduction has been identified as the dominant methanogenic pathway.

This report presents the largest 813CDIC and 8D dataset collected to date from CBNG produced waters in order to characterize multiple coal zones across the basin and describe methanogenic processes at work in different parts of the PRB. The carbon isotope composition is instructive both with respect to the hydrological characteristics and to

the biological characteristics within a reservoir. These stable isotopic analyses can provide inexpensive but valuable information during the exploration and the production phases of coal bed natural gas wells. Most importantly, the carbon isotope composition of a reservoir can be used to identify confined reservoirs and to identify efficient dewatering of coal beds in single and multiple zones: information that can help minimize unnecessary water production.

Acknowledgments

Analyses were funded through the Wyoming State Geological Survey and the Wyoming Water Development Commission. Funding for publication came through the Office of Research and Economic Development. We thank Anadarko Petroleum, Williams Petroleum, Fidelity Exploration, and Lance Petroleum who provided site access and water samples.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.coal.2013.10.006.

References

Ayers, W.B., Kaiser, W.R., 1984. Lacustrine-interdeltaic coal in the Fort Union Formation (Paleocene), Powder River Basin, Wyoming and Montana. In: Rahmani, RA., Flores, R.M. (Eds.), Sedimentology of Coal and Coal-bearing Sequences. International Association of Sedimentologists Special Publication, 7, pp. 61-84. Barker, J.F., Fritz, P., 1981. Carbon isotope fractionation during microbial methane

oxidation. Nature 293, 289-291. Bartos, T.T., Ogle, K.M., 2002. Water quality and environmental isotopic analyses of ground-water samples collected from the Wasatch and Fort Union formations in areas of coalbed methane development—implications to recharge and groundwater flow, eastern Powder River Basin, Wyoming. U.S. Geological Survey Water-Resources Investigations Report 2002-4045 (88 pp.). Campbell, C.E., Pearson, B.N., Frost, C.D., 2008. Strontium isotopes as indicators of aquifer communication in an area of coal bed natural gas production, Powder River Basin, Wyoming and Montana. Rocky Mt Geol. vol. 43 (2), 149-175. Carothers, W.W., Kharaka, Y.K., 1980. Stable carbon isotopes of HCO— in oil-field waters:

implications for the origin of CO2. Geochim. Cosmochim. Acta vol. 44, 323-332. Carter, SA, 2008. Geochemical Analysis of the Powder River, Wyoming/Montana and an Assessment of the Impacts of Coalbed Natural Gas Co-produced Water. M.S. Thesis University of Wyoming Department of Geology and Geophysics (May 2008). Clark, I., Fritz, P., 1997. Environmental Isotopes in Hydrogeology. CRC Press LLC, Boca Raton, Florida.

Coleman, D.D., Liu, C.-L., Riley, K.M., 1988. Microbial methane in the shallow Paleozoic

sediments and glacial deposits of Illinois, U.S.A. Chem. Geol. vol. 71 (1-3), 23-40. Coplen, T.B., 1970. Isotopic Fractionation of Water by Ultrafiltration. (96 pp. Ph.D.

Dissertation) University of Chicago. Coplen, T.B., Hanshaw, B.B., 1973. Ultrafiltration by a compacted clay membrane. I. Oxygen

and hydrogen isotope fractionation. Geochim. Cosmochim. Acta 37,2295-2310. Craig, H., 1961. Isotopic variations in meteoric waters. Science vol. 133 (3465), 1702-1703. Faiz, M., Stalker, L, Sherwood, N., Saghafi, A., Wold, M., Barclay, S., Choudhury, J., Barker, W., Wang, I., 2003. Bio-enhancement of coal bed methane resources in the southern Sydney Basin. Australian Petroleum Production and Exploration Association., 43 595-610. Faure, G., 1991. Mixing and Dilution, Principles and Applications of Inorganic Geochemistry: A comprehensive Textbook for Geology Students, 2nd ed. Prentice-Hall, Inc., New Jersey, p. 626. Flores, R.M., 1999. Wyodak-Anderson coal zone in the Powder River Basin Wyoming and Montana: A tale of uncorrelatable coal beds. In: Miller, Roger, W. (Eds.), Coalbed Methane and Tertiary Geology of the Powder River Basin Wyoming and Montana, Wyoming Geological Association Guidebook Flores, R.M., Bader, L.R., 1999. Fort Union coal in the Powder River Basin, Wyoming and Montana. A Synthesis: US Geological Survey Professional Paper 1625-A (Chapter PS). Flores, R.M., Ochs, A.M., Bader, L.R., Johnson, R.C., Vogler, D., 1999. Framework geology of the Fort Union coal in the Powder River Basin, Chapter PF. In: Fort Union Coal Assessment Team (Ed.), 1999 Resource Assessment of Selected Tertiary Coal Beds and Zones in the Northern Rocky Mountains and Great Plains Region. U.S. Geological Survey Professional Paper, 1625-A, p. 58. Flores, R.M., Rice, C.A., Stricker, G.D., Warden, A., Ellis, M.S., 2008. Methanogenic pathways of coalbed gas in the Powder River Basin, United States: the geologic factor. In: Flores, R.M. (Ed.), Microbes, Methanogenesis and Microbial Gas in Coal. International Journal of Coal Geology Special Issue, vol. 76, pp. 52-75. Freeze, RA, Cherry, J.A., 1979. Groundwater: Englewood Cliffs. Prentice Hall, New Jersey. Fritz, S., Hinz, D.W., Grossman, E.L., 1987. Hyperfiltration induced fractionation of carbon

isotopes. Geochim. Cosmochim. Acta 51,1121-1134. Frost, C.D., Brinck, E., 2005. Strontium isotopic tracing of the effects of coal bed natural gas (CBNG) development on shallow and deep groundwater systems in the Powder River

Basin, Wyoming. Wyoming State Geological Survey Report of Investigations No. 55, pp. 93-107.

Frost, C.D., Pearson, B.N., Ogle, K.M., Heffern, E.L., Lyman, R.M., 2002. Sr isotopic tracing of aquifer interactions in an area of coal and methane production, Powder River Basin, Wyoming. Geology vol. 30, 923-926.

Gillespie, A.R., Porter, S.C., Atwater, B.F., 2004. The Quaternary Period in the United States [of America]. Developments in Quaternary Science.Elsevier, Amsterdam (ISBN 978-0444-5147-4).

Goolsby, A., Finley, A., 2000. Correlation of Fort Union Coals, Powder River Basin, WY: a proposed new concept. Wyoming Geological Association, 5Ist Annual Field Conference Guidebook, pp. 51-74.

Gorody, A.W., 1999. The origins of natural gas in the Tertiary Coal Seams on the Eastern Margin of the Powder River Basin. In: Miller, W.R. (Ed.), Coalbed Methane and the Tertiary Geology of the Powder River Basin, Wyoming and Montana, Wyoming Geological Association 50th Annual Field Conference Guidebook, pp. 89-102.

Gref, D.L., Friedman, I., Meets, W.F., 1965. The origin of saline formation waters, isotopic fractionation by shale micropore systems. Ill. State Geol. Surv. Circ. 393,32.

Grossman, E., Coffman, B.K., Fritz, S.J., Wada, H., 1989. Bacterial production of methane and its influence on ground-water chemistry in east-central Texas aquifers. Geology 17,495-499.

Hitchon, B., Friedman, I., 1969. Geochemistry and origin of formation waters in the Western Canada sedimentary basin. Stable isotopes of hydrogen and oxygen. Geochim. Cosmochim. Acta 33,1321-1349.

Jenden, P.D., Kaplan, I.R., 1986. Comparison of microbial gases from the Middle America Trench and Scripps Submarine Canyon: implications for the origin of natural gas. Appl. Geochem. vol. 1,631-646.

Jones, N.R., 2008. Coal bed nomenclature and distribution. In: Copeland, D.A., Weald, M.L. (Eds.), Water Associated with Coal Beds in Wyoming's Powder River Basin: Geology, Hydrology, and Water Quality. Wyoming State Geological Survey Exploration Memoir No. 2.

Jones, N.R., 2010. Genesis of thick coal deposits and their unique angular relationships: Powder River Basin, Wyoming. Wyoming State Geological Survey Report of Investigations No. 60-10.

Love, J.D., and Christiansen, A.C., compilers, 1985. Geologic map of Wyoming. U.S. Geological Survey, 3 sheets, scale 1:500,000.

Martini, A.M., Walter, L.M., Budai, J.M., Ku, T.C.W., Kaiser, C.J., Schoell, M., 1998. Genetic and temporal relations between formation waters and biogenic methane: Upper Devonian Antrim Shale, Michigan Basin, USA. Geochim. Cosmochim. Acta vol. 62 (10), 1699-1720.

Martini, A.M., Walter, L.M., Ku, T.C.W., Budai, J.M., Mcintosh, J.C., Schoell, M., 2003. Microbial production and modification of gases in sedimentary basins: a geochemical case study from a Devonian Shale gas play, Michigan Basin. AAPG Bull. vol. 87,1355-1375.

Mavor, M.J., Russell, B., Pratt, T.J., 2003. Powder River Basin Ft. Union Coal reservoir properties and production decline analysis. Society of Petroleum Engineers, SPE. (84427).

McClurg, J.E., 1988. Peat forming wetlands and the thick Powder River Basin coals. Wyoming Geological Association Guidebook, 39th Field Conference, pp. 229-236.

Mcintosh, J., Martini, A., Petsch, S., Huang, R., Nussleain, K., 2008. Biogeochemistry of the Forest City Basin coalbed methane play. Int. J. Coal Geol. vol. 76,111-118.

McLaughlin, J.F., Frost, C.D., Sharma, S., 2011. Isotopic analysis of Atlantic Rim waters, Carbon County, Wyoming: a new tool for characterizing coalbed natural gas systems. AAPG Bull. vol. 95 (2), 191-218.

Montgomery, S.L., 1999. Powder River basin, Wyoming: an expanding coalbed methane (CBM) play. AAPG Bull. vol. 83 (8), 1207-1222.

Mook, W.G., Tan, F.C., 1991. Stable carbon isotopes in rivers and estuaries. In: Degens, E.T., Kempe, S., Richey, J.E. (Eds.), Biogeochemistry of Major World Rivers. John Wiley and Sons, Chichester, UK, pp. 245-264.

Moore, T.A., 2012. Coalbed methane: a review. Int. J. Coal Geol. vol. 101, 36-81.

Pearson, B.N., 2002. Sr isotope ratio as a monitor of recharge and aquifer communication, Paleocene Fort Union Formation and Eocene Wasatch Formation, Powder River Basin, Wyoming and Montana. University of Wyoming Department of Geology and Geophysics 151 (MS Thesis).

Phillips, F.M., Bentley, H.W., Davis, S.N., Elmore, D., Swanick, G., 1986. Chlorine-36 dating of very old groundwater, 2. Milk River aquifer, Alberta, Canada, Water Resources Research 22,2013-2016.

Pocknall, D.T., Flores, R.M., 1987. Coal palynology and sedimentology in the Tongue River Member, Fort Union Formation, Powder River Basin, Wyoming. Society of Economic Paleontologists and Mineralogists. Palaios vol. 2 (2), 133-145.

Quillinan, SA, 2011. Geochemical and Isotopic Characterization of Water Associated with Wyoming's Coalbed Natural Gas Development. University of Wyoming, Laramie (M.S. thesis, 205 pp.).

Quillinan, S.A., Frost, C.D., 2012. Spatial variability of coalbed natural gas produced water quality, Powder River Basin, Wyoming: implications for future development. Wyoming State Geological Survey Report of Investigations No. 64 (56 pp.).

Quillinan, S.A., McLaughlin, J.F., Frost, C.D., 2012. Geochemical and stable isotopic analysis of the Tongue River and associated tributaries in the Powder River Basin: an analysis of the cause of annual elevated salinity in spring runoff. Wyoming State Geological Survey Report of Investigations No. 63-2012 (15 pp.).

Rice, D.D., 1992. Controls, habitat, and resource potential of ancient bacterial gas. In: Vialy, R (Ed.), Bacterial Gas Editions Technip, pp. 91-118.

Rice, D.D., 1993. Composition and origins of coalbed gas. In: Law, B.E., Rice, D.D. (Eds.), Hydrocarbons from Coal. AAPG Studies in Geology, vol. 38, pp. 159-184 (Tulsa, Oklahoma).

Rice, D.D., Claypool, G.E., 1981. Generation, accumulation and resource potential of biogenic gas. Am. Assoc. Pet. Geol. Bull. vol. 74,1343.

Rice, D.D., Threlkheld, C.N., 1982. Occurrence and origin of natural gas in ground water, southern Weld County, Colorado. U.S. Geological Survey Open-File Report 1982-496 (6 pp.).

Rice, C.A., Ellis, M.S., Bullock Jr., J.H., 2000. Water Co-produced with coalbed methane in the Powder River Basin, Wyoming: preliminary compositional data. U.S. Geological Survey Open File Report 2000-372 (20 pp.).

Schoell, M., 1980. The hydrogen and carbon isotopic composition of methane from natural gases of various origins. Geochim. Cosmochim. Acta vol. 44, 649-661.

Seeland, DA., 1993. Origin of thick lower Tertiary coal beds in the Powder River Basin, Wyoming and Montana—some paleographic constraints. U.S. Geol. Surv. Bull. 1917-Q(13 pp.).

Sharma, S., Frost, C., 2008. An innovative approach for tracing coalbed natural gas co-produced water using stable isotopes of carbon and hydrogen. Groundwater vol. 46 (2), 329-334 (March-April 2008).

Sholes, M.A., Cole, G.A., 1981. Depositional history and correlation problems of the Anderson-Dietz coal zone, southeastern Montana. Mt. Geol. vol. 18 (2), 35-45.

Shurr, G.W., Ridgley, J.L., 2002. Unconventional shallow biogenic gas systems. AAPG Bull. vol. 86 (11), 1939-1969 (November 2002).

Siegel, D.T., Stoner, D., Byrnes, T., Bennett, P., 1990. A geochemical process approach to identify inorganic and organic ground-water contamination. Ground Water Management, Number 2, Fourth National Outdoor Action Conference on Aquifer Restoration, Ground Water Monitoring and Geophysical Methods, Las Vegas, Nevada. National Ground Water Association, Columbus, OH, pp. 1291-1301.

Waldron, S., Hall, A.J., Fallick, A.E., 1999. Enigmatic stable isotope dynamics of deep peat methane. Glob. Biogeochem. Cycles vol. 13 (1), 93-100.

Whiticar, M.J., Faber, E., Schoell, M., 1986. Biogenic methane formation in marine and freshwater environments: CO2 reduction vs. acetate fermentation-isotope evidence. Geochim. Cosmochim. Acta 50,693-709.

Woltemate, I., Whiticar, M.J., Schoell, M., 1984. Carbon and hydrogen isotopic composition of bacterial methane in a shallow freshwater lake. Limnol. Oceanogr. vol. 29, 985-992.

Wyoming Oil Gas Conservation Commission, 1987-2010. Coalbed production. http:// wogcc.state.wy.us/coalbedchart.cfm.

Wyoming State Geological Survey, 2010. at WSGS.uwyo.edu (accessed June 2010).