ABSTRACT Title of Document: MONITORING LEVELS OF DISSOLVED METHANE AND METALS IN MARYLAND STREAMS OVERLYING THE MARCELLUS SHALE PRIOR TO HYDRAULIC FRACTURING Caroline G. Coulter, Master of Science, 2015 Directed By: Associate Professor Johan Schijf, Marine Estuarine Environmental Sciences In western Maryland above the Marcellus Shale, 25 streams were monitored for baseline concentrations of dissolved CH4, Sr 2+, and Ba2+, constituents that may be affected by fracking activity. Migrated shale CH4 may intrude streams, and Sr 2+ and Ba2+ may be introduced by fracking fluid leakages or spills. Stream constituents were also measured in Maryland’s coastal plain for reference. In western Maryland, CH4 concentrations are significantly variable yet in agreement with concentrations reported for other North American pristine rivers. Measurements of δ13C-CH4 suggest CH4 is primarily biogenic. Dissolved Sr 2+ and Ba2+ are spatially and temporally variable, although Sr/Ba ratios are relatively stable at most sites, indicating these may be useful fracking fluid tracers. Major ions Na+, K+, Mg2+, Ca2+, Cl−, 24SO  and 3HCO were measured. These were elevated relative to Sr2+ and Ba2+ and are not suitable fracking fluid tracers. Major ions were highly variable, indicative of variable bedrock geology in western MD. MONITORING LEVELS OF DISSOLVED METHANE AND METALS IN MARYLAND STREAMS OVERLYING THE MARCELLUS SHALE PRIOR TO HYDRAULIC FRACTURING By Caroline G. Coulter Thesis submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Master of Science 2015 Advisory Committee: Associate Professor Johan Schijf, Chair Professor Keith Eshleman Assistant Research Professor Hali Kilbourne Assistant Professor Laura Lapham © Copyright by Caroline G. Coulter 2015 ii Acknowledgements First I acknowledge my major advisor Dr. Johan Schijf as well as Dr. Laura Lapham, both of whom provided much guidance, help, and patience in this process. Acknowledgement is also given to committee members Dr. Hali Kilbourne and Dr. Keith Eshleman for their guidance and their enthusiasm for my project. I also thank Dr. Andrew Heyes who contributed financial support as well as insight during my time as a student. Acknowledgement is given to Maryland Department of Natural Resources for supporting this work and allowing access to sites. Thanks to the IRMS lab of Dr. Jeff Chanton at Florida State University, including Claire Langford who oversaw sample analysis. Thanks are owed to former students and Faculty Research Assistants in the Schijf and Lapham labs who provided advice and assistance in both lab and field activities: Emily Christenson, Kathleen Marshall, Dr. Cedric Magen, Katherine Hoffman, and Joshua Condon. Thanks are in order for the many CBL friends who provided encouragement and laughs, in particular my office mates and house mates: Lauren Gelesh, Yuan Yuan Xu, Alex Atkinson, Christina Goethel, Kathleen Marshall, Jenna Luek, MengJie Zhang, and Rachel Clark. Finally, thanks to my family members in Arkansas for their continuous love and support. iii Table of Contents Acknowledgements ....................................................................................................... ii List of Tables ................................................................................................................ v List of Figures ............................................................................................................. vii Chapter 1: Hydraulic fracturing in the U.S. and its impact on the environment .......... 1 1.1 Introduction ........................................................................................................ 1 1.2 Oil and natural gas production ........................................................................... 4 1.2.1 Origins of oil and natural gas ...................................................................... 4 1.2.2 Drilling of conventional oil and gas reservoirs ........................................... 5 1.2.3 Drilling of unconventional oil and gas reservoirs ....................................... 6 1.3 Environmental challenges related to fracking.................................................... 9 1.3.1 Range of environmental impacts ................................................................ 9 1.3.2 Impacts of fracking fluids on water resources .......................................... 11 1.3.3 Methane release associated with fracking................................................. 14 1.4 Environmental management in the Marcellus Shale region of Maryland ....... 17 1.5 Study of dissolved methane and metals in Maryland surface waters .............. 20 Chapter 2: Baseline monitoring of dissolved metals in western Maryland streams ... 26 2.1 Introduction ...................................................................................................... 26 2.1.1 Hydraulic fracturing fluids as sources of water contamination ................ 26 2.1.2 Characteristics of fracking fluids and wastes............................................ 27 2.1.3 Fates of fracking fluids in surface and groundwater ................................. 28 2.1.4 Use of metals as fracking fluid tracers ...................................................... 32 2.1.5 Monitoring of Sr2+ and Ba2+ in western and southern MD streams .......... 35 2.1.6 Characterizing waters by major ion composition ..................................... 37 2.2 Materials and methods ..................................................................................... 40 2.2.1 Sampling sites ........................................................................................... 40 2.2.2 Field sampling ........................................................................................... 46 2.2.3 Analysis of Sr2+ and Ba2+ .......................................................................... 47 2.2.4 Analysis of major cations.......................................................................... 48 2.2.5 Analysis of major anions .......................................................................... 49 2.2.6 Analysis of Acid Neutralizing Capacity ................................................... 51 2.2.7 Analysis of Dissolved Organic Carbon..................................................... 52 2.3 Results .............................................................................................................. 53 2.3.1 Dissolved Sr2+ and Ba2+ concentrations .................................................... 53 2.3.2 Major ion concentrations .......................................................................... 59 2.4 Discussion ........................................................................................................ 61 2.4.1 Charge balance in western MD streams.................................................... 61 2.4.2 Major ions as tracers of fracking fluids .................................................... 63 2.4.3 Sr2+ and Ba2+ as tracers of fracking fluids ................................................ 71 2.4.4 Sr and Ba solubility and its effect on tracer use ........................................ 76 2.4.5 Implications and Conclusions ................................................................... 81 iv Chapter 3: Monitoring of dissolved methane in western Maryland streams .............. 84 3.1 Introduction ...................................................................................................... 84 3.1.1 Sources of methane in natural waters ....................................................... 84 3.1.2 Methane in groundwater ........................................................................... 85 3.1.3 Methane in surface waters ........................................................................ 87 3.1.4 The isotopic composition of carbon in dissolved CH4 .............................. 89 3.1.5 Methane behavior in relation to fracking .................................................. 90 3.1.6 Monitoring baseline dissolved CH4 in western and southern MD streams ............................................................................................................................. 94 3.2 Materials and methods ..................................................................................... 98 3.2.1 Field sampling methods ............................................................................ 98 3.2.2 Analysis of CH4 partial pressure in the headspace ................................. 101 3.2.3 Theory ..................................................................................................... 102 3.2.4 Calculation of stream CH4 saturation ..................................................... 103 3.2.5 Analysis of δ13C-CH4 .............................................................................. 104 3.3 Results ............................................................................................................. 106 3.3.1 Temporal Variability ............................................................................... 106 3.3.2 Spatial variability .................................................................................... 107 3.3.3 Carbon isotopic composition of CH4 ...................................................... 113 3.4 Discussion ...................................................................................................... 116 3.4.1 Temporal variability of stream CH4 ........................................................ 117 3.4.2 Spatial variability of stream CH4 ............................................................ 119 3.4.3 Carbon Isotope Measurements ................................................................ 128 3.4.4 Monitoring Implications and Conclusions .............................................. 134 Chapter 4: Conclusions ............................................................................................. 138 4.1 Use of Sr2+, Ba2+, and major ions as fracking fluid tracers ............................ 138 4.2 Use of dissolved CH4 for tracing fracking contamination ............................. 140 4.3 Further implications for fracking regions ...................................................... 142 Appendices ................................................................................................................ 145 Appendix A ........................................................................................................... 145 Appendix B ........................................................................................................... 156 Bibliography ............................................................................................................. 160 v List of Tables Table 2.1 Names and locations of sampling sites in western MD for the Feb 2012– Feb 2013 and Apr 2013–Apr 2014 sets and sampling sites in Calvert County from the Summer 2014 set. Site name SPRS represents the Savage River Pumping Station and Reservoir represents the Piney Creek Reservoir. WTP Intake and WTP Finished represent the intake of the Frostburg City water treatment plant and the treated water from the plant, respectively. Sites in the Apr 2013–Apr 2014 set are named according to catchment, where CASS indicates the Casselman River catchment, GEOR the Georges Creek catchment, PRUN the Potomac River Upper North Branch catchment, SAVA the Savage River catchment, and YOUG the Youghiogheny River catchment. Sites in this set are numbered based on stream order, with numbers in the 100s representing 1st order streams, those in the 200s representing 2nd order, etc. Designation ‘S’ indicates long-term monitoring sites which have historical data sets, and those with ‘D’ are newly established monitoring sites. Summer 2014 site names are associated with the stream where the site is found, with HELL representing Hellen Creek, GRAY representing Grays Creek, PARK representing Parkers Creek, and THOM representing Thomas Branch. Note that for site THOM01, June and July sampling locations are slightly different due to higher water levels in July at Thomas Branch. The characteristic site environment is given in the ‘Environment’ column. Table 2.2 Saturation state (Ω) of western MD samples found to be oversaturated in BaSO4, with calculated free Ba 2+ concentration (nM), free 24SO  concentration (µM), and Ion Concentration Product (ICP) of each sample. Table A1 Stream concentrations of Na+, K+, Mg2+, and Ca2+ (µM), Sr2+ and Ba2+ (nM), Cl−, 42SO  , 3NO , 3HCO , and DOC (µM), and pH. Values of 3HCO were approximated from charge balance, unless marked by * which indicates 3HCO measured by Gran Titration. Table B1 Dissolved CH4 concentrations (nM) in western MD streams from the Sep 2012–Feb 2013 sample set and the Apr 2013–Apr 2014 sample set. The meaning of site names is described in Chapter 2. Concentration values are taken as the mean of sample duplicates, and relative errors, determined as half the difference between duplicate samples, is <8% for most values. The designated site environment is given in the column labeled ‘Env.’ with ‘A’ indicating anthropogenically influenced sites, ‘F’ indicating forested sites, and ‘W’ indicating wetland sites. vi Table B2 Values of δ13C-CH4 (‰) measured in western MD streams from the Sep 2012–Feb 2013 sample set as well as the Apr 2013–Apr 2014 sample set. Values were determined from the mean δ13C-CH4 measurement of sample duplicates, and the relative error, determined as half the difference between duplicate samples, was <2% for most samples. Table B3 Dissolved CH4 concentrations (nM) and δ 13C-CH4 values (‰) in Calvert County streams in June and July of 2014. Site names were designated based on the stream in which sites were located, as defined in Chapter 2. Concentration values were determined from the mean of sample duplicates and showed relative error <8% for most measurements. δ13C-CH4 values were taken from the mean of sample duplicates and relative errors were <2% for most samples. Table B4 Longitudinal gradient of dissolved CH4 concentrations (nM) measured in Sep 2013. Duplicate samples were measured at the CASS101D site (0 m), as well as from locations downstream (−30 m and −20 m) and upstream (+10 m and +20 m). Errors were determined as half the difference between duplicate samples. vii List of Figures Figure 1.1 Map of shale formations with natural gas access in the lower 48 states. http://www.eia.gov/oil_gas/rpd/shale_gas.pdf Figure 1.2 Diagram showing conventional and unconventional drilling techniques for oil and natural gas. Conventional drilling accesses reservoirs of oil and gas trapped in reservoirs below a seal. Unconventional drilling utilizes a lateral wellbore and hydraulic fractures to extract oil and gas from shale formations. http://www.eia.gov/todayinenergy/detail.cfm?id=110 Figure 1.3 Map of western Maryland showing areas underlain by the Marcellus Shale in blue. http://www.wmrcd.org/ Figure 1.4 Map of the western Maryland study region with the study sites represented as purple points. Catchment areas studied are labeled, with the Youghiogheny River catchment shown in green, Casselman River catchment in yellow, Savage River catchment in blue, Georges Creek catchment in pink, and Potomac River Upper North Branch (‘Potomac River U N Branch’) catchment in orange. The wider Youghiogheny River basin is indicated in green, and the wider Potomac River North Branch basin in orange. Figure 2.1 Map of western MD stream sites for the Feb 2012–Feb 2013 and Apr 2013–Apr 2014 sampling sets. Background and symbol colors indicate catchment, with yellow indicating the Casselman River catchment, blue indicating the Savage River catchment, pink the Georges Creek catchment, orange the Potomac River Upper North Branch catchment, and green the Youghiogheny River catchment. Black symbol indicates the Frostburg WTP. The greater Youghiogheny River basin is indicated in green and the greater Potomac River North Branch in orange. Figure 2.2 Map of stream sites sampled in Calvert County, MD, in June and July 2014. Background and symbol colors correspond to stream catchment, with blue indicating the Patuxent River Lower catchment and purple indicating the West Chesapeake Bay catchment. Figure 2.3 Temporal variability in dissolved Sr2+ and Ba2+ concentrations (nM) across 1 year. Panels A, B, D, and F are sites in the Casselman River catchment, and panels C and E are sites in the Savage River catchment. SRPS indicates Savage River Pumping Station. Sites in Panels B, D, and E are from the 2012−2013 sample set, and those in Panels A, C, and F are from the 2013−2014 sample set. viii Figure 2.4 Values of log([Ba]/[Sr]) at each sampling site, with Feb 2012–Feb 2013 sites shown in the left panel, Apr 2013–Apr 2014 sites shown in the center panel, and June−July 2014 Calvert County sites shown in the right panel. Bars plotting above 0 indicate Ba2+ is elevated in concentration relative to Sr2+ and those plotting below 0 indicate Sr2+ is elevated. Figure 2.5 Concentrations of Sr2+ vs. Ca2+ in all stream waters. Figure 2.6 Concentrations of Ca2+ vs. 3HCO in western MD stream samples collected Oct’ 2013. Figure 2.7. Trilinear diagram showing relative contributions (%) of Mg2+, Ca2+, and Na++K+ cations in western MD and Calvert County streams. Marker color represents catchment as in Fig. 2.1. Average cation composition of shallow groundwater in a region overlying the Marcellus Shale (New York state) is represented by the red star (Lautz et al., 2014). Figure 2.8. Geologic map of western MD by Brezinksi et al. (2013). Map background shades represent the underlying bedrock type. Green shades represent Mississippian and younger formations, and purple shades represent Devonian and older formations. All western MD stream sites are marked, with marker color indicating catchment as in Fig. 2.1 Figure 2.9. Bilinear plot of K+ vs. Mg2+ in western MD stream samples, with samples categorized by the underlying bedrock era, either Devonian or Pennsylvanian. Figure 2.10 Plot of Na+ vs Cl– concentrations in deep Appalachian brines which occur in fracking fluids (Warner et al., 2012), in road salt effluent (Lautz et al., 2014), and in western MD streams. Figure 2.11 Plot of Mg2+ vs. Cl– concentrations in deep Appalachian brines which occur in fracking fluids, in shallow groundwater overlying the Marcellus Shale (Warner et al., 2012), and in western MD streams. Figure 2.12 Values of log([Ba]/[Sr]) found in western MD streams (light gray) and in fracking fluids (dark gray). Western MD values were determined from average molar Sr2+ and Ba2+ concentrations from each time of sampling. Fracking fluid values were determined from molar concentrations from fracking fluid analyses as reported by Barbot et al. (2013) (A), Hayes et al. (2009) (B), He et al., (2014) (C, D, and E), Warner et al. (2012) (F), and Warner et al. (2013b) (G). ix Figure 3.1 Diagram showing a fracking well drilled into the Marcellus Shale, with red arrows indicating mechanisms for CH4 contamination of water. Arrow A represents leakage of Marcellus Shale CH4 through failed wellbore casing and B represents transport of intermediate formation CH4 through annuli. Arrow C represents transport of CH4 from shallow groundwater to streams. Arrow D represents a potential pathway of Marcellus Shale CH4 from depth and E represents CH4 migration from depth facilitated by pre-existing faults Figure 3.2 Temporal variability in western MD stream CH4 concentrations (nM) across 1 year. Top panels (A and B) show sites from the 2012–2013 set, with high CH4 concentration sites shown in A and low concentration sites in B. Bottom panels (C and D) show sites from the 2013–2014 set, with high concentration sites shown in C and intermediate concentration sites in D. SRPS indicates the Savage River Pumping Station. Note that the scale in panels B and D is an order of magnitude lower than in A and C. Bar color indicates site catchment, with yellow indicating Casselman River, blue Savage River, and green Youghiogheny River. Error bars indicate half of the difference between the sample duplicates, <8% of the concentration value in most cases. Figure 3.3. Methane concentrations (nM) in Calvert County from June and July 2014 measurements, grouped by stream setting (forested, wetland, or anthropogenically influenced). Figure 3.4 Spatial distribution of CH4 concentrations (nM) according to site setting, with sites characterized as forested, wetland, or anthropogenically influenced. Concentrations measured in Aug/Sep sampling in western MD (of both the 2012−2013 and 2013−2014 sets) are shown in (A) and those measured in July in Calvert County shown in (B). Concentrations shown in B are the same as those shown in Fig. 3.3 for July 2014. Note the order of magnitude higher concentrations in (B) which shows Calvert County measurements relative to (A) which shows those of western MD. Error bars indicate half of the difference between duplicate samples, <8% of the concentration values for most measurements. Figure 3.5 Longitudinal variability in dissolved CH4 (nM) along the CASS101D stream reach in Sep 2013. Sampling occurred at CASS101D (0 m), 2 upstream locations (+10 m, +20 m), and 2 downstream locations (−20 m, −30 m). Error bars indicate half of the difference between duplicate samples. Figure 3.6 Measurements of δ13C-CH4 values (‰) for sites YOUG105D and YOUG106D across time (Apr 2013–Apr 2014). x Figure 3.7 δ13C-CH4 values (‰) in Oct samples in western MD streams (A) and July samples in Calvert County streams (B), separated by stream setting (forested, wetland, or anthropogenically influenced). The range of δ13C-CH4 values characteristic of biogenic CH4, –110‰ to –50‰ (Whiticar, 1996), is indicated by the green bar. The range characteristic of Marcellus Shale CH4, from –43‰ to –27‰ (Hakala, 2014), is indicated by the gray bar. Figure 3.8 Plot of CH4 concentration (nM) vs. water temperature (°C) for the 2012–2013 and 2013–2014 western MD sampling sets. Figure 3.9 Diagram of Principle Component Analysis (PCA) based on variables CH4 concentration, temperature, and streamflow for western MD samples. Color separates samples by environmental setting. Figure 3.10 Plot of CH4 concentration (nM) vs. streamflow for the 2013–2014 western MD sampling set. Streamflow values (ft3/s) obtained from each sampling trip were taken from the USGS gauge nearest to each site. Figure 3.11 Plot of CH4 concentration vs. Upstream Agricultural Land Use (%) for each site. April, August, and October samples are plotted, with each month including samples of the Sep 2012–Feb 2013 and the Apr 2013–Apr 2014 sample sets. Figure 3.12 Plot of δ13C-CH4 (‰) vs. CH4 concentration (nM) for each site’s highest measured CH4 concentration, apart from site THOM02, for which δ 13C-CH4 was measured in a sample with slightly lower CH4 concentration (July ’14). 1 Chapter 1: Hydraulic fracturing in the U.S. and its impact on the environment 1.1 Introduction Rates of energy use by humans have steadily increased since the Industrial Revolution and continue to do so today (Carpenter, 2014). With high energy demands in the 21st century, novel resources and technologies are being developed and utilized with the intent of sustaining energy demands of the growing human population on a long-term scale. Renewable resources have generated substantial interest and include hydropower, biofuels, geothermal energy, wind, and solar power. These energy sources are ultimately the most sustainable for human energy demands due to their very low greenhouse gas outputs to the atmosphere as well as their long-term energy potential. Solar power offers the largest reserve of energy, although it may be decades before developments occur to place solar energy in a competitive price range with hydrocarbon energy sources (Armor, 2014). It is estimated that currently approximately 11% of total world energy consumption comes from renewable resources, with the projection that the proportion of renewables will reach approximately 15% in 2040 (Letcher, 2014). For a significant emergence of renewable resource use, a combination of technological advances, cost breakthroughs, and legislation will be needed (Armor, 2014). Continued worldwide exploration of hydrocarbon repositories, including oil and natural gas, along with recent advancements in extraction techniques have allowed fossil fuels to remain the mainstay of world energy production and human consumption. 2 Oil, a prominent fossil fuel which is the largest contributor to world energy supplies, constitutes 31.4% of world energy sources (Key World Energy Statistics, 2014). Rates of discovery of oil have been declining since the 1960s (Hoffmann, 2014), however, and production of oil from conventional extraction methods is expected to drop by more than 50% over the next 20 years (Rhodes, 2014). Natural gas contributes 21.3% of world energy supplies and is growing in importance (Key World Energy Statistics, 2014). Between 1990 and 2009, global production of natural gas rose by almost 42%, nearly twice the 22% increase in global oil production over this period. Much of this growth is related to continuing discovery and expansion of gas reserves worldwide, and high natural gas production rates will likely be sustained. It is estimated that, outside of the U.S., less than 11% of natural gas reserves have been recovered (Moniz, 2011). Of particular interest are novel techniques for stimulation of natural gas wells from impermeable deep shale formations, or shale gas drilling, which has vastly increased the availability of natural gas. Recent projections estimate that worldwide accessibility of shale-contained gas increases global recoverable gas resources by more than 40% (Rahm et al., 2013), with indications that more than 169 trillion m3 of shale gas resources are potentially available (Soeder, 2012). Plans for drilling of newly available shale resources are underway in many countries that have access to the deep gas reservoirs, including Poland, Ukraine, Russia, China, and Australia (Malakoff, 2014; Vidic et al., 2013); however, some countries, such as England and France, are presently suspending or banning these drilling methods due to their potential environmental repercussions (Lewis, 2012). 3 The U.S. has been the first location of widespread natural gas extraction from deep shale formations, and is now the world’s leading natural gas producer (Malakoff, 2014). With 29 known shale basins in the U.S. (Fig. 1.1), including the Marcellus Shale spanning 246,000 km2 in the eastern U.S., shale gas is the fastest growing energy sector (Entrekin et al., 2011). Shale gas accounts for approximately 40% of total U.S. natural gas production and is expected to grow to 53% of the country’s gas production by 2040 (Malakoff, 2014). Figure 1.1 Map of shale formations with natural gas access in the lower 48 states. http://www.eia.gov/oil_gas/rpd/shale_gas.pdf In addition to the widespread availability of natural gas, there is also worldwide interest in this energy resource because it is relatively efficient and clean-burning. When considering greenhouse gases like carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O), released during fuel combustion, natural gas emits smaller 4 quantities per megajoule (MJ) of fuel combusted relative to both oil and coal (Burnham et al., 2012). One study has predicted that accelerated shale drilling of natural gas will reduce U.S. electricity sector emissions by 17% (Sovacool, 2014). Also, compared with oil and coal, natural gas combustion emits lower amounts of sulfur oxides, nitrogen oxides, and mercury (Sovacool, 2014). This has been an impetus for increasing utilization of natural gas worldwide. In China, for example, natural gas supplies 4% of total energy demands, and the country plans to increase its proportion of natural gas use to 8% by 2020 (Gregory et al., 2011). 1.2 Oil and natural gas production 1.2.1 Origins of oil and natural gas Oil and natural gas are hydrocarbon fuels, often referred to as fossil fuels, which are formed from organic matter that deposits with lake or marine sediments and becomes buried and compacted over geologic time scales. This process occurs in aquatic areas of high productivity and low circulation, such as lagoons, deltas, or marshes, where high amounts of organic matter accumulate and stagnant conditions allow bottom waters to become anoxic. Settled organic matter may avoid decomposition and become buried as younger sediment layers are subsequently deposited. Buried layers are then subjected to increasingly higher temperatures and pressures which convert organic matter within sediments to hydrocarbon compounds over time (Keller, 1982). With high pressure from the overlying rock as well as high temperature from the earth below, organic matter is first converted to kerogen, the high-molecular-weight 5 organic precursor to oil and gas. The next phase of breakdown of organic matter, in which kerogen undergoes thermal alteration to smaller organic compounds, is referred to as catagenesis and proceeds over thousands to millions of years (Schobert, 2013). When kerogen reaches temperatures of 60–170 °C, typically occurring at depths of 4–6 km, it is thermally broken down into hydrocarbons of oil, which mainly consists of compounds in the C3–C34 size range (Marshak, 2005). With further increases to temperatures of 170–225 °C and depths of approximately 6–9 km, the gas window is reached (Marshak, 2005). Hydrocarbons are broken down to form compounds of natural gas, which include low-molecular-weight compounds in the C1–C7 size range, with methane (CH4) being the predominant product at the hottest end of the gas window (Schobert, 2013). 1.2.2 Drilling of conventional oil and gas reservoirs With exposure to high pressures and temperatures at depth, the organic material in the form of oil or natural gas is also prone to migrate to lower-pressure environments which have increased porosity. The oil and gas tend to migrate upward through the source rock, by way of faults, fractures, and more porous rock units, and upward migration continues until it is impeded by an impermeable overlying rock unit, or cap rock, under which oil and gas then accumulate (Letcher, 2014). These accumulations of “trapped” oil and natural gas create reservoirs which may be accessed by vertical drilling through the impervious cap rock into the reservoir, as illustrated by the image of a conventional gas well in Fig. 1.2. The compressed gas of the reservoir is then transported through the wellbore to the surface (Suarez, 2012). 6 These types of reservoirs have been accessed and drilled this way since the late nineteenth century, and the International Energy Agency (IEA) defines these resources as conventional because they are relatively easy and inexpensive to produce compared to oil and natural gas within deeper formations of low permeability (Letcher, 2014). 1.2.3 Drilling of unconventional oil and gas reservoirs In some environments, sediments deposited with the organic matter are very fine-grained and the rock formed from the compaction of these sediments consequently has very low permeability. Oil and gas formed within this rock are unable to migrate and remain trapped within the formation (Kolb, 2014). If the impermeable rock formation is rich in organic matter and subjected to high temperature and pressure with very deep burial, the organic matter undergoes sufficient alteration over geologic time to form primarily natural gas within the rock (Judson, 1987). One type of low-permeability sedimentary rock is shale, which derives from fine sediments deposited under anaerobic conditions in biologically productive marine environments. Within shale formations there is an abundance of organic matter, and, with exposure to elevated temperatures and pressures, large amounts of natural gas are generated (Soeder, 2012). 7 Figure 1.2 Diagram showing conventional and unconventional drilling techniques for oil and natural gas. Conventional drilling accesses reservoirs of oil and gas trapped in reservoirs below a seal. Unconventional drilling utilizes a lateral wellbore and hydraulic fractures to extract oil and gas from shale formations. http://www.eia.gov/todayinenergy/detail.cfm?id=110 These shale sources of natural gas were previously difficult and uneconomical to access due to shale’s low permeability and also due to the relatively small vertical extent of shale formations, which are typically tens of meters thick, allowing only limited contact with a single conventional well that is drilled vertically (Soeder, 2012). In the 1990s, horizontal drilling techniques were developed, enabling extensive wellbore contact with shale throughout the formation. Around this time, methods were refined for extraction of hydrocarbons from rock by the high-pressure pumping of fluids into a formation, termed hydraulic fracturing. The development of hydraulic fracturing techniques facilitated stimulation of gas from low-permeability shale. These techniques were then combined with horizontal drilling, leading to 8 emergence of the first commercially viable shale gas production wells in North America’s Barnett Shale in the early 21st century (Soeder, 2012). The term hydraulic fracturing, commonly referred to as fracking, is now used to refer to the combination of horizontal drilling and hydraulic fracturing for shale gas extraction. Shale gas accessed by fracking is considered an unconventional hydrocarbon resource because stimulation of gas trapped within shale is more difficult and requires advanced techniques, while conventional methods of gas production typically involve simpler withdrawal from a belowground gas reservoir within a more porous rock (Suarez, 2012). In the typical fracking process for access to unconventional resources, a well is first drilled vertically. Upon reaching the shale, which may be between 600–4,000 m deep depending on the formation, drilling is redirected horizontally for well access within the formation, as is seen in the lateral wellbore diagram in Fig. 1.2 (Kell, 2009). The horizontal portion, called a lateral, may extend 300–3,000 m and multiple laterals may originate from one well, extending in different directions (Dale et al., 2013). Before drilling begins, the lateral is perforated using explosive charges on a wireline, generating holes which allow enhanced contact with the shale reservoir (Soeder, 2012). For gas extraction, hydraulic fluids, containing water and numerous chemical additives, are pumped into the wellbore and then pressurized at 480–850 bar to generate new fractures in the shale and extend existing fractures (Soeder, 2012; Vidic et al., 2013). In the next phase, gas pressure in the rock is used to push the injected fluids back out of the well. Returning fluids are collected and referred to as flowback 9 water during this phase of drilling, which is termed completion flowback and may last up to several days (Allen et al., 2013). Any gas which returns to the surface during the completion flowback phase is sent to a flare, in which the majority of gas compounds are combusted (Soeder, 2012). After the initial period of completion flowback, gas production begins, although some fluid flow to the wellhead continues during the gas production phase, termed produced water (Soeder, 2012; Vidic et al., 2013). In total,10–40% of injected fluids return and are collected at the surface wellhead during the life of the well (Gregory et al., 2011). The recovered fluids, including flowback water and produced water, are collectively referred to as fracking fluids. Unconventional oil reserves, in addition to those of gas, are also found worldwide (Letcher, 2014). Accessibility of oil and gas reserves due to developments of unconventional drilling have significantly changed the global energy market and countries around the world planning for fracking development are looking to the U.S. and Canada, the first to extensively use fracking techniques, to take the lead in environmental research and Best Management Practices with regard to the drilling process (Soeder, 2012). 1.3 Environmental challenges related to fracking 1.3.1 Range of environmental impacts While fracking provides access to significant fuel resources, there are several potential environmental impacts related to fracking operations, some of which are not sufficiently understood. Infrastructure and well construction for this technology have 10 seen rapid growth in many areas of the U.S., and policies and enforcing personnel in many regions of fracking have not been able to expand regulatory activity at a pace to match that of drilling development (Soeder, 2012). Aquatic, terrestrial, and atmospheric systems all may be affected by fracking development and its industrial nature. Compared with conventional drilling wells, well pads for fracking have a larger footprint, drilling time is typically longer, and transport of fluids leads to more extensive road traffic (Kolb, 2014). Environmental disturbances which may occur include light and noise pollution, atmospheric release of CH4, land fragmentation, non-native species introduction, habitat destruction, and degradation of water quality (Soeder, 2012). Although some of these impacts, such as light and noise pollution, may have short-term influence, others, such as land fragmentation and water quality degradation, may be more long-term and lead to a variety of effects on the surrounding ecosystem (Kolb, 2014). Another aspect of fracking that may have long-term consequences is air pollution from various compounds emitted during operations. Heavy machinery is used considerably during construction and drilling, releasing a variety of compounds to the atmosphere including aromatic hydrocarbons benzene, toluene, ethylbenzene, and xylene (BTEX), carbon monoxide (CO), nitrous oxides (NOx), and fine particulate matter PM2.5 (McKenzie et al., 2012). In addition, light alkanes, including propane and butane, and other volatile organic compounds (VOCs) are released from shale formations with gas extraction and may be emitted to the atmosphere during well operations (McKenzie et al., 2012; Petron et al., 2012). 11 Induction of earthquakes and seismic activity is also a potential risk of fracking processes. In 2012 the National Research Council (NRC) reviewed 154 worldwide reports of induced seismicity and found that there was no direct link between fracking and seismic events, but fracking fluid disposal by deep well injection was linked to induced seismicity in some areas (Hitzman, 2012). This finding has been supported by other studies in the U.S. which have linked seismic activity with deep well injection (Kim, 2013; van der Elst et al., 2013). 1.3.2 Impacts of fracking fluids on water resources Surface and groundwater impacts constitute a major environmental concern surrounding fracking. Large quantities of water (between 1–100 million liters) are required for drilling of each well (Clark et al., 2013). Local water supplies, including surface water and groundwater sources, are employed in the process, leading to significant freshwater usage which may be deleterious in regions with stressed water supplies, particularly in periods of drought (Adair, 2012). In addition to water quantity, water quality may also be jeopardized in areas of fracking activity. In recent years there has been increasing public consideration of water quality risks posed by fracking (Gregory et al., 2011). Aqueous fluids used for fracking contain numerous chemicals added prior to well injection; the major classes of additives include proppant material for keeping fractures open, acids which clean the wellbore, biocides for bacterial control, antiscalants for reducing precipitation of salts and metals, and surfactants which reduce surface tension to improve fluid recovery (Vidic et al., 2013). Although classes of chemical additives are known, 12 many specific compounds used in the fluids are considered proprietary and are not reported publicly. Some states, however, do require all fracking fluid compounds to be disclosed to regulatory authorities (Adair, 2012). A 2011 congressional investigation found that, over a 4-year period, a total of 750 chemicals were used in fracking fluids utilized by fourteen leading gas companies (Waxman et al., 2013), and many of these chemicals are not regulated by the U.S. Safe Drinking Water Act (Vidic et al., 2013). Fracking fluids which are recovered at the wellhead after the fracking process typically contain compounds added before injection, degradation products of these compounds, and elements from brines and solutions within the shale and adjacent formations (Haluszczak et al., 2013). Due to interactions with formation solutions, high levels of total dissolved solids (TDS) are found in fracking fluids, with TDS commonly in the range 10,000–180,000 mg/L, while surface waters typically have TDS << 1,000 mg/L (Vengosh et al., 2014). High levels of TDS may directly impair the normal metabolism of fish, amphibians, and aquatic plants (Canedo-Arguelles et al., 2013) and are also linked to the growth of harmful algae (Patino et al., 2014). In domestic and municipal water sources, elevated TDS may cause damage to water treatment equipment, increased water treatment costs, and undesirable drinking water taste (Anning, 2014). Fracking fluids also contain radioactive material derived from the targeted formation in the form of radium (Ra) isotopes, 226Ra and 228Ra. These isotopes have been found in fracking fluids at levels 1,000–3,200 times higher than the EPA Maximum Contaminant Level (MCL) for radium in drinking water (Warner et al., 13 2013a). Also present in fracking fluids are hydrocarbons including benzene, xylene, toluene, and naphthalene originating from the shale formation (Hayes, 2009); in surface and groundwater these have been linked to negative ecological and human health effects (Gross et al., 2013). In addition to the variety of contained compounds, fracking fluids raise concern due to the large quantities which are collected at well sites and also due to the many possible pathways for their introduction to natural waters. During their flow to the surface from depth, fracking fluids may affect groundwater through potential belowground well casing failures or natural or induced fractures (Vidic et al., 2013). Upon their return to the surface, fracking fluids have potential for water contamination during on-site storage, transport, and disposal. Transport of large fluid quantities, required for many disposal methods, presents significant contamination risks (Skalak et al., 2014). Also particularly hazardous is disposal by fluid treatment and discharge, which releases many fracking fluid compounds to surface waters (Warner et al., 2013b). Deep underground injection of fracking fluids is another common disposal method which poses a contamination threat to groundwater in addition to the potential seismic effects (Vengosh et al., 2014). Fracking fluids are not federally regulated, and states in which fracking occurs have addressed management of fluids in a variety of ways. In some states modifications of existing oil and gas regulations have been made for improved management of fracking fluids. In Colorado, requirements have been increased for pits storing hazardous substances, which are now more prevalent due to fracking. Ohio and Pennsylvania, two states with a high prevalence of fracking (Vidic et al., 14 2013), have modified rules for treatment plants which receive fracking fluids (Adair, 2012). In Ohio, received fracking fluids must be below a threshold TDS level, and in Pennsylvania, treatment plant discharge is subject to average monthly TDS standards (Adair, 2012). With a range of synthetic and naturally occurring constituents found in fracking fluids, as well as many opportunities for leakage and contamination, it is important to understand fates of these fluids and to be able to detect them in natural waters. A tool that has been used for this is the analysis of specific constituents which occur at very high levels in fracking fluids, on the order of thousands of mg/L, including chloride (Cl–), bromide (Br–), calcium (Ca2+), sodium (Na+), strontium (Sr2+), and barium (Ba2+) (Haluszczak et al., 2013; Warner et al., 2013a). Although some studies have reported high Cl– and TDS levels in areas where treated fracking fluids are discharged (Olmstead et al., 2013; Wilson and VanBriesen, 2012), Cl– and TDS also may become elevated in streams due to other human activities including suburban development and agriculture (Kaushal et al., 2005). Constituents such as Ba2+, Sr2+, and Br–, which occur at negligible or low levels in surface waters, are the most promising for detecting the presence of fracking fluids (Vidic et al., 2013). 1.3.3 Methane release associated with fracking Due to the high-pressure stimulation of gas, there is potential for mobilizing methane (CH4) from the target shale formation or surrounding formations and this may lead to CH4 contamination of surface or groundwater as well as emission of CH4 to the atmosphere. In domestic water sources, CH4 contamination may be an 15 explosive hazard (Vidic et al., 2013). Methane in surface water environments and in the atmosphere is of concern since CH4 is a powerful greenhouse gas (GHG) with global warming potential 28–32 times greater than that of CO2 on the 100-year time scale (Caulton et al., 2014). Although the predominant mechanisms for CH4 contamination associated with fracking are not known, there are a number of suggested mechanisms through which CH4 movement and contamination may occur with fracking activity (Osborn et al., 2011). One potential mechanism is the displacement of CH4 present in the target shale formation leading to its upward migration into aquifers through faults or fracture networks. These networks are extensive in many shale basin regions, particularly that of the Marcellus Shale, and drilling may generate new or enlarged fractures which increase potential for CH4 migration. In addition, numerous abandoned, uncased, and inadequately plugged wells exist in many areas of fracking development, further increasing systems of cross-formational connectivity (Osborn et al., 2011). Also possible with fracking is that the drilling stimulates upward migration of CH4 from formations above the targeted shale formation (Osborn et al., 2011). Finally, fracking wells with faulty well casings or seals may allow CH4 to leak from within the wellbore and to migrate laterally or vertically to water reservoirs, a phenomenon which has been documented in the conventional oil and gas industry (Vengosh et al., 2014). While there are several pathways by which fracking may lead to CH4 introduction to waters, dissolved CH4 also occurs naturally in surface and groundwater. Methane forms at depth in sedimentary basins by high-temperature 16 maturation of organic matter, and this thermogenic CH4 is able to migrate vertically to aquifers and also migrates laterally at faults and fractures (Etiope and Klusman, 2002). In addition, biogenic CH4 is formed through microbial processes at shallow depths; this CH4 may similarly reach groundwater or surface water through vertical or lateral migration (Vidic et al., 2013). Because CH4 may occur in surface and groundwater due to both natural processes and hydraulic fracturing activity, studies have been unable to clearly link fracking to elevated surface or groundwater CH4 levels (Vidic et al., 2013). Long-term pre-drilling measurements of dissolved CH4 in waters are needed to determine if dissolved CH4 contamination correlates with fracking, and this type of long-term baseline data has not been available in the majority of active fracking areas (Adair, 2012). Management actions have been taken in some states, however, to better understand drilling risks of CH4 contamination and to better protect the water table. Michigan, Pennsylvania, and West Virginia have established Presumptive Liability, which assumes that the well operator is responsible for water contamination in the well vicinity unless there is evidence to the contrary. This incentivizes operators to monitor groundwater CH4 prior to drilling. For reduction of groundwater contamination risks, several states have recently updated regulations to specify that surface casing must be in place to a depth of at least 50 ft below the deepest fresh groundwater sources (Trimble, 2012). Although it is unknown whether there is a link between fracking and CH4 contamination of waters, it is known that gas extraction by fracking releases gaseous 17 CH4 to the atmosphere (Miller et al., 2013), which, as noted above, has climate implications because of the potency of CH4 as a GHG in the atmosphere. Significant amounts of CH4 are emitted at several points during the gas extraction process, including well completion, transport through seals and pipelines, and intentional venting practices (Allen et al., 2013; Eshleman, 2013). The EPA reports natural gas extraction systems to be the top source of anthropogenic CH4 emissions (Hockstad, 2013). Estimates of CH4 emission rates from a fracking well vary between 0.6–7.7% of the produced gas, with some estimates calculated from localized atmospheric sampling at well sites and others determined from aircraft or remote sensing measurements (Allen et al., 2013; Caulton et al., 2014; Miller et al., 2013). Recently it has been shown that if emission rates are greater than 3.2% of production gas, then, overall, natural gas combustion for fuel becomes worse than coal with respect to the global warming potential of released gases (Alvarez et al., 2012). 1.4 Environmental management in the Marcellus Shale region of Maryland Development of fracking wells has proceeded very rapidly in the U.S. over the last decade, particularly in eastern U.S. areas underlain by the Marcellus Shale, a gas-bearing formation of middle-Devonian age (Vengosh et al., 2014). Many states overlying the Marcellus Shale have moved forward quickly with fracking infrastructure and gas production, primarily West Virginia and Pennsylvania, in which a total of 8,000 gas wells have been drilled between 2008 and 2011 alone (Soeder, 2012). The Marcellus Shale also underlies Maryland in the western portion of the state, as is shown in Fig. 1.3, with areas underlain by the Marcellus Shale 18 shown in blue. Maryland is unique, however, in that it has not allowed fracking to date. A 2011 executive order was given for the state, The Marcellus Shale Safe Drilling Initiative, which suspended the licensing of fracking permits in Maryland for 3 years, allowing time for further study of potential environmental and public health effects of fracking (O’Malley and McDonough, 2011). Figure 1.3 Map of western Maryland showing areas underlain by the Marcellus Shale in blue. http://www.wmrcd.org/ The suspension of drilling under this policy initiated a 3-year pre-fracking period in Maryland and, in accordance with recommended management practices, monitoring of baseline conditions is being carried out during this period. Throughout the western Maryland region, the Maryland Department of Natural Resources (MD DNR) and researchers at the University of Maryland Center for Environmental Science (UMCES) have monitored streams at sites located downstream of areas leased for gas production. For the streams of interest, monitoring of stream water chemistry monitoring began in 2011, with continuous measurements of stream 19 temperature and specific conductance collected at all sites. Regular measurements of stream water pH, total suspended solids (TSS), and nutrients were taken at all sites starting in 2012. To assess stream biological conditions, regular monitoring of benthic macroinvertebrates, fish, mussel, and herpetofauna began in 2012. Monitoring of polycyclic aromatic hydrocarbons (PAHs) and surfactant compounds in streams was initiated in 2013 as these compounds are associated with fracking operations and may lead to water quality degradation. From 2012–2014, regular monitoring of dissolved CH4, Sr 2+, Ba2+, and major ions occurred in view of their potential to be displaced and enter surface waters during fracking; the monitoring of the latter four constituents are the topic of this work. The pre-drilling characterization of stream conditions and water chemistry in this region is unique and provides multi-faceted baseline data that were not collected in most U.S. states with fracking development. The pre-drilling stream water data for western Maryland inform resource managers in the state, and it will be most useful if collection of this data continues during drilling, production, and post-drilling periods, as was recommended in Maryland’s Best Management Practices for Marcellus Shale Gas Development (Eshleman, 2013), so that potential changes in stream water quality may be observed over time. These datasets will aid in comprehension and potential mitigation of watershed impacts from fracking. They may also reveal the particular stages of the fracking process which are the most likely to have water quality impacts. 20 1.5 Study of dissolved methane and metals in Maryland surface waters During the pre-fracking period in Maryland, I monitored surface water dissolved CH4, Sr 2+, Ba2+, and major ions for the current study, as these dissolved constituents may be indicators of watershed impacts from fracking. In particular, they may indicate two major types of water quality perturbations: contamination from fracking fluids and CH4 migration in proximity to fracking wells. The western Maryland region examined in this study comprises the area underlain by the Marcellus Shale, shown in Fig. 1.3, and includes the state’s westernmost county, Garrett, and the western portion of adjacent Allegany County. This region is primarily in the Allegheny Plateau physiographic province of the Appalachian mountains, with elevation ranging from 600–900 m (Roth, 1999). It remains largely undeveloped, with forest making up >75% of land cover and agriculture approximately 15% (Eshleman, 2013). There is a history of coal mining with many underground coal mines, both active and abandoned, as well as active and inactive gas wells and a gas storage field covering 138 km2 (Eshleman, 2013). Existing mine and well infrastructure in areas of fracking development may compromise the integrity of drilled wells and may enable belowground migration of CH4 or fracking fluids to aquifers (Vidic et al., 2013). Also prevalent in the region are headwater streams, which warrant careful study before, during, and after drilling because of their strong aquatic–terrestrial linkages and the inherent vulnerability of their fauna (Pond, 2012). These streams contribute to the biological diversity and integrity of downstream systems (Meyer et al., 2007). 21 For monitoring streams in western Maryland in the current study, regular sampling was carried out in 25 stream reaches of varying topography and land use. These sites spanned five catchment areas across the region which are labeled on the map in Fig. 1.4: the Youghiogheny River catchment (green), Casselman River catchment (yellow), Savage River catchment (blue), Georges Creek catchment (pink), and Potomac River Upper North Branch catchment (‘Potomac River U N Branch,’ orange). Other areas of the greater Youghiogheny River basin are indicated in green on the map and areas of the greater Potomac River North Branch basin are shown in orange. Although fracking is not occurring in the western MD region, there is fracking activity just to the north in Pennsylvania as well as in West Virginia which borders the region on the south and west. Activity in West Virginia is notable since Maryland sites in the Youghiogheny River and Potomac River Upper North Branch catchments are located downstream of West Virginia land. The Casselman River, Savage River, and Georges Creek catchments drain Maryland land, with the Casselman and Savage Rivers flowing northward into Pennsylvania. Baseline measurements in the current study thus may not represent ‘pristine’ conditions but represent the current condition of western MD streams prior to MD fracking development. Metals Sr2+ and Ba2+ occur at very high concentrations in fracking fluids as they are transported from in situ shale formations and are highly enriched in the fluids relative to surface waters (Barbot et al., 2013; Watmough, 2014). Strontium and Ba2+ may therefore act as tracers for the presence of fracking fluids and were monitored in 22 Figure 1.4 Map of the western Maryland study region with the study sites represented as purple points. Catchment areas studied are labeled, with the Youghiogheny River catchment shown in green, Casselman River catchment in yellow, Savage River catchment in blue, Georges Creek catchment in pink, and Potomac River Upper North Branch (‘Potomac River U N Branch’) catchment in orange. The wider Youghiogheny River basin is indicated in green, and the wider Potomac River North Branch basin in orange. western Maryland surface waters. Specifically, objectives with this monitoring were: 1. To determine temporal and spatial variability in stream concentrations of Sr2+ and Ba2+ throughout western Maryland across a 1-year period. This allowed determination of baseline concentrations in the region prior to fracking 23 development. With collection of baseline measurements, the applicability of Sr2+ and Ba2+ as fracking fluid tracers was also assessed. Monitoring of Sr2+ and Ba2+ stream concentrations is discussed in Chapter 2 of this work. 2. To determine relative concentrations in Sr2+ and Ba2+ using the Ba/Sr molar ratios in monitored stream waters. The Ba/Sr ratios were of interest as they may indicate different geologic Sr2+ and Ba2+ sources or may indicate relative patterns of Ba2+ and Sr2+ saturation in streams. In addition, stream ratios of Ba/Sr were assessed for their applicability as fracking fluid tracers. Evaluation of Ba/Sr ratios is similarly discussed in Chapter 2. Major ions Na+, K+, Mg2+, Ca2-, Cl–, and 24SO  were measured in all streams for further characterization of stream waters. Nitrate ( 3NO ), bicarbonate ( –3HCO ), and dissolved organic carbon (DOC) were also measured in select sampling months. Goals of the monitoring of major ions were: 1. To determine temporal and spatial variability of major ion concentrations in western Maryland streams monitored across 1 year. Monitoring of major ion levels allowed their assessment as potential fracking fluid tracers. Major ion measurements also enabled calculations of the ionic strength and charge balance of stream waters. Additionally, this monitoring allowed further interpretation of Sr2+ and Ba2+ concentration patterns,and, in particular, facilitated investigation of the potential for oversaturation with respect to BaSO4 or SrSO4. 24 2. To determine major ion ratios in stream waters across both space and time. Major ion ratios in streams were used to determine probable geologic and anthropogenic sources to streamflow. The monitoring of major ion ratios also allowed determination of their viability as fracking fluid tracers. The assessment of major ion ratios is discussed in Chapter 2 of this work. With the potential for upward CH4 migration associated with fracking activity, dissolved CH4 may shift from baseline surface water concentrations following fracking development, thus, in the current study, naturally occurring dissolved CH4 was monitored in western Maryland headwater streams. The objectives of this part of the study were: 1. To determine temporal and spatial variability in stream concentrations of dissolved CH4 over a year-long time span. Assessment of variability across time and space allowed determination of pre-fracking baseline concentrations which have not been previously measured in streams of western MD. With determining spatial and temporal concentration variability, I intended to verify whether CH4 concentrations will be useful indicators of fracking-related stream contamination. Monitoring of dissolved CH4 concentrations is discussed in Chapter 3 of this work. 2. To determine whether stream CH4 is primarily supplied by microbial production or by migration from deep reservoirs of thermogenic CH4. The presence of thermogenic CH4 was of particular interest since thermogenic inputs to streams may be affected by future fracking activity. The major CH4 25 stream sources were evaluated using analyses of the carbon isotopic composition of CH4, since microbially produced CH4 has lower 13C/12C ratios than thermogenic CH4. Carbon isotopic compositions of stream CH4 and their use in determining CH4 sources is presented in Chapter 3. The characteristic stream baseline with respect to CH4, Sr 2+, Ba2+, and major ions indicates expected ranges and variability of these constituents in streams unaffected by MD fracking. It allows for examining shifts in surface water chemistry over time related to fracking fluid contamination or to CH4 migration to surface waters. Monitoring will be valuable for Maryland in the event that fracking is allowed, as potential short- and long-term impacts to surface waters may be observed. The identification of water quality impacts of fracking will also be useful outside of Maryland and will inform drilling practices so that they may be carried out with minimal impacts to the environment. 26 Chapter 2: Baseline monitoring of dissolved metals in western Maryland streams 2.1 Introduction 2.1.1 Hydraulic fracturing fluids as sources of water contamination The use of hydraulic fracturing (fracking) fluids constitutes a major component of the fracking process, with 1–100 million L injected for the fracking process of each well (Clark et al., 2013). Fracking fluids may compromise surface and groundwater quality through a variety of contamination pathways. It is therefore necessary to detect and track the occurrence of fracking fluids in natural waters, for which tracer compounds may be useful. Tracking the fate of fracking fluids and its constituents requires that baseline conditions in natural waters are known. Baseline levels of dissolved strontium (Sr2+) and barium (Ba2+) in western Maryland (MD) streams are determined in this study as these metals may be used as tracers of fracking fluids in the future. In addition, stream water chemistry is assessed with respect to cations sodium (Na+), potassium (K+), magnesium (Mg2+), and calcium (Ca2+); anions chloride (Cl−), sulfate ( 24SO  ), bicarbonate ( 3HCO ), and nitrate ( 3NO ); and dissolved organic carbon (DOC). Some of these ions were assessed for their applicability as fracking fluid tracers. They were also used to characterize stream waters and to determine ionic strength, charge balance, and the tendency for oversaturation with respect to BaSO4 and SrSO4 in streams of this region. 27 2.1.2 Characteristics of fracking fluids and wastes Fracking fluids contain a variety of synthetic compounds added to enhance the drilling process. Although it is not required to disclose fracking fluid compounds in all U.S. states, the range of added compounds may include hydrochloric acid, ammonium persulfate, guar gum, isopropanol, citric acid, glutaraldehyde, ethylene glycol, acrylic polymers, and sand (Vengosh et al., 2014). During the fracking process fluids acquire natural compounds including dissolved solids, radioactive material, and hydrocarbon compounds from brines within the deep shale and adjacent formations (Vengosh et al., 2014). The natural compounds, synthetic compounds, and various degradation products all remain within the fracking fluid stream that is collected as waste at the end of the drilling process (Lester et al., 2015). The level of total dissolved solids (TDS) in recovered fracking fluids, which includes dissolved inorganic and organic constituents, varies widely among shale formations throughout the U.S., with Marcellus Shale drilling fluids typically having TDS of 180,000 mg/L or greater (Vengosh et al., 2014), well above TDS of seawater (35,000 mg/L) (Perlman, 2014). Some drilled shale formations, including the Marcellus Shale, produce fluids that have high levels of toxic metals such as arsenic (As) and selenium (Se) due to the high metal content of the shale (Vengosh et al., 2014). In addition to recovered fluids, the fracking process produces drill cuttings, which consist of drilled rock and clay remnants generated during wellbore drilling. Drill cuttings contain entrained salts, hydrocarbon compounds, and radionuclides. Like fracking fluids, the drill cuttings pose a contamination threat to surface and groundwater and are treated and disposed of off-site (Brantley et al., 2014; Kargbo et al., 2010). 28 Recovered fracking fluids and drill cuttings, as oil and gas industrial waste products, are exempt from U.S. Resource Conservation and Recovery Act (RCRA) hazardous waste provisions including “cradle to grave” surveillance of waste products (Hammer et al., 2012). Additionally, the 2005 Energy Policy Act exempted fracking operations from the Safe Drinking Water Act in the U.S. (Vengosh et al., 2014). State governments are the primary vehicles for regulating and documenting water contamination from fracking fluids and other waste materials (Vengosh et al., 2014). Maryland currently has in place regulations applicable to conventional oil and gas industries but does not have existing regulations and protocols addressing fracking activities and wastes (Eshleman, 2013). 2.1.3 Fates of fracking fluids in surface and groundwater Throughout the life cycle of fracking fluids, many opportunities exist for fluid introduction to surface and groundwater. Groundwater may be affected upon fluid injection into wells. Fracking fluids may leak across failed well casings and migrate through underground systems of faults and fractures (Vidic et al., 2013). Such systems, which occur in the Marcellus Shale region, connect deep formation fluids with shallow aquifers (Llewellyn, 2014; Warner et al., 2012). Even without fracking fluid leakage, high-pressure fluid injections during drilling may stimulate upward migration of deep brines through the pre-existing fault systems (Vengosh et al., 2014). Abandoned oil and gas wells, which are prevalent in many areas of the Marcellus Shale, may also facilitate flow of brines or fracking fluids to shallow aquifers (Jackson et al., 2013b). 29 After fracking fluids are recovered at the wellhead, groundwater contamination may occur in association with spills or leaks of fluids during their storage or transport. Fracking fluids are often disposed of by deep well injection, which may be particularly hazardous to groundwater because of the potential for cement failure or compromised well integrity (Vengosh et al., 2014). Fracking fluid contamination of groundwater has been suggested in some studies. A groundwater study in Colorado showed a link between fluid spill events and elevated levels of benzene, toluene, ethylbenzene, and xylene (BTEX) compounds in groundwater near fracking sites (Gross et al., 2013). Studies in an active fracking area in Wyoming showed elevated levels of specific conductance in groundwater, indicating increased concentrations of dissolved ionic constituents. High levels of ionic constituents as well as elevated groundwater dissolved methane, ethane, and propane found in this area indicated subsurface leakage of fracking fluids, although mechanisms of contamination were not determined (Wright, 2012). In the Barnett Shale region of Texas, a study reported elevated levels of As, Se, Sr, and Ba in groundwater located near fracking wells, with significantly lower concentrations of these constituents found at reference sites outside the fracking area (Fontenot et al., 2013). Contamination of surface water may occur by inflow of contaminated groundwater (Heilweil et al., 2013), or it may occur via surface spills during fracking fluid recovery, storage, transport, or disposal. Recovered fluids are initially stored at the drilling site, which may allow surface water contamination in cases of storage pit breaching, insufficient pit linings, or other leakage incidents (Harkness et al., 2015; 30 Vengosh et al., 2014). After on-site storage, fracking fluids may be treated and reused at the well site or may be transported for off-site treatment and disposal (Shaffer et al., 2013). Long transport distances are often required, particularly for fluids disposed of in specified deep injection wells (Skalak et al., 2014). Long transport distances for large fluid volumes increase the risk of spills and impacts to surface waters. Disposal of fracking fluids at municipal or industrial wastewater treatment plants influences surface waters because treatment plant outflow is discharged to streams and rivers, and wastewater treatment plants do not effectively remove all dissolved solids during treatment (Wilson and VanBriesen, 2012). Signals of fracking fluids have been found downstream of treatment plant discharge, with multiple studies showing elevated concentrations of dissolved constituents including +4NH , Na +, Mg2+, Ca2+, Sr2+, Ba2+, Cl−, Br−, and I− in downstream waters (Ferrar et al., 2013; Harkness et al., 2015). In some regions, fluids are disposed of by land application or road spreading as deicer, and these practices affect surface waters through runoff (Vengosh et al., 2014). Road spreading of conventional drilling waste fluids has been linked to high levels of extractable 226Ra, Sr, Ca, and Na in soils and sediments (Skalak et al., 2014). Illegal disposal of untreated fracking fluids is an additional threat to surface waters and has been documented in some U.S. regions (Vengosh et al., 2014). Surface water intrusion of inorganic ions such as those found in fracking fluids causes significant ecological problems. Increased ion levels interfere with normal osmoregulation and metabolism of aquatic organisms and may compromise functioning of aquatic plant species (Canedo-Arguelles et al., 2013). In addition, high 31 levels of dissolved ions may stimulate blooms of golden algae (Prymnesium parvum) which thrive in brackish environments and produce toxins lethal to fish (Patino et al., 2014). Presence of the ions Br− and I− in surface waters are of particular concern for public health. Even slightly elevated concentrations of Br− and I− may lead to the formation of carcinogenic brominated and iodinated disinfection byproducts (DBPs) during drinking water chlorination or ozonation processes (Harkness et al., 2015; Warner et al., 2013a). Simulations show that surface waters receiving 0.01% fracking fluids by volume generate increased amounts of brominated and iodinated DBPs in drinking water (Parker et al., 2014). The contaminant burden of fracking fluids is amplified in many regions because surface waters may already receive significant contaminant loads from sources including acid mine drainage, road salts, coal ash impoundment fluids, brines from abandoned oil wells, septic system effluent, and agricultural runoff. In such regions, surface water inputs of dissolved ions from fracking fluids may not be as readily diluted to safe levels (Ferrar et al., 2013). A recent study examining synergistic water quality impacts from mixtures of acid mine drainage and fracking fluids showed that the blended fluids were lower in metals and radionuclides; however, the mixtures formed precipitates that were high in radioactivity (Kondash et al., 2014). Fracking fluid presence is also a major concern in surface waters because of exposure to naturally occurring radioactive material (NORM) derived from drilled shale formations (Vengosh et al., 2014). Discharge of treated fracking fluids to Pennsylvania streams has been linked to accumulation of radium isotopes 226Ra and 32 228Ra in sediments, with radioactivity levels approximately 200 times greater than those of background sediment samples (Warner et al., 2013a). The presence of Ra in freshwater sediments may allow significant Ra bioaccumulation in benthic organisms (Warner et al., 2013a). Impacts of fracking fluids may be especially pronounced in headwater stream systems. Well pads for fracking sites are commonly in close proximity to headwater stream systems, allowing direct impacts from spills and other incidents (Brantley et al., 2014). Further, forest and vegetation are cleared for well construction and drilling, effectively removing buffers which normally protect high-quality headwater streams (Trexler et al., 2014). Small headwater streams, which are characterized by relatively low flow volumes, are innately sensitive to contaminants and thus highly vulnerable to water quality changes during fracking development. 2.1.4 Use of metals as fracking fluid tracers There are clearly numerous pathways for contamination of natural waters by fracking fluids as well as many serious ramifications of such contamination. It is therefore essential to understand the environmental fate of fracking fluids, particularly in headwater streams where their impacts may be greatest. Towards this objective, tracer compounds or elements are expected to be useful as indicators of fluid intrusion of surface waters (Mair et al., 2012). Tracers for identification of fracking fluids should be compounds that are known to occur in high concentrations in the fluids and found at very low concentrations in natural waters. A tracer compound should have relatively low variability in freshwater and should have no 33 major inputs from common contaminants or anthropogenic effluents, allowing the tracer to be definitively linked to fracking fluids. In addition, a tracer compound should be conducive to regular monitoring. Several classes of synthetic organic compounds, such as surfactants, biocides, and scale-inhibiting species, are associated with fracking fluids and may signal fracking fluid presence since they otherwise do not occur in natural waters (Vengosh et al., 2014). Such organic compounds are difficult to use as tracers, though, since individual compounds in fluids are not disclosed and may vary widely among well operators (Arnaud, 2015). The degradation products of the compounds also vary depending on conditions and are not well known (Lester et al., 2015). Further, the organic compounds found in fluids are generally expensive to analyze and may require large sample volumes, making regular monitoring less practical. Several major ions are elevated in fracking fluids and would be observed in surface waters in cases of fracking fluid contamination, including Na+, Mg2+, Ca2+, and Cl− (Barbot et al., 2013; Warner et al., 2013a). However, detection of these ions alone would not be distinctive of fracking fluids since such ions are supplied to surface waters from many different anthropogenic and geologic sources (Godwin et al., 2003; Likens et al., 1967). Still, a cumulative measure of total major ions, assessed as stream water specific conductance, can show spikes immediately following fracking fluid contamination (Brantley et al., 2014). Since specific conductance can be continuously recorded in streams by in situ data loggers (Stranko et al., 2013), such a tool would allow continuous observation and prompt detection of spikes signaling possible fracking fluid contamination. The specific conductance 34 measurements signaling potential contamination can be immediately followed by stream measurements of a tracer that is unique to fracking fluids. The ions Sr2+, Ba2+, and Br− occur at very high levels in fracking fluids (Barbot et al., 2013) and are expected to be distinctive tracers of the fluids in streams. Concentrations of these elements are very low in natural waters relative to concentrations in fracking fluids. Based on reported Sr2+ and Ba2+ concentrations in fracking fluids and in streams (Brantley et al., 2014; Watmough, 2014), fracking fluids are expected to be enriched in Sr2+ and Ba2+ by a factor between 37,000 and 560,000 relative to streams. In low-order streams Br− is not detected or detected at very low levels (Rai and Iqbal, 2015). The mere presence of Br− therefore may be indicative of fracking fluid contamination. Regular monitoring on the order of weeks to months is feasible for these three constituents, with monitoring frequencies of once per 1–2 months assessed in the current study. In this study the 1–2 month sampling frequency was utilized for observation of natural concentration ranges in the stream constituents. With the potential for fracking fluid contamination events, however, a higher frequency monitoring tool, such as a field sensor of specific conductance, may be utilized for initial alert of stream contamination. Of the three elemental tracers Sr2+, Ba2+, and Br− which would distinguish fracking fluid presence, Sr2+ and Ba2+ were favored over Br− in the current western MD study since they could be regularly analyzed at the Chesapeake Biological Laboratory. 35 2.1.5 Monitoring of Sr2+ and Ba2+ in western and southern MD streams In order to use Sr2+ and Ba2+ as indicators of fracking fluid presence, it is important to understand their natural concentrations and ranges in aquatic environments. Dissolved Sr2+ and Ba2+ in freshwater are most commonly derived from weathering of compounds containing Sr and Ba in the mineral lattice, primarily carbonate rocks and weathered clay minerals, (Kravchenko et al., 2014; Negrel et al., 1993), both of which are abundant in western MD (Roth, 1999). Groundwater interacts considerably with these minerals and is enriched in Sr2+ and Ba2+, acting as a source of Sr2+ and Ba2+ to overlying freshwater (Peek and Clementz, 2012). Surface runoff may acquire Sr2+ and Ba2+ through interactions with soil minerals, providing another freshwater source of these elements (Hogan and Blum, 2003). Strontium is more abundant than Ba in crustal minerals (Turekian and Wedepohl, 1961), and its salts are also more soluble than those of Ba (Peek and Clementz, 2012). In some active fracking regions in Pennsylvania, surface water Sr2+ and Ba2+ concentrations have been measured and show high spatial and temporal variability (Brantley et al., 2014). This demonstrates the need for baseline measurements of Sr2+ and Ba2+ which can be compared with concentration measurements taken during and after fracking activity. Natural variability in surface water Sr2+ and Ba2+ in the region before the onset of fracking may be accounted for with baseline measurements at 1–2 month intervals in surface waters draining each potential well site, as was assessed in the current study. With the suspension of permitting for fracking in MD under the Marcellus Shale Safe Drilling Initiative, an initial 3-year pre-drilling period occurred from 2011 to 36 2014, allowing for monitoring of dissolved Sr2+ and Ba2+ in the prospective drilling region of western MD, the topic of the current work. The first goal of this work was to determine both temporal and spatial variability of stream Sr2+ and Ba2+ concentrations. This allowed for characterization of baseline Sr2+ and Ba2+ concentrations prior to fracking activity. In addition, assessment of the spatial and temporal variability was used to determine the applicability of Sr2+ and Ba2+ as stream tracers for fracking fluids. It was hypothesized that Sr2+ and Ba2+ temporal patterns are dependent upon relative inputs of surface and groundwater to a stream reach. Stream reaches which receive Sr2+ and Ba2+ from groundwater and which are strongly influenced by surface runoff were expected to show high concentrations of the metals in late summer and fall with lower concentrations in spring when runoff is greatest (Ahearn et al., 2004). Stream reaches which are predominantly groundwater fed with low influence from runoff were expected to show relatively low temporal variability. It was hypothesized that the spatial distribution of Sr2+ and Ba2+ concentrations follows patterns of bedrock geology which are highly variable in western MD (Brezinski, 2013). Despite the anticipated variability in concentrations, concentrations of these elements were expected to be very low relative to concentrations in fracking fluids based on previous Sr2+ and Ba2+ measurements in minimally disturbed streams of North America, as reported in the literature (Elias et al., 1982; Watmough, 2014); Sr2+ and Ba2+ were thus projected to be good fracking fluid tracers in streams. To evaluate naturally occurring temporal variability in stream Sr2+ and Ba2+, regular concentration measurements were carried out at 4–8 week intervals over a 1-year 37 period. Baseline spatial variability in Sr2+ and Ba2+ was assessed through regular concentration measurements at 25 sites across western MD. Another goal of this work was to assess naturally occurring stream ratios of Ba/Sr across time and space. These ratios, determined from Sr2+ and Ba2+ concentration measurements, allowed characterization of stream baselines with respect to the Ba/Sr ratio. Monitoring of Ba/Sr and its variability over time and space also allowed assessment of the Ba/Sr ratio as a potential tracer of fracking fluid. Finally, Ba/Sr ratios were used to interpret patterns observed in Sr2+ and Ba2 concentrations across space and time. It was hypothesized that stream ratios of Ba/Sr vary based on underlying geology with which groundwater interacts, surface runoff interactions with soils, and extent of BaSO4 and SrSO4 precipitation. Ratios of Ba/Sr were assessed from Sr2+ and Ba2+ concentrations measured at 4–8 week sampling frequencies over 1 year, with measurements across the 25 western MD stream sites. 2.1.6 Characterizing waters by major ion composition In addition to baseline Sr2+ and Ba2+ concentrations, major ion concentrations were also of interest. A primary goal in this study then was to determine temporal and spatial variability in stream concentrations of major ions Na+, K+, Mg2+, Ca2+, Cl−, and 24SO  . The monitoring of major ion concentrations enabled assessment of major ions as potential tracers of fracking fluids. The major ion measurements were also used to determine the ionic strength and relative charge contributions in stream waters. Additionally, major ion measurements allowed interpretation of Sr2+ and Ba2+ levels in streams since precipitation and dissolution of Sr2+ and Ba2+ salts are dictated 38 by the presence of certain ions; for example, Ba2+ concentrations in natural waters are affected by 24SO  levels because of the tendency for barite (BaSO4) formation (Kravchenko et al., 2014). Because of such effects of the major ions, observed temporal and spatial trends in Sr2+ and Ba2+ were compared with those of major cation and anion concentrations. Acid neutralizing capacity (ANC), or the proton deficiency of a water sample which can be titrated to an equivalence point by a strong acid, was measured in some stream waters in this study. ANC is primarily controlled by 3HCO -content in river waters and was analyzed in stream waters for comparison with the unbalanced positive charge. Dissolved 3NO and DOC were measured in some stream samples for assessment of their anionic contributions. With measuring of major ions, an additional goal of this study was to assess stream major ion ratios over time and space. Major ion ratios are useful for detecting various natural and anthropogenic impacts to streams; they have been used to indicate the presence in streams of migrated geologic brines, road salt effluent, abandoned mine drainage, conventional oil and gas production fluids, coal power plant wastewater, septic effluent, and agricultural wastewater (Dresel and Rose, 2010; Lautz et al., 2014; Wilson, 2014). Relative ion concentrations were compared across sites to determine similarities or differences among stream waters prior to fracking development. Also, major ion ratios were assessed for their viability as fracking fluid tracers. As with stream Sr2+ and Ba2+ concentrations, it was hypothesized that absolute and relative major ion concentrations vary largely according to bedrock geology 39 underlying stream catchments. Seasonal influences were predicted to affect temporal variability of ion concentrations and ion ratios, with road salt effluent expected to be a major seasonal influence. Road salt effluent was expected to cause increases in stream levels of Na+ and Cl− during winter, as well as slight increases in Mg2+ and Ca2+. With major ions supplied to streams from several potentially variable sources, major ions as measured from grab sampling were not expected to be useful fracking fluid tracers. Spatial and temporal variability of major ions and major ion ratios were assessed through regular measurements of major cations and anions in streams throughout western MD, with sampling at 4–8 week time points over a 1-year period. In addition to analysis of Sr2+ and Ba2+ in western MD streams, I compared stream concentrations in western MD with those of southern MD’s coastal plain setting which is not underlain by the Marcellus Shale formation, with analysis of stream water Sr2+ and Ba2+ in Calvert County in June and July 2014. Along with Sr2+ and Ba2+, major ion concentrations were assessed in Calvert County. Major ion measurements in Calvert County may further elucidate contrasts in major stream water influences between western and southern MD. Monitoring of spatial and temporal ranges of Sr2+, Ba2+, and major ions in streams allows for characterization of baseline water chemistry, which is particularly important for a region such as western MD that may be impacted by fracking in the future. These constituents and their relative concentrations are useful for categorizing stream water influences both currently and in the future following fracking development. 40 2.2 Materials and methods 2.2.1 Sampling sites Samples for metals analysis were collected from first- to fourth-order stream reaches. As defined by Strahler (1952), first-order streams are perennial headwater streams with no tributaries, second-order streams are those formed from the joining of two first-order streams, third-order streams form from the joining of two second-order streams, and so on. Stream reaches sampled were downstream of potential fracking well pad locations in Garrett and Allegany Counties of western MD. In this region of MD, approximately 76% of land cover is forest. Agriculture and urban land use comprise approximately 16% and 3% of the region, respectively (Eshleman, 2013). Sampling occurred in 5 catchment areas which are indicated by color in Fig 2.1: Casselman River, Georges Creek, Savage River, Potomac River Upper North Branch, and Youghiogheny River. Although the western MD study region does not presently have fracking activity, some sites within the Potomac River Upper North Branch and Youghiogheny River catchments may currently be influenced by fracking as they are downstream of West Virginia land where there is fracking activity. Fracking also occurs in Pennsylvania, although Pennsylvania land is downstream of all study sites. Also monitored within the current study were the Savage River Pumping Station (SRPS), a groundwater reservoir within the Savage River catchment, and the Frostburg Water Treatment Plant (WTP), which provides drinking water for the city of Frostburg (Fig. 2.1). The WTP combines waters received from SRPS and from Frostburg’s Piney Creek Reservoir. The treatment plant intake water monitored in this work is referred to as ‘WTP Intake,’ and the finished water from the plant which 41 was monitored is referred to as ‘WTP Finished’. The Piney Creek Reservoir, located in the Casselman River catchment, was regularly sampled in this study, along with several of its minor tributaries, and it is referred to as ‘Reservoir’. In southern MD, stream samples for analysis of metals were collected from streams in the two major catchments of Calvert County: West Chesapeake Bay and Patuxent River Lower, both of which are delineated by background color on the map in Fig. 2.2. The Calvert County region is approximately 46% forest, 38% developed land, and 14% agricultural land (Maryland Dept of Planning, 2010). Samples for analysis of dissolved metals were collected in three sets: two in western MD and one in southern MD’s Calvert County. The first set consists of six sites within Allegany and Garrett Counties which were sampled monthly during the period Feb 2012–Feb 2013. The second set includes fourteen sites in Garrett County which were sampled every 6–8 weeks in the period Apr 2013–Apr 2014. Samples within this set were collected with assistance from the Maryland Department of Natural Resources (MD DNR), who obtained landowner permission for site access, as necessary. Also within the second set are five sites which were sampled biannually (once each in winter and summer): PRUN301D, SAVA302D, YOUG201D, YOUG301D, and YOUG302D. All sites from the first and second sets are shown on the western MD map in Fig. 2.1. Apart from WTP Intake, WTP Finished, and SRPS, I designated all western MD sites as forested, wetland, or anthropogenically influenced, as indicated in Table 2.1, based on the observed surrounding environment of the site. Forested stream sites are characterized by significant tree and vegetation cover surrounding the stream. Wetland sites are characterized by inundated soils 42 surrounding the stream with vegetation and fewer tree stands than found at forested sites. Streamflow is relatively slow at wetland sites. Anthropogenically influenced sites are found in the immediate vicinity of anthropogenic developments including major roads, bridges, residential areas, and coal facilities. Figure 2.1 Map of western MD stream sites for the Feb 2012–Feb 2013 and Apr 2013–Apr 2014 sampling sets. Background and symbol colors indicate catchment, with yellow indicating the Casselman River catchment, blue indicating the Savage River catchment, pink the Georges Creek catchment, orange the Potomac River Upper North Branch catchment, and green the Youghiogheny River catchment. Black symbol indicates the Frostburg WTP. The greater Youghiogheny River basin is indicated in green and the greater Potomac River North Branch in orange. The third set of samples collected in this study includes seven sites in Calvert County which were sampled in June and July of 2014. These are shown on the map 43 in Fig. 2.2, with symbol color indicating the sampling stream and background color indicating catchment. Sites from all sampling sets are listed by site name in Table 2.1 with the corresponding stream name, catchment, geographic coordinates of the reference coordinate system WGS 84, and the site environment. Figure 2.2 Map of stream sites sampled in Calvert County, MD, in June and July 2014. Background and symbol colors correspond to stream catchment, with blue indicating the Patuxent River Lower catchment and purple indicating the West Chesapeake Bay catchment. Table 2.1 Names and locations of sampling sites in western MD for the Feb 2012–Feb 2013 and Apr 2013–Apr 2014 sets and sampling sites in Calvert County from the Summer 2014 set. Site name SPRS represents the Savage River Pumping Station and Reservoir represents the Piney Creek Reservoir. WTP Intake and WTP Finished represent the intake of the Frostburg City water treatment plant and the treated water from the plant, respectively. Sites in the Apr 2013–Apr 2014 set are named according to catchment, where CASS indicates the Casselman River catchment, GEOR the Georges Creek catchment, PRUN the Potomac River Upper North Branch catchment, SAVA the Savage River catchment, and YOUG the Youghiogheny River catchment. Sites in this set are numbered based on stream order, with numbers in the 100s representing 1st order streams, those in the 200s representing 2nd order, etc. Designation ‘S’ indicates long-term monitoring sites which have historical data sets, and those with ‘D’ are newly established monitoring sites. Summer 2014 site names are associated with the stream where the site is found, with HELL representing Hellen Creek, GRAY representing Grays Creek, PARK representing Parkers Creek, and THOM representing Thomas Branch. Note that for site THOM01, June and July sampling locations are slightly different due to higher water levels in July at Thomas Branch. The characteristic site environment is given in the ‘Environment’ column. Site Name Stream Catchment Area Latitude (N) Longitude (W) Environment Feb 2012–Feb 2013 SRPS Savage River Savage River 39° 40' 17.88" 78° 58' 14.94" Reservoir Piney Creek Casselman River 39° 42' 15.00" 79° 0' 40.80" Anthropogenic Tributary T1 Piney River Upper Tributary Casselman River 39° 42' 52.62" 78° 59' 52.08" Forested Tributary C6 Piney Run Creek Casselman River 39° 42' 55.38" 78° 59' 44.10" Wetland Tributary B7 Geatz Run Casselman River 39° 42' 45.48" 78° 59' 35.22" Wetland Tributary G8 Blandy Run Casselman River 39° 42' 21.6" 79° 0' 7.62" Wetland WTP Intake Reservoir/SRPS Casselman/Savage 39° 40' 19.55" 78° 56' 25.07" WTP Finished Reservoir/SRPS Casselman/Savage 39° 40' 19.55" 78° 56' 25.07" Apr 2013–Apr 2014 CASS101D Piney Creek Upper Tributary Casselman River 39° 42' 47.10" 79° 0' 40.70" Forested CASS103D Twomile Run Casselman River 39° 40' 33.10" 79° 3' 43.20" Anthropogenic CASS105D Meadow Run Casselman River 39° 41' 13.40" 79° 6' 0.60" Anthropogenic GEOR101D Mill Run Georges Creek 39° 32' 53.10" 79° 3' 56.70" Forested Table 2.1 (cont.) Site Name Stream Catchment Area Latitude (N) Longitude(W) Environment PRUN301D Nydegger Run Potomac River Upper North 39° 17' 51.10" 79° 21' 0.60" Anthropogenic SAVA101D Savage River Upper Tributary Savage River 39° 38' 45.6" 79° 1' 1.30" Forested SAVA202D Big Run Savage River 39° 34' 27.40" 79° 9' 35.40" Forested SAVA204S Crabtree Creek Savage River 39° 30' 13.10" 79° 9' 20.38" Forested SAVA301D Savage River Savage River 39° 38' 30.90" 79° 1' 18.40" Forested SAVA302D Savage River Savage River 39° 38' 35.50" 79° 1' 12.10" Anthropogenic YOUG102D Youghiogheny River Upper Trib. Youghiogheny River 39° 41' 48.50" 79° 22' 51.20" Forested YOUG103D Youghiogheny River Upper Trib. Youghiogheny River 39° 41' 48.40" 79° 22' 56.30" Forested YOUG104D Buffalo Run Upper Tributary Youghiogheny River 39° 41' 10.70" 79° 24' 36.50" Wetland YOUG105D Laurel Run Youghiogheny River 39° 22' 57.40" 79° 29' 28.60" Anthropogenic YOUG106D Salt Block Run Youghiogheny River 39° 33' 55.84" 79° 28' 6.38" Wetland YOUG201D Mill Run Youghiogheny River 39° 42' 37.8" 79° 22' 11.60" Anthropogenic YOUG301D Buffalo Run Youghiogheny River 39° 41' 20.10" 79° 25' 17.00" Forested YOUG302D Cherry Creek Youghiogheny River 39° 22' 6.90" 79° 27' 14.90" Anthropogenic YOUG432S Bear Creek Youghiogheny River 39° 38' 33.07" 79° 16' 47.28" Forested Summer 2014 HELL01 Hellen Creek Patuxent River Lower 38° 22' 37.20" 76° 27' 35.82" Anthropogenic GRAY01 Grays Creek West Chesapeake Bay 38° 24' 6.96" 76° 24' 39.42" Wetland GRAY02 Grays Creek West Chesapeake Bay 38° 23' 45.12" 76° 25' 27.54" Forested PARK01 Parkers Creek West Chesapeake Bay 38° 31' 16.32" 76° 34' 18.48" Anthropogenic PARK02 Parkers Creek West Chesapeake Bay 38° 31' 56.40" 76° 32' 30.84" Forested THOM01 June Thomas Branch West Chesapeake Bay 38° 24' 13.20" 76° 24' 55.02" Wetland THOM01 July Thomas Branch West Chesapeake Bay 38° 24' 13.74" 76° 24' 55.74" Wetland THOM02 Thomas Branch West Chesapeake Bay 38° 24' 0.78" 76° 25' 21.66" Forested 46 2.2.2 Field sampling All field sampling was carried out when streams were at base flow stage, defined as normal flow conditions for the given season, with avoidance of sampling during high-flow conditions and following storm events. Prior to field sample collection, 60-mL Teflon sampling bottles were acid cleaned by immersion in sub-boiling 8 M HNO3 for 24 hours and subsequently rinsed with Milli-Q water (18.2 MΩ cm). Bottles were transported to and from the field inside trace metal clean Ziploc® bags. Water samples were collected from the stream bank by hand at each site, except at site THOM01, where samples were collected from the wooden footbridge which crosses the stream. In June 2014, the sample was collected from the footbridge in the middle of the stream, and in July 2014, due to higher water levels, sample collection from the footbridge took place close to the north bank of the stream (Table 2.1). Because of the proximity of Calvert County surface waters to the Chesapeake Bay, salinity at each Calvert County site was tested with a refractometer and found to be <1 everywhere. For sample collection, a plastic 60-mL Luer-lok syringe (Becton Dickinson) was used to collect stream water from mid-depth at each stream site. The syringe was first rinsed three times with 60 mL stream water. Following the rinses, 60 mL of water was taken into the syringe and a 0.2-µm filter (Corning) was attached to the syringe tip. The filter was rinsed by expelling 10 mL water through the syringe filter as waste. The Teflon sample bottle was rinsed three times with 10 mL of filtered stream water and was then filled to the bottle neck. After being transported to the laboratory, each sample was acidified with 50 µL concentrated HNO3 (TraceMetal Grade, Fisher 47 Scientific) to keep metals in solution. Acidified samples were allowed to equilibrate for 24 hours before analysis. Stream pH values in Apr 2013 were determined from measurements by MD DNR, where water samples were collected in polyethylene syringes without exposure to the atmosphere. Upon returning to the laboratory, samples were analyzed using a combination pH electrode and Orion 611 pH meter. Stream pH values of the Sep 2012–2013 period were measured with an in situ field meter (values provided by Tom Kozikowski, Mountain Ridge High School, Frostburg, MD). 2.2.3 Analysis of Sr2+ and Ba2+ All preparation of samples and standards for elemental analyses was conducted in a class-100 laminar flow bench. Solutions were prepared with Milli-Q water (18.2 MΩ cm) from a Millipore Direct-Q 3UV purification system. For analysis by inductively coupled plasma mass spectrometry (ICP-MS), an aliquot of acidified field sample was diluted 1 in 10 in 1% HNO3 and was spiked with an internal standard solution with 2 µg/L each of In and Cs. Diluted samples were then analyzed for Sr and Ba concentrations on an Agilent 7500cx ICP-MS, with concentrations measured against an external 5-point calibration line which was 0, 1, 2, 5, and 10 µg/L in both Sr and Ba. For Sr, masses 86 and 88 were measured, which were normalized to 115In. For Ba, masses 135, 137, and 138 were measured, which were normalized to 133Cs. For each element, measurements of separate isotopes were averaged to determine the element concentration. The set of diluted samples was analyzed by ICP-MS twice in random order. After each sample injection, the autosampler probe and sample 48 introduction system were rinsed by a 10 s aspiration of Milli-Q water followed by a 30 s aspiration of 1% HNO3 rinse solution. After running the sample batch, stream samples were prepared in a second round of dilutions by diluting sample aliquots 1 in 20 in 1% HNO3. Three samples of high elemental concentrations from Aug 2013 were prepared with a 1 in 100 dilution in 1% HNO3. The second round of sample dilutions was analyzed by the same method on the ICP-MS. The average of concentration measurements of the two dilution rounds, each of which included two ICP-MS runs, was then used to determine sample Sr and Ba concentrations. Relative precision based on repeated measurements of samples was <1% for most samples. 2.2.4 Analysis of major cations Analysis of major cations Na+, K+, Mg2+, and Ca2+ was carried out using a Perkin Elmer Optima 8300 Inductively Coupled Plasma Atomic Emission Spectrometer (ICP-AES) in Chesapeake Biological Laboratory’s Nutrient Analytical Services Laboratory (NASL). Five-mL aliquots of each acidified sample were used for ICP-AES analysis. Cations Na+, Mg2+, and Ca2+ were measured at two different emission wavelengths for each sample: Na+ was measured in the visible spectrum at wavelengths 589.0 and 589.6 nm, Mg2+ in the UV spectrum at wavelengths 279.1 and 285.2 nm, and Ca2+ at wavelength 317.9 nm in the UV spectrum and at 422.7 nm in the visible spectrum. Potassium was measured at one visible wavelength, 766.5 nm, for each sample. For elements measured at two wavelengths, the measurements were averaged. Errors in each of the two wavelength measurements were propagated in the 49 average. All Na+ and K+ wavelengths were measured in the radial view and all Ca2+ and Mg2+ wavelengths measured in the axial view. Cation concentrations were measured against a 5-point external calibration line, which was 0, 0.2, 0.4, 2, and 4 mg/L in Mg2+ and K+ and 0, 1, 2, 10, and 20 mg/L in Ca2+ and Na+. Calibration standards were made from a multi-element stock solution containing Na+, K+, Mg2+, and Ca2+ (SPEX CertiPrep). Samples whose concentrations exceeded the highest calibration standard were diluted in Milli-Q water for re-analysis, and cation concentrations of these samples were taken from the average of measurements of non-diluted and diluted runs. River water SLRS-5 SRM was analyzed for concentrations with each ICP-AES run. Concentration measurements agreed within 1.1%, 2.2%, 2.6%, and 0.9% of certified SRM concentrations of Na+, K+, Mg2+, and Ca2+, respectively. River water SRM was analyzed 27 times in total, and overall the concentration measurements gave relative precision of 1.3%, 2.7%, 1.2%, and 1.3% for Na+, K+, Mg2+, and Ca2+, respectively. Relative error of cation concentration measurements was <5% for most samples. 2.2.5 Analysis of major anions For analysis of Cl− and 24SO  concentrations, samples were run on a Dionex ICS-3000 Reagent-Free Ion Chromatography system with AS40 autosampler. An IonPac AG22 (4×50 mm) guard column and an ERS 500 suppressor with suppressor current of 50 mA were used in the system. As eluent, a carbonate/bicarbonate buffer was used, made fresh before each run from sodium carbonate and sodium bicarbonate (J.T. Baker). Buffer was made with deionized water and was allowed to degas by 50 sparging with He before IC runs. In the IC system the eluent flow rate was 1.2 mL/min. Six-mL aliquots of each acidified sample were used for analysis, and anion concentrations were measured against a 6-point calibration curve with concentrations 0, 0.8, 2, 4, 8, and 20 mg/L in 24SO  and concentrations 0, 2, 5, 10, 20, and 50 mg/L in Cl−. The calibration standard solutions were made using a stock solution which was 1,000 mg/L in Cl− and a stock solution which was 1,000 mg/L in 24SO  (VHG Labs). Calibration standards were acidified with HNO3 to match the sample matrix in order to account for effects of HNO3 on 24SO  sensitivity in the system. This precluded separate analysis of nitrate concentrations. The calibration curves for Cl− and 24SO  were fitted with quadratic functions. Samples whose concentrations exceeded the highest calibration standard were diluted in deionized, acidified water for re-analysis by IC and final concentrations were determined from averages of diluted and non-diluted sample runs. Verification standards of 6.0 mg/L and 35.0 mg/L Cl– were made from a separate lot of 1,000 mg/L stock Cl– solution (VHG Labs) and, upon analysis, verification standards were found to be within 2.5% of the nominal concentrations. The two verification standards for 24SO  , similarly made from a separate lot of 1,000 mg/L stock 24SO  solution (VHG Labs), were 6.0 mg/L and 35.0 mg/L in 24SO  . Upon analysis, these were found to be within 0.9% of the nominal concentration. The SLRS-5 SRM was not used for verification of Cl– and 24SO  concentrations since it is not certified for these ions. To determine reproducibility, SLRS-5 was analyzed for Cl− and 24SO  with every IC run and gave 51 relative precisions of 1.6% and 0.8% for Cl− and 24SO  , respectively, for a total of 33 measurements each. 2.2.6 Analysis of Acid Neutralizing Capacity During the Oct 2013 sampling in western MD, samples were collected for analysis of ANC to assess the levels of bicarbonate ( –3HCO ) anions in stream samples. Samples were collected in 200-mL polyethylene bottles which had been previously washed in deionized water. At each site, bottles were rinsed three times with stream water. A sample of water from mid-depth of the stream was then collected in the bottle, filling it to the rim without headspace in order to avoid gas exchange. Sample bottles were stored in a cooler during transport to the laboratory and were refrigerated until ANC analysis. Analysis of ANC was carried out by Gran Titration at University of Maryland Center for Environmental Science’s Appalachian Laboratory in Frostburg, MD. For each sample, 50 mL was poured into a vial and 300 µL of 2 M KCl was added to each to increase the ionic strength. While the sample was continuously stirred, a pH electrode was immersed in the sample and initial pH measured. The 0.01 M HCl titrant was then used to lower pH to approx. 4.7. Next, titrant was added in 8 increments to lower sample pH from 4.7 to 3.5. After each titrant addition, volume of cumulative acid added and pH were recorded as data pairs. After titration of the sample to pH 3.5 and recording of titration data, the pH electrode was rinsed thoroughly in deionized water before analysis of the next sample. 52 For each titrated sample, the Gran function F1a was determined for each data pair of cumulative acid volume Va (mL) and pH values. F1a was determined by: ])[HV(VF as1a  (2.1) where Vs = total initial sample volume (mL). The F1a value was then plotted against volume titrant added, Va, to obtain a curve expressed by the linear equation 1a aF b aV  (2.2) The equivalence volume V1 of the titration (mL) was determined as the point of interception with the Va axis, calculated by 1 bV a  (2.3) The ANC (µeq/L) of the sample was then calculated from the equivalence volume by: 61 a s V CANC = ×10V (2.4) where Ca = concentration of acid titrant (M). 2.2.7 Analysis of Dissolved Organic Carbon During April 2014 sampling, western MD samples were collected for measurement of DOC in order to assess anionic contributions of DOC in stream waters. Prior to sampling, 30 mL polyethylene vials were immersed in 10% HCl for 12 hours and were rinsed with deionized water and dried. In the field, vials were 53 rinsed three times with stream water and then filled with water collected from mid-depth at each stream site. For laboratory analysis, DOC was measured on a Shimadazu TOC-5000 in NASL at Chesapeake Biological Laboratory using a high-temperature combustion method (Sugimura and Suzuki, 1988). Samples were treated with HCl and then sparged with ultra-pure carrier grade air (a synthetic blend of oxygen and nitrogen) to remove inorganic carbon. In a high-temperature combustion (680°C), dissolved organic carbon was quantitatively oxidized on a catalyst bed of platinum-coated alumina to break down organic carbon to CO2. The CO2 was carried by ultra-pure carrier grade air to the non-dispersive infrared detector (NDIR) for detection. Concentrations of DOC were measured against a calibration line using the external standard potassium hydrogen phthalate (KHP). Standards in the calibration line were 0, 0.5, 1.0, 5.0, and 10.0 mg/L in KHP. The Continuing Calibration Verification (CCV) solution which contained 5.0 mg/L KHP was analyzed 5 times during sample analysis, with all measurements within 1.4% of the standard KHP concentration. 2.3 Results 2.3.1 Dissolved Sr2+ and Ba2+ concentrations Dissolved Sr2+ and Ba2+ concentrations vary significantly in western MD streams throughout the Feb 2012–Feb 2013 and Apr 2013–Apr 2014 sample sets. Stream waters of the two western MD sample sets range between 100 nM and 700 nM in Ba2+ across space and time. Strontium concentrations at sites in the region show a 54 greater range, with concentrations across both sample sets varying between 100 nM and 1,400 nM in space and time for most stream waters, although one site (PRUN301D) reached 3,200 nM in August 2013 (Table A1). Spatially, Sr2+ and Ba2+ concentrations are significantly variable both within and among catchments. Strontium shows greater variability in space, with concentrations across the region varying by 1,000–3,000 nM in summer months when spatial variability is greatest (June and August). Variability in Ba2+ throughout the region is lower, with concentrations varying by 550 nM in August when spatial variability of Ba2+ is greatest. At the site PRUN301D, which was monitored biannually, exceptionally high Sr2+ concentrations were observed in Aug 2013 and Jan 2014, with concentrations of 3,151 nM and 1,453 nM, respectively (Table A1). The Ba/Sr ratio was accordingly very low relative to all other sites. Even when not including this site, the spatial variability of Sr2+ concentrations is greater than that of Ba2+ concentrations. Significant temporal variability in dissolved Sr2+ and Ba2+ is also observed for western MD sites (Fig. 2.3). Concentrations range by 100 to 600 nM during one year in most streams. Certain sites, including Tributary T1 and Tributary C6, show strong seasonality for both metals, with concentrations highest between late summer and fall and lowest in early spring, while other sites show no distinct seasonal patterns. Figure 2.3 shows four sites in the Casselman River catchment (Panels A, B, D, and F) and two in the Savage River catchment (Panels C and E), illustrating the different temporal patterns both within and between catchments. 55 Most sites show consistent relative enrichment of either Sr2+ or Ba2+ throughout the year (Fig. 2.3). This is indicated by the logarithm of [Ba]/[Sr], the ratio of molar Ba and Sr concentrations, where high Ba2+ concentration relative to Sr2+ gives a value >0 and high Sr2+ concentration relative to Ba2+ give a value <0. Values of log([Ba]/[Sr]) determined for western MD stream sites ranged between –0.85 and +0.37, as shown in Figure 2.4. Although most sites show consistent enrichment of one element, a number of sites in the Casselman River and Youghiogheny River catchments were found to alternate between enrichment of Sr2+ and Ba2+ throughout the year (Fig. 2.4). Among these sites, enrichment of Sr2+ was typically observed in fall and enrichment of Ba2+ observed in winter, spring, and summer. Among sites with continual enrichment of Sr2+ or Ba2+, the dominance of a particular constituent—either Sr2+ or Ba2+—was not found to be related to catchment but varied within catchments as much as throughout western MD. In Calvert County streams, Ba2+ concentrations are between 100 nM and 300 nM, within the range of concentrations measured in western MD. Strontium concentrations in Calvert County have a slightly greater range than concentrations of most western MD streams, with concentrations between 100 nM and 1,800 nM (Table A1). The Parkers Creek sites in this region show Sr2+ concentrations at the high end of this range, varying between 1,400 nM and 1,800 nM, while all other Calvert County sites are between 100 nM and 500 nM. For both Sr2+ and Ba2+, significant spatial variability is observed across Calvert County as well as for multiple sites within the same stream, although Grays Creek is an exception with similar Sr2+ concentrations at both stream sites (GRAY01 and GRAY02) in June and July 2014. 56 F M A M J J A S O N D J F M A B. Tributary C6 200 400 600 800 F. CASS105D 200 400 600 800 A. CASS103D C o n c e n t r a t i o n ( n M ) 200 400 600 800 Sr (nM) Ba (nM) C. SAVA301D C o n c e n t r a t i o n ( n M ) 200 400 600 800 D. Tributary T1 200 400 600 800 E. SRPS 200 400 600 800 C o n c e n t r a t i o n ( n M ) F M A M J J A S O N D J F M A F M A M J J A S O N D J F M A F M A M J J A S O N D J F M A F M A M J J A S O N D J F M A F M A M J J A S O N D J F M A Figure 2.3 Temporal variability in dissolved Sr2+ and Ba2+ concentrations (nM) across 1 year. Panels A, B, D, and F are sites in the Casselman River catchment, and panels C and E are sites in the Savage River catchment. SRPS indicates Savage River Pumping Station. Sites in Panels B, D, and E are from the 2012−2013 sample set, and those in Panels A, C, and F are from the 2013−2014 sample set. 57 Between June and July, most sites in Calvert County were stable in Sr2+ concentrations, although Parkers Creek sites increased. Barium concentrations increased between the two months for most sites, apart from sites in Grays Creek which were relatively constant. At Parkers Creek and Grays Creek sites, Sr2+ and Ba2+ showed the same concentration pattern between June and July, with both metals temporally constant at Grays Creek sites and both metals showing increases over time at Parkers Creek. Relative concentrations of Ba2+ and Sr2+ in Calvert County show that sites have consistent enrichment of either Sr2+ or Ba2+ across June and July, indicated in the plot of log([Ba]/[Sr]) (Fig 2.4). As in western MD, the dominance of a particular element varies throughout Calvert County. The dominant element also varies within a single stream, as is observed in Grays Creek and Thomas Creek, although Parkers Creek shows dominance of Sr2+ at both sampled sites. As Sr2+ concentrations in Parkers Creek were relatively high and above the range measured for most western MD sites, the Ba/Sr ratios are very low and the log([Ba]/[Sr]) values, between –0.8 and –1.0, are the most negative of both western MD and Calvert County samples. All other Calvert County sites show log([Ba]/[Sr]) values between –0.2 and +0.13 (Fig. 2.4), within the range of values observed for western MD sites. W T P I n t a k e W T P F i n i s h e d R e s e r v o i r T r i b u t a r y T 1 T r i b u t a r y C 6 S R P S C A S S 1 0 1 D C A S S 1 0 3 D C A S S 1 0 5 D G E O R 1 0 1 D P R U N 3 0 1 D S A V A 1 0 1 D S A V A 2 0 2 D S A V A 2 0 4 S S A V A 3 0 1 D S A V A 3 0 2 D Y O U G 1 0 2 D Y O U G 1 0 3 D Y O U G 1 0 4 D Y O U G 1 0 5 D Y O U G 1 0 6 D Y O U G 2 0 1 D Y O U G 3 0 1 D Y O U G 3 0 2 D Y O U G 4 3 2 S H E L L 0 1 G R A Y 0 1 G R A Y 0 2 P A R K 0 1 P A R K 0 2 T H O M 0 1 T H O M 0 2 l o g ( [ B a ] / [ S r ] ) -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 Apr Jun Aug Oct Dec Jan Apr Jun Jul 2012 2013 2014 Figure 2.4 Values of log([Ba]/[Sr]) at each sampling site, with Feb 2012–Feb 2013 sites shown in the left panel, Apr 2013–Apr 2014 sites shown in the center panel, and June−July 2014 Calvert County sites shown in the right panel. Bars plotting above 0 indicate Ba2+ is elevated in concentration relative to Sr2+ and those plotting below 0 indicate Sr2+ is elevated. 59 2.3.2 Major ion concentrations Major ions are highly variable across western MD streams. Potassium concentrations show the lowest range among major cations, with concentrations found between 10−160 µM (Table A1). Sodium (Na+) concentrations are highly variable and span a range of 10−2,600 µM throughout space and time, with most values in the range of 10−1,100 µM. Concentrations of Ca2+ and Mg2+ in western MD streams are primarily between 30 µM and 800 µM, although Ca2+ levels reach as high as 1,350 µM in some sites and seasons. Sodium and Ca2+ are the dominant cations in all streams, with relative levels of each varying throughout western MD. Anion concentrations are highly variable over space and time throughout western MD streams. Concentrations of 24SO  range between 10−1,600 µM, although concentrations are most frequently between 10−400 µM (Table A1). Levels of Cl− show greater variability through space and time, with concentrations found between 10–3,200 µM. Chloride concentrations are predominantly in the range of 10−1,300 µM (Table A1). Values of ANC measured for Oct 2013 samples ranged between 38.5 µeq/L and 2,334 µeq/L and are taken to be equivalent to concentrations of 3HCO (µM) based on the pH values of stream waters, which range between 5.5 and 7.7 (Table A1). Also within this range are 3HCO concentrations estimated from the excess of cation charge in all samples not analyzed for 3HCO by ANC (Section 2.4.1). The positive charge excess was attributed to 3HCO anions, as samples which were analyzed for 3HCO show a balance of ionic charge within 10%, with the spread of 10% associated 60 with charge contributions of DOC and 3NO as well as variability within major cation and anion analyses (section 2.4.1). Measurements of DOC taken in western MD streams in April 2014 ranged between 49.1–166 µM C, an order of magnitude lower range than is found in 3HCO concentrations (Table A1). The range is lowest in the Youghiogheny River catchment, with concentrations spanning 49.1–76.6 µM C, while within other catchments concentrations range between 50.0–166 µM C. Anionic charge contributed from DOC was determined for samples in which DOC was analyzed (Apr 2014) and used in calculations of ion balance. Anionic charge equivalents [A–] (meq/L) of DOC were determined from the following equation: T + – K[C ][A ] = 1000K +[H ]  (2.5) where K is the pH-dependent mass action quotient determined from titration of humic and fulvic organic acids (Oliver et al., 1983), and CT is the organic acid concentration (eq/L), which was determined by assigning a carboxyl concentration of 10 µeq/mg C in DOC (Oliver et al., 1983). Measured values of pH from April 2013 were used in the calculation. Concentrations of 3NO were measured in western MD samples from Feb 2012−May 2012. Ranging between 44−244 µM (Table A1), 3NO concentrations are also an order of magnitude lower in range than those of 3HCO . Nitrate 61 concentrations were temporally variable at each site, with highest concentrations observed in February and March and lower concentrations in April and May. Ancillary analysis of Br– concentrations was carried out by ion chromatography at NASL. Six samples (WTP Intake, WTP Finished, Reservoir, SPRS, Trib. T1, and Trib. C6) were analyzed from collections in Feb 2012–June 2012. All samples showed Br– concentrations below the detection limit of 1 µM. Major cation concentrations in Calvert County streams are similar to those in western MD. Levels of K+ are relatively low, as in western MD, and fall within the range of 10−100 µM (Table A1). Sodium is the dominant cation at most sites, with concentrations ranging between 100 µM and 2,000 µM in concentration, although Ca2+ is the dominant cation at Parkers Creek sites. Concentrations of both Mg2+ and Ca2+ range from 20−800 µM among sites in Calvert County. Stream concentrations of Cl− in Calvert County waters range from 100−2,600 µM, within the range of western MD Cl− concentrations (Table A1). Concentrations of 24SO  are between 30−140 µM, falling at the lower end of the 24SO  concentration range found in western MD despite the proximity to Chesapeake Bay. 2.4 Discussion 2.4.1 Charge balance in western MD streams Charge contributions of all measured cations and anions were determined in each stream sample. Sodium and Ca2+ are the greatest contributors to the cationic charge, with Na+ contributing up to 75% of positive charge and Ca2+ contributing up to 80% 62 among all samples. Magnesium contributes between 8–40% of positive charge, and K+ contributes less than 8% among all samples. Total equivalents of cationic charge, in meq/L, ranged between 0.2–4.4. Relative to total equivalents of positive charge, Cl– is the dominant anion in 43% of samples and contributes up to 85% of anionic charge. For low Cl– samples in which it is not dominant, Cl– may contribute as little as 2% of anionic charge. Sulfate is dominant in 18% of samples and contributes up to 95% of the balance of negative charge, although it also may contribute as little as 2% in samples for which it is not dominant. In all other waters where Cl– and 24SO  are not dominant, it is expected that 3HCO is the dominant anion. In samples not analyzed for 3HCO , an excess of cations is found for the charge balance. It is expected that the unaccounted for negative charge from this calculation is primarily attributable to 3HCO , as samples for which 3HCO was analyzed by ANC titration show relatively large contributions from this constituent. Samples analyzed for 3HCO show a balance of charge within 10% of 0 for most waters, which is likely attributable to charge contributions of DOC and 3NO and to variability from error in the major cation and anion analyses. Charge contributions of DOC are estimated to account for <5% of anionic charge in samples, and 3NO contributes <10% for most samples in which it was analyzed. Other analyses of headwater streams in several Appalachian highlands regions have similarly shown low contributions of 3NO and DOC relative to other ions (Mast, 1999). 63 The charge contributed from 3HCO , estimated from the excess of positive charge in samples, is found to be dominant in 38% of samples and contributes up to 90% of anionic charge, although, in samples for which it is not dominant it may contribute negligible charge (Table A1). Analyses of charge contributions in western MD stream waters indicate the dominant ions in this region, demonstrating the current pre-fracking conditions of western MD streams. 2.4.2 Major ions as tracers of fracking fluids All major ions were found to be 1–3 orders of magnitude greater in concentration than Sr2+ and Ba2+ in streams and were highly variable throughout space and time, with each measured ion ranging over 1–2 orders of magnitude in concentration, suggesting that individual major ion measurements may not be useful as tracers of fracking fluids. To investigate possible patterns in variability, relative concentrations of major ions were evaluated. Sodium and Ca2+ were found to be the dominant cations in all western MD and Calvert County streams although the dominance of Na+ vs. Ca2+ varies spatially both within and among catchments. Other studies in Appalachian highland streams similarly show dominance of Na+ and Ca2+ with relative abundance of each varying within single catchments (Johnson et al., 2000; Timpano et al., 2015). The dominance of Ca2+ in certain streams is likely associated with areas of carbonate mineral bedrock (Roth, 1999). Stream water Sr2+ and Ca2+ showed correlation among all sampled streams (Fig. 2.5), a reasonable finding since Sr2+ substitutes for Ca2+ in 64 Ca-bearing minerals (Peek and Clementz, 2012). The strength of this correlation varied among western MD waters, though, likely related to varying mineral ratios of 0 200 400 600 800 1000 1200 1400 S r 2 + ( n M ) 0 500 1000 1500 2000 2500 3000 3500 Ca 2+ (M) Figure 2.5 Concentrations of Sr2+ vs. Ca2+ in all stream waters. 0 500 1000 1500 2000 2500 0 200 400 600 800 1000 1200 1400 HCO 3 - (M) C a 2 + (  M ) Figure 2.6 Concentrations of Ca2+ vs. 3HCO in western MD stream samples collected Oct 2013. 65 Sr/Ca (Land et al., 2000) (Fig 2.5). Calcium concentrations were also correlated with 3HCO concentrations during Oct 2013 when 3HCO was measured (Fig. 2.6), supporting the theory that high-Ca2+ sites are associated with locations of underlying carbonate rock. In Calvert County where carbonate bedrock is less prominent (Roth, 1999), Ca2+ is not found to be a dominant cation, although Parkers Creek sites, which show relatively high Ca2+ and Sr2+, are an exception and may be influenced by localized carbonate minerals in proximity to Parkers Creek (Carnevale et al., 2011). Variability in major cation composition was further evaluated by plotting stream waters on a trilinear diagram of relative cation concentrations which shows distinctions of stream waters in western MD and Calvert County (Fig. 2.7). Major cation composition may be dictated in part by catchment area in which the stream is found, as some stream samples in the diagram plot in groups according to catchment, which is indicated by color. Some stream waters within the same catchment, though, plot in separate areas of the diagram. For example, some waters in the Savage River catchment plot in an area indicating relatively high Ca2+ contribution (45%−60%) and fairly low Na++K+ contribution (20%−25%), and other Savage River waters show a grouping of relatively high Na++K+ (60%−70%) with fairly low contributions of Ca2+ (20%−30%). This suggests that another major factor affects stream water chemistry. It is suspected that variable bedrock geology of the region has a predominant influence on ion composition as groundwater is expected to be a major stream water source under base flow conditions (Ahearn et al., 2004). In the trilinear diagram, the point marked with a star represents average cation composition of shallow groundwater overlying the Marcellus Shale in southern New York (Fig. 2.7) N a + + K + ( % ) 0 10 20 30 40 50 60 70 80 90 100 M g 2 + ( % ) 0 10 20 30 40 50 60 70 80 90 100 Ca 2+ (%) 0102030405060708090100 Frostburg City Water Treatment Plant Piney Reservoir Casselman River Catchment Savage River Pumping Station (groundwater) Savage River Catchment Youghiogheny River Catchment Georges Creek Catchment Potomac River Upper North Catchment West Chesapeake Bay Catchment Patuxent River Lower Catchment Shallow groundwater, NY (Lautz et al., 2014) Figure 2.7. Trilinear diagram showing relative contributions (%) of Mg2+, Ca2+, and Na++K+ cations in western MD and Calvert County streams. Marker color represents catchment as in Fig. 2.1. Average cation composition of shallow groundwater in a region overlying the Marcellus Shale (New York state) is represented by the red star (Lautz et al., 2014). 67 (Lautz et al., 2014), with the shallow groundwater plotting among most stream water samples in the diagram. This demonstrates the similarity in composition of shallow groundwater and stream water, further suggesting that groundwater is an important contributor to western MD streams. Bedrock geology is variable in western MD, as illustrated in the Geologic Map of Western Maryland (Brezinski, 2013) (Fig. 2.8), with the two map background colors representing different bedrock formations. In this region, bedrock geology was previously shown to be an important factor in stream chemistry responses to acidic deposition (Roth, 1999). In the current study, plotting of sites on the geologic map Figure 2.8. Geologic map of western MD by Brezinksi et al. (2013). Map background shades represent the underlying bedrock type. Green shades represent Mississippian and younger formations, and purple shades represent Devonian and older formations. All western MD stream sites are marked, with marker color indicating catchment as in Fig. 2.1. 68 (Fig. 2.8) shows that underlying bedrock type varies within and among catchments, with marker color corresponding to catchment as in Fig. 2.1. To assess the influence of underlying bedrock on stream water chemistry, study sites were categorized according to underlying bedrock era and stream waters were plotted bilinearly with K+ vs. Mg2+ concentrations (Fig. 2.9). The bilinear plot of major cation composition shows separation of stream waters according to bedrock formation era, either Devonian, shown in purple points, or Pennsylvanian, shown in green points in Fig. 2.9. The differentiation of waters based on underlying bedrock demonstrates the influence of bedrock geology on overlying stream water composition. 0 200 400 600 800 1000 0 20 40 60 80 100 120 140 160 K ( µ M ) Mg (µM) Devonian bedrock Pennsylvanian bedrock Figure 2.9. Bilinear plot of K+ vs. Mg2+ in western MD stream samples, with samples categorized by the underlying bedrock era, either Devonian or Pennsylvanian. 69 Although geology and its impact on groundwater affect stream water chemistry, runoff from road salt is a major factor as well. A plot of Na+ vs. Cl−, the main constituents of road salt, shows that stream waters generally plot along a line of mixing with road salt effluent (Lautz et al., 2014) (Fig. 2.10). This is in agreement with speculation following other northern U.S. studies that consistent seasonal road salt application has increased groundwater stores of mobile Na+ and Cl− which drive surface water levels of these ions (Thompson et al., 2011). N a + (  M ) Western MD streams Deep Appalachian brines (Warner et al., 2012) Road salt effluent (Lautz et al., 2014) 10,000,000 100,000 1,000 10 1 1 100 10,000 1,000,000 Cl - (mM) 100,000,000 Figure 2.10 Plot of Na+ vs. Cl– concentrations in deep Appalachian brines which occur in fracking fluids (Warner et al., 2012), in road salt effluent (Lautz et al., 2014), and in western MD streams. The plot in Fig. 2.10 also shows Na+ vs. Cl− concentrations of deep Appalachian brines, a major influence in fracking fluids from the Marcellus Shale. As observed in the plot, Na+/Cl− ratios are similar for western MD streams, fracking fluids, and road 70 salt effluents. Ratios of Na+/Cl− will therefore not be useful as fracking fluid tracers. The plot also demonstrates that stream water variability of Na+ and Cl− concentrations is about 3 orders of magnitude, indicating that the individual ions will not be suitable as fracking fluid tracers. The bivariate plot of Mg2+ vs. Cl− shows similarities in composition of western MD streams and fracking fluids containing Appalachian brines (Warner et al., 2012) (Fig. 2.11). The similarity means that Mg2+/Cl− ratios cannot be used to track fracking fluids in streams. Shallow groundwater overlying the Marcellus Shale in Pennsylvania is also plotted in Fig. 2.11 and shows close resemblance to headwater streams in MD, indicative of a groundwater origin of these streams (Lautz et al., 2014). M g 2 + (  M ) Western MD streams Deep Appalachian brines (Warner et al., 2012) Shallow groundwater (Warner et al., 2012) 1,000,000 10 1 1,00 10,000 1,00 10,000 1,000,000 100,000,000 Cl - (mM) Figure 2.11 Plot of Mg2+ vs. Cl– concentrations in deep Appalachian brines which occur in fracking fluids, in shallow groundwater overlying the Marcellus Shale (Warner et al., 2012), and in western MD streams. 71 With the variability in major cation and anion concentrations across western MD streams, these constituents are not useful tracers of fracking fluids. Values of major ion ratios are also not desirable as tracers without other constraining constituents, although analysis of ion ratios indicates predominant stream water sources. In addition to the variability, the major ions are relatively abundant, particularly in comparison to trace metal tracers such as Sr2+ and Ba2+. Relative to streams, fracking fluids are enriched in Sr2+ and Ba2+ by factors of 25,000−50,000, while fluids are enriched in the major ions by factors of 100−1,000 (Barbot et al., 2013; Warner et al., 2013a). Major ion levels in streams are likely to show lower shifts in cases of fracking fluid intrusion compared to other tracers. Dissolved Br−, which was measured irregularly for samples of the Feb 2012−Feb 2013 set, was found to be below detection limits in all analyzed stream samples. Although it could not be regularly measured in this study due to instrument limitations, the preliminary analyses show that Br− may be the optimal fracking fluid tracer in western MD since it is enriched in fracking fluids and occurs at very low levels in streams of this region. 2.4.3 Sr2+ and Ba2+ as tracers of fracking fluids Strontium and Ba2+ concentrations in western MD streams, although variable over space and time, agree with concentrations measured for a range of other minimally disturbed catchments in North America. Stream waters were found to have ranges of 140−800 nM in Sr2+ and 10−250 nM in Ba2+ (Elias et al., 1982; Fitzpatrick et al., 2007; Watmough, 2014), with a greater range for Sr2+ as was found 72 in the current study. This observation is due to higher solubility of Sr2+ than Ba2+ in minerals as well as the greater abundance of Sr2+ in nature relative to Ba2+(Nesbitt et al., 1980; Turekian and Wedepohl, 1961). Stream concentrations of Sr2+ and Ba2+ in Calvert County are in the same ranges as concentrations found in western MD streams, although Sr2+ concentrations show a greater range in Calvert County with Parkers Creek sites showing high Sr2+ outside the concentration range of most western MD streams. This occurrence may be associated with marine carbonates which have outcrops along the mouth of Parkers Creek (Carnevale et al., 2011). In western MD, a distinct temporal pattern in Sr2+ and Ba2+ levels is observed for sites Tributary T1, Tributary C6, and SAVA301D which have concentrations highest in late summer and fall and lowest concentrations in spring (Fig. 2.3). This indicates that these streams may be influenced by seasonal inputs of surface runoff. With groundwater supplying dissolved Sr2+ and Ba2+ year-round, metal concentrations are lowest in spring when surface runoff is high and highest in fall when runoff is minimal (Stranko et al., 2013). A number of other sites, including SAVA204S, YOUG102D, and YOUG104D, similarly show maximum concentrations in fall with large concentration decreases by spring, suggesting that Sr2+ and Ba2+ are primarily from groundwater and that streamflow is influenced by surface runoff in these streams. The temporal pattern may be less clear at these sites as there may be a portion of Sr2+ and Ba2+ supplied by surface runoff that has interacted with soil minerals (Hogan and Blum, 2003). Also, the sites with less distinct temporal patterns were sampled bimonthly (Apr 2013–Apr 2014), allowing lower resolution in 73 observed patterns, while the sites showing the most distinct temporal cycles are those sampled monthly (Feb 2012–Feb 2013). This suggests that monthly measurements of Sr2+ and Ba2+ may be optimal. Many western MD streams show temporal variability in Sr2+ and Ba2+ but no distinct temporal pattern (Fig. 2.3, panels A and E). This may indicate that these streams receive relatively low amounts of flow from surface runoff and are predominantly fed by groundwater throughout the year. Other sites in the region show relatively low Sr2+ and Ba2+ concentrations as well as relatively low temporal variability in concentrations, such as CASS105D (Fig 2.4, panel F), SAVA101D, GEOR101D, YOUG105D, and YOUG106D. For these sites, streamflow may be substantially supplied by surface runoff which is less enriched in Sr2+ and Ba2+ relative to groundwater (Peek and Clementz, 2012). Alternatively, for sites of this type, low concentrations and low temporal variability may indicate a groundwater source which has short aquifer residence time and therefore lower Sr2+ and Ba2+ concentrations, yet still provides these species at constant levels to streams (Peek and Clementz, 2012). Spatial variability of Sr2+ and Ba2+ concentrations is found both within and among catchments. Groundwater, the major contributor of Sr2+ and Ba2+ to surface waters (Peek and Clementz, 2012) is influenced by bedrock geology, a factor which is highly variable across space in western MD (Brezinski, 2013). Bedrock minerals found in western MD, including carbonates and weathered clay minerals, vary in abundance of Sr2+ and Ba2+ (Brezinski, 2013; Turekian and Wedepohl, 1961). In addition to variability in abundances of these trace metals, minerals vary in their 74 weathering rates, affecting the release of Sr2+ and Ba2+ to groundwater (Land et al., 2000). Heterogeneity of mineral Sr2+ and Ba2+ sources to streams is reflected in the Ba/Sr ratios, assessed as log([Ba]/[Sr]) values here, which show greater variability over space than over time. Relatively stable Ba/Sr ratios across time are likely indicative of a temporally consistent geologic source of Sr2+ and Ba2+. In sites where seasonal shifts in concentrations are observed despite a stable ratio, dilution from runoff is likely affecting these constituents. The variability in log([Ba]/[Sr]) values across the region may demonstrate the spatial variability of mineral sources underlying stream beds (Fig. 2.4). The irregularity of weathering of Sr- and Ba-bearing minerals also affects log([Ba]/[Sr]) values across the region. Land et al (2000) found that, within a forested catchment, groundwater derived from deeper flowpaths showed lower Ba/Sr ratios compared to shallower groundwater and soil water. This was attributed to less intensive weathering environments at greater depths, where Sr is released at greater rates than is Ba, since Sr-containing minerals are overall more soluble. In western MD streams, molar Ba/Sr ratios reported as log([Ba]/[Sr]) are in agreement with log([Ba]/[Sr]) ranges occurring in other first-order North American streams (Hogan and Blum, 2003; Land et al., 2000). Values obtained in these studies, though, show lower ranges of log([Ba]/[Sr]) than observed here, which may be indicative of the high spatial variability of bedrock underlying western MD streams. Although stream dissolved Sr2+ and Ba2+ vary throughout western MD, stream concentrations, in the range of 100–1,400 nM for most sites, are very low in relation 75 to fracking fluids, which may allow observation of concentration spikes with spill events. Fracking fluid spill events are reported to be 2,900 L on average (Gross et al., 2013), as determined in a western U.S. shale region, although most documented spills in the Marcellus Shale region from 2009−2013 were larger (Brantley et al., 2014). Considering a 2,900 L spill from a well pad which drains to a nearby headwater stream over five hours, the discharge of fluids to the stream would comprise approximately 0.2% of streamflow, based on average discharge of 66 L/s, as determined on average from long term (2000–2014) monitoring stream sites in western MD (Saville et al., 2014). With this 0.2% contribution of fracking fluid at the site of discharge to the stream, the Sr2+ and Ba2+ levels in the stream may reach 38,690 nM and 32,400 nM, respectively, based on reported Sr2+ and Ba2+ levels in fluids (Barbot et al., 2013). As monitored stream sites are positioned just downstream of proposed well pad locations, these high concentrations which are 1–2 orders of magnitude greater than background concentrations would likely be observable above normal variability. Still, if stream sampling is not carried out within 1–2 days of a spill event, signals of Sr2+ and Ba2+ may be dampened and would be within normal concentration variability. It may therefore be best to use a continuous online monitoring tool for preliminary detection of potential fracking fluid contamination. Stream specific conductance, as noted, may be monitored this way using in situ sensors, where the specific conductance, a measure of stream water dissolved ions, would show significant spikes immediately following fracking fluid contamination (Brantley et al., 2014). Upon observance of spikes in conductivity, measurements of Sr2+ and Ba2+ concentrations could then be taken for verification that contamination is 76 from fracking fluids and not from other anthropogenic effluents which would be relatively low in Sr2+ and Ba2+ (Vidic et al., 2013). Measurements of Sr2+ and Ba2+ would show elevated concentrations following spill events and would confirm the presence of fracking fluids. 2.4.4 Sr and Ba solubility and its effect on tracer use The saturation states of BaSO4 and SrSO4 were calculated for western MD and Calvert stream waters in order to determine whether oversaturation affects Ba2+ and Sr2+ levels. The solubility product (Ksp) values were determined from literature values for 25°C and 0 ionic strength conditions, where Ksp = 1.102 × 10 –10 for BaSO4 (Templeton, 1960) and Ksp = 2.344 × 10 –7 for SrSO4 (Prieto et al., 1997). Ionic strength was determined according to the equation 2 i i i 1I = c z2 (2.6) where ci is the concentration of the i th ionic species and zi is the charge of the species. Ionic strength was found to be below 0.007 for stream waters and was approximated to be zero. The ion concentration product (ICP) was determined with respect to BaSO4 and SrSO4 in each stream sample and was compared to Ksp to determine saturation state Ω according to the equation sp ICPȍ K (2.7) The ICP values for BaSO4 and SrSO4 were calculated according to the equations 4 2+BaSO f 4 f2ICP = [Ba ] [SO ] (2.8) 77 4 2+SrSO f 4 f2ICP = [Sr ] [SO ] (2.9) where total measured molar concentrations of Ba2+, Sr2+, and 24SO  were taken to be equal to the free concentrations [Ba2+]f, [Sr 2+]f, and 24 f[SO ] , respectively, or [Ba 2+]T ≈ [Ba2+]f , [Sr 2+]T ≈ [Sr 2+]f, and 24 T[SO ] ≈ 24 f[SO ] . This was a conservative approximation since free concentrations are slightly lower than total concentrations due to complexation of Ba2+ and Sr2+ with Cl– and OH–, and of Na+ with 24SO  . Ratios of 2+ f 2+ T Ba Ba       and 2+ f 2+ T Sr Sr       , which account for this complexation, were calculated from the following equations: 2 f 2 * T BaOH BaCl + [Ba ] 1= [Ba ] ȕ1+[Cl ]ȕ  [H ]         - (2.10) 2 f 2 * T SrOH SrCl + [Sr ] 1= [Sr ] ȕ1+[Cl ]ȕ  [H ]         - (2.11) where βBaCl = 0.363, *BaOHȕ = 4.603 × 10 –14, βSrCl = 0.603, and *SrOHȕ = 6.653 × 10 –14 (Martell, 2004). Values of stream pH collected for April 2013 were used for approximating pH values in each sample. The ratios of 24 f 2 4 T – – SO SO       accounting for 4NaSO complexation were determined by: 4 2 4 f 2 4 T NaSO [SO ] 1=[SO ] (1+[Na ]ȕ )   (2.12) 78 where 4NaSOȕ = 5.495 (Martell, 2004). Ratios of 2+ f 2+ T Sr Sr       were found to be between 0.998–1.000 for all samples. Ratios of 2+ f 2+ T Ba Ba       were between 0.999–1.000 for all samples, and ratios of 24 f 2 4 T – – SO SO       were between 0.986–0.999 for all samples. Using total concentrations [Ba2+]T, [Sr 2+]T, and 24 T[SO ] , to calculate ICP values, no samples were found to be oversaturated with respect to SrSO4, whereas possible oversaturation with respect to BaSO4 (Ω ≥ 0.9) was found to occur for 15 samples. For the 15 samples which showed oversaturation of BaSO4, true [Ba 2+]f and 24 f[SO ] concentrations were calculated from Equations (2.10) and (2.12) to verify the oversaturation state when accounting for BaCl+, BaOH+, and 4NaSO ion pairs. Calculation of ICP using free concentrations confirmed that oversaturation occurred for each of the 15 stream samples, with saturation states of ~1.1−7.5 as shown in Table 2.2. All occurrences of oversaturation were found at four sites located in the Youghiogheny River and Potomac River Upper North Branch catchments. In these stream reaches, 24SO  concentrations are high relative to all other sites (Table A1), which largely affects the degree of saturation. Each of these sites also shows relatively high Sr2+ and Ca2+ levels. Waters from three of the four sites (YOUG102D, YOUG103D, and PRUN301D) are elevated in 24SO  relative to 3HCO , indicating that underlying anhydrite minerals (CaSO4) may influence these areas. Relatively 79 high 3HCO as well as Ca 2+ at site YOUG104D suggests that carbonate bedrock underlies this stream area. At all four sites, Ba2+ concentrations are relatively high, although oversaturation is not solely dependent upon Ba2+ concentration since other sites with similar Ba2+ levels, such as CASS101D and SAVA301D, are not oversaturated. All stream samples which show oversaturation were also found to be those with greatest enrichment of Sr2+ relative to Ba2+ with log([Ba]/[Sr]) values between −0.85 and −0.36 (Fig. 2.4). No sites in Calvert County were found to be oversaturated, likely related to the relatively low 24SO  concentrations in this region. Table 2.2 Saturation state (Ω) of western MD samples found to be oversaturated in BaSO4, with calculated free Ba 2+ concentration (nM), free 24SO  concentration (µM), and Ion Concentration Product (ICP) of each sample. Sample [Ba]f (nM) 24 f[SO ] (μM) ICP Ω YOUG102D Apr 2013 285 757 2.16 × 10– 10 1.96 YOUG102D Jun 2013 404 1011 4.09 × 10– 10 3.71 YOUG102D Aug 2013 430 1174 5.04 × 10– 10 4.58 YOUG102D Oct 2013 525 1575 8.27 × 10– 10 7.50 YOUG102D Dec 2013 216 530 1.15 × 10– 10 1.04 YOUG102D Apr 2014 234 549 1.28 × 10– 10 1.17 YOUG103D Apr 2013 270 535 1.45 × 10– 10 1.31 YOUG103D Jun 2013 390 667 2.60 × 10– 10 2.36 YOUG103D Aug 2013 400 832 3.33 × 10– 10 3.02 YOUG103D Oct 2013 476 1072 5.11 × 10– 10 4.64 YOUG104D Jun 2013 469 311 1.46 × 10– 10 1.33 YOUG104D Aug 2013 474 302 1.43 × 10– 10 1.30 YOUG104D Oct 2013 495 336 1.66 × 10– 10 1.51 PRUN301D Aug 2013 448 1541 6.90 × 10– 10 6.27 PRUN301D Jan 2014 258 650 1.68 × 10– 10 1.52 Although fracking fluids do not have a distinct log([Ba]/[Sr]) signature (Fig. 2.12), findings in this study suggest that log([Ba]/[Sr]) values may be suitable tracers 80 of fracking fluids in western MD due to the tendency for BaSO4 oversaturation, which is expected to occur in streams with the intrusion of fracking fluids. The oversaturation with respect to BaSO4, which was found to occur in western MD streams, is expected to allow increased Sr2+ concentrations with fracking fluid stream contamination. With BaSO4 oversaturation, Ba 2+ concentrations would remain relatively constant in such contamination events. A significant decrease in log([Ba]/[Sr]) would then be observed in the affected stream, particularly as log([Ba]/[Sr]) values are relatively stable across time at most sites. The Sr2+ and Ba2+ stream concentrations described for a 0.2% fracking fluid introduction, 38,690 nM and 32,400 nM, respectively, will likely cause oversaturation of BaSO4, particularly Western MD Sites W T P I n t a k e W T P F i n i s h e d R e s e r v o i r T r i b u t a r y T 1 T r i b u t a r y C 6 S R P S C A S S 1 0 1 D C A S S 1 0 3 D C A S S 1 0 5 D G E O R 1 0 1 D P R U N 3 0 1 D S A V A 1 0 1 D S A V A 2 0 2 D S A V A 2 0 4 S S A V A 3 0 1 D S A V A 3 0 2 D Y O U G 1 0 2 D Y O U G 1 0 3 D Y O U G 1 0 4 D Y O U G 1 0 5 D Y O U G 1 0 6 D Y O U G 2 0 1 D Y O U G 3 0 1 D Y O U G 3 0 2 D Y O U G 4 3 2 S 0 A B C D E F G -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 l o g ( [ B a ] / [ S r ] ) Fracking fluids Figure 2.12 Values of log([Ba]/[Sr]) found in western MD streams (light gray) and in fracking fluids (dark gray). Western MD values were determined from average molar Sr2+ and Ba2+ concentrations from each time of sampling. Fracking fluid values were determined from molar concentrations from fracking fluid analyses as reported by Barbot et al. (2013) (A), Hayes et al. (2009) (B), He et al., (2014) (C, D, and E), Warner et al. (2012) (F), and Warner et al. (2013b) (G). 81 with increased 24SO  supplied by fluids, with He et al (2014) having shown that 24SO  addition to fracking fluid samples results in rapid oversaturation with respect to BaSO4. Fracking fluid contamination following a spill event could thus be further verified with evaluation of the stream water log([Ba]/[Sr]) value. 2.4.5 Implications and Conclusions Headwater streams in western MD show significant variability in Sr2+ and Ba2+ over space and time. Variability is expected to be primarily related to variability of underlying geology in the region since groundwater is a predominant source of Sr2+ and Ba2+ in streams. In headwater streams near well pads, dissolved Sr2+ and Ba2+ concentrations may be useful as tracers which verify stream intrusion of fracking fluids, with initial detection of contamination indicated by a continuous stream monitoring tool. Measured log([Ba]/[Sr]) values vary across space but show lower variability across time for each site. These values may be useful as fracking fluid tracers. Due to BaSO4 oversaturation which is expected to occur with introduction of fracking fluids to streams, Sr2+ is expected to increase to much greater levels than Ba2+. The log([Ba]/[Sr]) in stream waters would then decrease from the baseline value with fracking fluid contamination. Use of Sr2+, Ba2+, and log([Ba]/[Sr]) values as tracers requires monitoring of concentrations of these metals before, during, and after fracking activity in order to observe departures from natural levels. In order to capture the naturally occurring temporal variability of Sr2+, Ba2+, and log([Ba]/[Sr]), it was determined that monthly baseline monitoring is suitable. Monthly concentration measurements are a sufficient 82 frequency for observation of patterns and variation in Sr2+ and Ba2+, as observed in Fig. 2.4, Panels B, D, and E, while sampling bimonthly, as in Panels A, C, and F, shows lower resolution in concentration patterns. Monthly monitoring should then continue during and after fracking development, allowing for continued observation of trends and patterns in Sr2+, Ba2+, and log([Ba]/[Sr]). In addition, during fracking development, these tracers should be used for verification of the presence of fracking fluids in stream contamination events, with measurements immediately following such events, as elevated Sr2+ and Ba2+ and decreased log([Ba]/[Sr]) will be observed with fracking fluid intrusion. Contamination events can be initially detected using continuous monitoring tools such as the in situ specific conductivity sensor. Similarly, with the indication that dissolved Br− will be valuable as a fracking fluid tracer, regular monitoring of Br− should be carried out on a monthly basis prior to fracking in order to capture temporal variability of this constituent. With the start of fracking activity in the region, monthly Br− measurements should continue for observation of trends, and, in addition, this constituent may be used to verify presence of fracking fluids in cases where contamination is suspected based on continuous monitoring data. For spatial considerations of Sr2+ and Ba2+, these constituents should be monitored in streams immediately downstream of well pads. This was the basis for site choice in the current study, and, as new well pad locations are leased in MD or in other areas of fracking development, new stream monitoring sites should be established. Entrekin et al (2011) found that Marcellus Shale gas wells are typically quite close to streams, with an average distance of 153 m to stream channels. It is 83 important to assess streams at these vulnerable locations which would likely be the first points to show water quality changes. For monitoring of Sr2+ and Ba2+, in particular, stream sites should be immediately downstream of well pads since spill signals could be diluted further downstream and may appear to be within baseline variability. With the use of Br− as a tracer, spatial distribution of monitoring would similarly include sites immediately downstream of well pads which may be most directly impacted by fracking activity. Ions Na+, K+, Mg2+, Ca2+, Cl−, and 24SO  all vary significantly across western MD streams, with bedrock geology and groundwater having considerable influence on variability. The high variability and relative abundance of major cations and anions demonstrate that periodic ion measurements are not applicable for detection of fracking fluids. Ion ratios allow assessment and comparison of stream waters, although major ion ratios also are not likely to be useful tracers due to their resemblance to fracking fluids. Monitoring of Sr2+, Ba2+, and log([Ba]/[Sr]) values suggest that these constituents will be most beneficial for recognizing fracking fluid presence in western MD streams, and localized Br− measurements in 2012 demonstrated this constituent should also be monitored as it will be an advantageous tracer of fracking fluids. 84 Chapter 3: Monitoring of dissolved methane in western Maryland streams 3.1 Introduction Methane (CH4) is found naturally in freshwater due to both groundwater inputs and in situ biological productivity, yet it may also be transferred to freshwater as a result of CH4 migration linked to hydraulic fracturing (fracking) operations. Little is known of the natural spatial and temporal variability of CH4 in streams which may be impacted by fracking (Brantley et al., 2014). To address this knowledge gap, naturally occurring dissolved CH4 was monitored across spatial and temporal ranges in western Maryland (MD) streams. This may enable future identification of CH4 contamination associated with fracking. 3.1.1 Sources of methane in natural waters Dissolved CH4 is found naturally in most aquatic systems, including ocean waters, estuaries, groundwater, streams, lakes, and swamps (Whiticar and Faber, 1986). In many of these systems, CH4 derives from biogenic production, referred to as methanogenesis (Schoell, 1988). In methanogenesis, microbes primarily produce CH4 under anoxic conditions through one of the following reactions: Carbonate reduction: CO2(g) + 8H + + 8e–  CH4(g) + 2H2O(l) (3.1) Acetate fermentation: CH3COOH(aq)  CH4(g) + CO2(g) (3.2) 85 where CO2 is carbon dioxide, H + represents protons, e– represents electrons, H2O is water, and CH3COOH is acetate, with CO2 and CH3COOH representing competitive substrates as they are more effectively used by other microbial assemblages (Whiticar, 1999). In some cases CH4 is also produced from non-competitive substrates such as methanol (CH3OH) and methylated amines (Whiticar, 1999). A methanogenic reaction utilizing CH3OH substrates is represented by the following (Liu and Whitman, 2008): 4 CH3OH(aq)  3 CH4(g) + CO2(g) + 2 H2O(l) (3.3) Methane is also produced thermogenically in deep subsurface environments (Schoell, 1988). Thermogenic CH4 is primarily generated in sedimentary basins due to the breakdown of hydrocarbons upon their exposure to high temperatures and pressures (Schoell, 1988). It has been shown to migrate upward from deep reservoirs, with major pathways along faults and fractures as well as lateral displacement and movement through permeable rock (Etiope and Klusman, 2002). Through these pathways, thermogenic CH4 may eventually intrude groundwater, surface water, and soils (Van Stempvoort et al., 2005). 3.1.2 Methane in groundwater Dissolved CH4 in groundwater originates from upward migration of thermogenic CH4 sources as well as from in situ biogenic production (Taylor et al., 2000; Vengosh et al., 2014). Concentrations in groundwater are found to be highly variable within and among regions of North America (Brantley et al., 2014). Across the Appalachian Basin of the eastern U.S., which contains the Marcellus Shale, groundwater dissolved 86 CH4 concentrations range from 0–4.42 mM (0–71 mg/L), depending on location (Vidic et al., 2013). Similar variability is reported for groundwater in other regions of North America (Aravena and Wassenaar, 1993; Zhang et al., 1998). Dissolved CH4 in groundwater raises environmental and public health concerns since groundwater is widely used for household and agricultural needs across the U.S. (Vidic et al., 2013). Occurrence of CH4 at concentrations above 0.623 mM (10 mg/L) in domestic and drinking water sources is an asphyxiation and explosion hazard (Revesz et al., 2010). More than 40 million U.S. citizens obtain drinking water from private wells fed by shallow groundwater (Vidic et al., 2013), and in Garrett County, MD, the area of interest for this thesis and an area of potential fracking development, 90% of citizens use private well water (Eshleman, 2013). Given the water quality risks associated with CH4, the U.S. Department of the Interior advises caution and an investigation if well water reaches dissolved CH4 concentrations of 0.623 mM (10 mg/L), and immediate action is necessary if concentrations reach 1.76 mM (28 mg/L) (Vidic et al., 2013). Variability of groundwater CH4 may be due to the irregularity of underlying geology. For example, cross-formational fractures and faults, which are prevalent in some areas (Sharma et al., 2014), facilitate upward flow of CH4 to aquifers driven by the buoyancy of gas-phase CH4 (Etiope and Klusman, 2002; Sharma et al., 2014). In several regions of North America, including the Michigan Basin (Long et al., 1988; Weaver et al., 1995), Appalachian Basin (Schedl et al., 1992), and the Williston Basin in Manitoba, Canada (Grasby and Betcher, 2002), evidence of cross-formational 87 migration of fluids has been found, and it is theorized that such cross-formational networks also facilitate upward migration of CH4 (Llewellyn, 2014). Cross-formational fluid migration has been shown to connect the Marcellus Shale, a Devonian formation of 1,200–2,600 m depths, to shallow groundwater (Warner et al., 2012). It is indicated that this hydraulic connectivity may promote upward CH4 migration, particularly in the presence of fracking activity (Warner et al., 2012). Some have contested this, however, theorizing that, with low permeability of the Marcellus Shale and overlying formations, the upward migration of shale CH4 is only possible over geologic time scales and is not of concern with regard to fracking development (Flewelling and Sharma, 2014; Flewelling et al., 2013). 3.1.3 Methane in surface waters In freshwater rivers and streams, CH4 is derived from in situ microbial production, transport from methanogenesis in riparian areas, and transport from groundwater through bedrock flow paths (Jones and Mulholland, 1998b). In a forested Appalachian stream, Jones and Mulholland (1998b) found inputs from riparian zones to be the predominant CH4 source; however, relative inputs of the stream CH4 sources vary within and among streams due to irregularity of factors such as hydrologic flow paths, organic matter storage in the catchment, and the availability of oxygen in riparian zones (Jones and Mulholland, 1998b). Consequently, in stream networks where CH4 has been measured, concentrations vary over space and time. Jones et al (1998a) showed that in summer, CH4 concentrations were positively related to soil organic matter storage for streams in southern Appalachian catchments. 88 In hemiboreal streams, Wallin et al. (2014) attributed measurements of high CH4 to locations of riparian peat or standing waters at low flow where CH4 production is prolific. Methane in terrestrial surface waters has particularly important implications with regard to CH4 emissions to the atmosphere (Etiope and Klusman, 2002; Jones and Mulholland, 1998b). It has been suggested that CH4 emissions from surface waters may offset as much as 25% of the estimated terrestrial greenhouse gas sink (Bastviken et al., 2011), based on a compilation of CH4 concentration measurements in freshwater systems. Although river systems were included in this estimate, they were represented by fewer data sets than were lakes and reservoirs, and in general, data on CH4 concentrations and evasion from streams are less prevalent. Recent studies also show that the areal extent of headwater streams has been underestimated by previous inventories, and stream CH4 emission estimates based on existing data may be too low (Benstead and Leigh, 2012; Downing et al., 2012). Estimates of CH4 emissions from streams are therefore not widespread or comprehensive (Wallin et al., 2014). To better inform terrestrial CH4 emission estimates, it is important to understand natural CH4 occurrence and variability in stream systems. Toward this objective, monitoring of CH4 concentrations in MD streams is carried out in the current study, and measurements of stable carbon isotope composition are also made to elucidate sources of CH4. 89 3.1.4 The isotopic composition of carbon in dissolved CH4 For determining the origin of CH4 in various aquatic and geologic environments, analysis of carbon isotopic composition is useful because the carbon isotope ratio of CH4 varies depending on whether it is formed biogenically or thermogenically (Schoell, 1980). The carbon isotopic composition of a substance is expressed relative to that of a standard reference material using delta notation (δ13C), which is defined as:  13 Sample Standardį & 5  5 1×1000   (3.4) where R represents the 13C/12C abundance ratio of the sample or standard reference material, and δ13C values are reported in units of per mil (‰) (Fry, 2006). Pee Dee Belemnite (PDB) is a carbonate which has a relatively high 13C/12C ratio and is used as the international standard reference material with δ13C = 0, per definition (Eq. 3.4). Natural samples with a lower abundance of 13C compared to PDB have negative δ13C values, with increasing 13C depletion corresponding to increasingly negative δ13C values and vice versa (Craig, 1957). When referring to CH4, the C isotopic composition is denoted by δ13C-CH4. Considering biogenic CH4 production, precursor C used in the process (i.e. organic matter) initially has δ13C values between –32‰ and −22‰ (Whiticar, 1996). In the methanogenic process, microbes preferentially utilize precursor compounds containing 12C to produce 12CH4 (Whiticar, 1996). This fractionation process results in CH4 that is characteristically depleted with respect to the heavy isotope and has δ13C-CH4 values < –50‰ (Hornibrook et al., 2000a; Whiticar, 1996). Contrary to this isotopically depleted CH4, there is little fractionation that occurs when CH4 is 90 formed from thermogenic processes, and thermogenic CH4 is therefore less depleted in 13C than biogenic CH4 (Schoell, 1980). Thermogenic CH4, including that produced in the Marcellus Shale, is characterized by values of δ13C-CH4 > –50‰ (Hakala, 2014; Schoell, 1980). Values of δ13C-CH4 between –64‰ and –50‰ may represent a mixture of thermogenic and biogenic CH4 (Schoell, 1980). 3.1.5 Methane behavior in relation to fracking Methane contamination has recently been explored in regions of fracking activity in view of the potential for CH4 flow between deep formations, groundwater, and surface water. In regions of the Marcellus Shale, analysis of well water CH4 has been carried out in areas near active fracking wells (within 1 km) and in areas outside of the fracking areas (Jackson et al., 2013a; Osborn et al., 2011). The authors found that average well water CH4 concentrations within active fracking areas were significantly higher than those away from fracking. It was also found that shallow groundwater near fracking wells with higher CH4 concentrations showed higher values of δ13C-CH4 than groundwater away from wells, suggesting a greater prevalence of thermogenic CH4. In contrast, a similar study in the Fayetteville Shale region found no difference between groundwater CH4 concentrations within and outside of active fracking areas (Warner et al., 2013b). Such studies comparing dissolved CH4 of active and inactive fracking areas are contested, though, since CH4 may be found in groundwater irrespective of fracking activity (Molofsky et al., 2013; Saba and Orzechowski, 2011). Baseline groundwater analyses of >1,700 wells in Pennsylvania (PA) regions where fracking has not 91 occurred found dissolved CH4 in 78% of water wells (Molofsky et al., 2013). Moreover, in several non-fracking regions of PA and New York (NY), concentrations of CH4 are relatively high, between 0.623−4.42 mM (10−71 mg/L) (Kappel and Nystrom, 2012; Moore and Buckwalter, 1996; Sloto, 2013), in the range of groundwater concentrations measured by Osborn et al. (2011). As pre-drilling groundwater CH4 data was not available in the study areas investigated by Osborn et al. (2011) and Jackson et al. (2013a), it is difficult to affirm that CH4 concentrations in these areas are elevated due to fracking. To further investigate the relationship between fracking activity and CH4 concentrations of natural waters, particularly since baseline concentration measurements were not available in studies to date, stable C isotope analysis has been used. In several analyses of groundwater wells throughout PA, δ13C-CH4 values ranged between –50‰ to –30‰, consistent with thermogenic CH4 from either Upper Devonian or Middle Devonian (Marcellus) formations (Jackson et al., 2013a; Molofsky et al., 2013; Osborn et al., 2011). Although Upper Devonian CH4 in groundwater has been linked to upward migration over geologic time in some regions (Vengosh et al., 2014), its migration may be associated with fracking in certain cases. It was found that some groundwater wells in PA located <1 km from fracking wells were elevated in dissolved CH4 and showed Upper Devonian isotopic signatures. These water samples also showed proportions of C2H6 and 4He that are inconsistent with migration over geologic time (Darrah et al., 2014). Alternative to migration of Upper Devonian CH4, some groundwater δ 13C-CH4 measurements overlying the Marcellus Shale suggest the presence of deeper 92 Marcellus Shale CH4, showing values between –45‰ and –30‰ characteristic of this reservoir (Hakala, 2014; Sharma et al., 2014). In one subset of groundwater wells in PA within 1 km of fracking wells, gas analysis showed δ13C-CH4 values within the range of –45‰ and –30‰ and also showed C2H6 isotopic signatures characteristic of Marcellus Shale gas (Jackson et al., 2013a). The diagram of a fracking well in Fig. 3.1 shows proposed CH4 flowpaths linked to fracking that may allow thermogenic CH4 to intrude groundwater and eventually surface waters. Arrow A in this figure indicates leakage through poorly joined casing near groundwater reservoirs, a pathway which has been suggested as an explanation for findings of characteristic Marcellus Shale CH4 in groundwater, as in PA (Darrah et al., 2014). Arrow B indicates migration of Upper Devonian CH4 facilitated by leaky well annuli, consistent with findings of Upper Devonian CH4 in groundwater in proximity to fracking (Darrah et al., 2014; Vengosh et al., 2014). Well annuli, the spaces between well casing layers, are typically filled with cement, although seal failures or cement permeability can occur, allowing Upper Devonian CH4 to be transported to shallow groundwater (Brantley et al., 2014; Darrah et al., 2014). Geochemical and isotopic evidence for CH4 transport via annuli has been found in regions of conventional gas drilling in North America (Harrison, 1983; Rowe and Muehlenbachs, 1999). Another potential mode of CH4 contamination is migration from the deep shale formation through newly initiated fractures, represented by Arrow D, although this mechanism is thought to be of low likelihood because of the thickness (1–2 km) of overlying formations (Osborn et al., 2011). Fracking may also Figure 3.1 Diagram showing a fracking well drilled into the Marcellus Shale, with red arrows indicating mechanisms for CH4 contamination of water. Arrow A represents leakage of Marcellus Shale CH4 through failed wellbore casing and B represents transport of intermediate formation CH4 through annuli. Arrow C represents transport of CH4 from shallow groundwater to streams. Arrow D represents a potential pathway of Marcellus Shale CH4 from depth and E represents CH4 migration from depth facilitated by pre-existing faults. 94 stimulate upward CH4 migration along pre-existing fault systems which are prevalent in many regions of the Marcellus Shale (Osborn et al., 2011). This pathway is represented by Arrow E in Fig. 3.1. Arrow C in Fig. 3.1 represents flowpaths by which CH4 is transported from groundwater to streams (Heilweil et al., 2013), such that each of the indicated pathways facilitating groundwater contamination may ultimately cause stream contamination. With the known flow of CH4 between groundwater and surface waters, monitoring of stream CH4, as is carried out in this study, is important since shifts in stream concentrations may indicate groundwater as well as surface water contamination following fracking activity. Notably, two major modes of CH4 contamination associated with fracking—transport through annuli (Arrow B) and leakage through failed casings (Arrow A)—are caused by defective well structures. Such structural problems have been documented at a rate of 1–2% for fracking wells in PA (Considine et al., 2012), although an even greater rate of 3.4% was indicated from a survey by the PA Department of Environmental Protection (DEP) (Vidic et al., 2013). Construction rates of fracking wells are high in the Marcellus Shale region, with 60,000 wells projected to be built in the next 30 years (Entrekin et al., 2011), so the problem of substandard well construction offers many opportunities for CH4 migration and contamination of natural waters. 3.1.6 Monitoring baseline dissolved CH4 in western and southern MD streams Given the connectivity of deep CH4 reservoirs with groundwater and surface water as depicted in Fig. 3.1, as well as the documented structural challenges of wells, 95 there are many opportunities for CH4 contamination associated with fracking. This contamination may be conflated, however, with in situ production of CH4 in groundwater and streams. It is therefore necessary to determine the natural distribution of surface and groundwater CH4 in areas surrounding fracking operations and to determine predominant CH4 sources. In the current study, I pursue this by monitoring baseline stream CH4 concentrations and C isotope compositions in rural catchments slated for fracking development. The first goal of this work was to determine temporal and spatial variability of stream CH4 concentrations overlying the Marcellus Shale. This allowed for establishment of baseline CH4 concentrations in western MD. Determining variability in CH4 concentrations also allowed for assessment of the viability of stream CH4 as an indicator of fracking contamination. In assessing temporal and spatial variability of stream CH4, concentrations were also measured in MD’s coastal plain region, allowing comparisons on a larger spatial scale. It has been shown in some laboratory and field studies that methanogenic activity, which can occur in soils and sediments surrounding streams (Jones and Mulholland, 1998b), is elevated at higher temperature (Kelly and Chynoweth, 1981; Zeikus and Winfrey, 1976). I therefore hypothesized that, temporally, CH4 concentrations would be greatest in summer and lowest in winter. Inherent in this projection is the hypothesis that CH4 in western MD streams is primarily biogenic, which was reasoned based on previous works finding methanogenesis to be a predominant CH4 source in freshwater systems (Jones et al., 1995; Jones and Mulholland, 1998a). 96 Previous research has also shown that methanogenesis occurring in anoxic riparian sediments and stream bottom sediments is a major source of stream CH4 (Jones and Mulholland, 1998b), and this led to the hypothesis that stream CH4 concentrations vary spatially based on the observed environmental setting of the stream. Forested stream sites were expected to be lowest in CH4 since bottom sediments are typically coarse with little organic matter, and water flow at these sites is relatively fast. Sites near anthropogenic influences, such as roadways and agricultural or residential areas, were expected to be higher in CH4 than forested sites since these streams likely receive higher amounts of allochthonous organic matter and nutrients which stimulate in-stream methanogenesis (Sanders et al., 2007). Sites in wetland-type settings were expected to be highest in CH4 due to an abundance of organic matter which facilitates high productivity and anoxic conditions in sediments adjacent to and within the streams. Also, wetland sites have relatively slow streamflow allowing lower rates of CH4 evasion and more widespread development of anoxia (Jones et al., 1995; Wallin et al., 2014). Although the stream sites in this study are spatially divided amongst five major catchments, it was expected that stream CH4 concentrations are more closely tied to the immediate landscape surrounding the site rather than larger spatial divisions such as catchment, particularly as land use and landscapes are heterogeneous within catchment areas (Appler, 2010). Considering large-scale variability across physiographic regions of MD, it was hypothesized that dissolved CH4 in coastal plain freshwater streams is primarily biogenic, as in western MD, although it was expected that concentrations in this region of MD are higher than those in western MD due to features of the coastal plain 97 streams. Streams in this region are characterized by finer sediments, with lower stream gradients which allow greater rates of sediment deposition (Paul, 2002; Roth, 1999); these characteristics then stimulate greater rates of methanogenesis relative to streams of western MD. To assess temporal variability, with the hypothesis of increased CH4 concentrations in summer and lower concentrations in winter, concentration measurements were taken across a period of 1 year in western MD, and stream samples were collected every 4–8 weeks. To evaluate spatial variability of dissolved CH4, concentrations were regularly monitored in 25 streams throughout western MD and the coastal plain region of Calvert County, with all streams characterized by either a forested, wetland, or anthropogenically influenced setting. Stream CH4 concentrations measured in both western MD and Calvert County were compared with concentrations reported for other North American headwater streams which do not overlie shale formations. The second major goal of this portion of the study was to verify the source of stream CH4, either biogenic or thermogenic, in both western MD and Calvert County, with the use of carbon isotopic analysis of CH4. Measurements of carbon isotopic composition of CH4 may distinguish biogenic and thermogenic CH4 in streams, and they may also demonstrate baseline δ13C-CH4 values in these settings. This is of particular importance in a region before the onset of fracking since fracking contamination may stimulate increased flow of thermogenic CH4 to streams. Stream CH4 in both MD regions was hypothesized to be primarily biogenic based on previous studies of CH4 in freshwater systems (Jones et al., 1995; Sansone 98 et al., 1999). To test this hypothesis, a subset of samples from the western MD and Calvert County regions was analyzed for C isotopic composition. Isotopic analysis was carried out for sites YOUG105D and YOUG106D, which were assessed across all seasons, and for Aug–Oct samples which were measured across all sites the region. As CH4 was presumed to be primarily biogenic, δ 13C-CH4 values were expected to be between –110‰ and –50‰ (Whiticar, 1996). Monitoring of both CH4 concentrations and C isotopic compositions across spatial and temporal ranges in this study increases understanding of natural behaviors of CH4 in streams which connect to both groundwater flow paths and major downstream river systems (Roth, 1999). This will enable detection of shifts in concentrations and C isotopic compositions of CH4 should CH4 migration or contamination occur in the future. 3.2 Materials and methods 3.2.1 Field sampling methods Water samples were collected from streams near potential fracking wells in Garrett and Allegany Counties in western MD. All samples were collected during base flow, which was defined as normal flow conditions for each season, not during high-flow conditions or following storm events. Sampling occurred in the five catchment areas delineated on the map in Fig. 2.1: Casselman River, Georges Creek, Savage River, Potomac River Upper North Branch, and Youghiogheny River. Groundwater from the Savage River Pumping Station (SRPS) and water from the Frostburg Water Treatment Plant (which blends SRPS water with intake from Piney 99 Creek Reservoir within the Casselman River catchment) were also monitored. As described for the metals analyses in Chapter 2, samples for CH4 were collected in three separate sets: two in western MD and one in southern MD. The western MD sites are shown on the map in Fig 2.1. The first set includes eight sites in Garrett and Allegany Counties that were sampled monthly during the period Sep 2012–Feb 2013. The second set includes fourteen sites in Garrett County that were sampled every 6–8 weeks during the period Apr 2013–Apr 2014. Also within this set are five sites sampled biannually, once in winter and once in summer, between Apr 2013–Apr 2014 (PRUN301D, SAVA302D, YOUG201D, YOUG301D, and YOUG302D). Samples in this set were collected with assistance from the Maryland Department of Natural Resources (DNR), who obtained landowner permission for site access, where necessary. The third set includes seven sites in Calvert County which were sampled in June and July of 2014. Sites are found within the West Chesapeake Bay and the Patuxent River Lower catchments. The seven stream sites sampled in Calvert County are shown on the map in Fig 2.2. In both western MD and Calvert County sample sets, streams were found in environments that I designated as either forested, wetland, or anthropogenically influenced, where anthropogenically influenced sites were adjacent to features such as major roads, residential areas, lawns, or industrial activities. Before each field trip, 8 M KOH solution (Sigma-Aldrich) for sample preservation was made in the lab at CBL and allowed to equilibrate with air for one day. Within 24 hours of sampling, 10-mL syringes were filled with CH4-free air (‘zero air’; Airgas) slightly past the 10-mL mark and closed with a 3-way stopcock. 100 These syringes were used to add headspace to the CH4 sample bottles in the field. One syringe was prepared for each sample bottle, with some spares, in addition to another ten to be used as blanks. The blank syringes were brought to the field sites and back with the other equipment, remaining securely closed throughout, and subsequently analyzed for CH4 content. In a preliminary test, syringes were exposed to field conditions and showed negligible syringe intrusion of atmospheric CH4 during 60 hours of exposure. For collection of CH4 samples in the field, 125-mL glass serum bottles (Wheaton) were first carefully rinsed with stream water three times. The vial was then submerged and filled to the rim with no headspace, avoiding bubbling while filling to prevent entrainment of ambient air in the sample. The sample was preserved by adding 0.5 mL of 8 M KOH solution into the bottom of the sample bottle while letting excess sample overflow the top. Next, a 1-cm thick black butyl rubber stopper (Geo-Microbial Technologies) pierced with a steel needle attached to a 3-mL syringe was inserted firmly into the top of the sample bottle, taking care not to introduce an air bubble. Excess sample was simultaneously expelled into the syringe, which was subsequently removed and drained into a waste container. Finally, an aluminum cap was crimped over the stopper to secure it and make a gas-tight seal. Triplicate samples were collected at each site. After collection of samples at each site, a 10-mL headspace of zero air was added to each bottle at ambient pressure (1 atm) for equilibration with the water sample and subsequent analysis of CH4 content. This addition prevented bottles from breaking due to thermal expansion during transport. To ensure the injection of exactly 10 mL, 101 each syringe was first tested for leaks by applying pressure to the plunger with the stopcock closed. The stopcock was then opened, and the plunger pushed in to the 10-mL mark, using the excess zero air to flush the needle before inserting it into the bottle through the stopper. Another steel needle was inserted through the stopper to drain excess sample into the waste container while the headspace was injected. After removing the syringe and second needle, sample bottles were stored upside down in a cooler during transport from the field to minimize contact of the headspace with the stopper. To assess the small-scale spatial variability in CH4 concentrations at a single site, five samples were collected within a 50 m reach upstream and downstream of CASS101D, in Sep 2013. Triplicate samples were taken at CASS101D, as well as 10 m and 20 m upstream, and 20 m and 30 m downstream. For Calvert County sites, in view of their proximity to the Chesapeake Bay, the salinity of all water samples was tested using a refractometer and found to be <1 everywhere, indicating the absence of any seawater influence. 3.2.2 Analysis of CH4 partial pressure in the headspace Field samples were returned to the lab for analysis of dissolved CH4. An additional 10 mL of zero air was added to each bottle, without displacing sample, increasing the headspace pressure to 2 atm and enabling subsequent removal of headspace gas for analysis. The bottles were vigorously shaken for 2 minutes to re-equilibrate the headspace with the water. For analysis of CH4 partial pressure in the headspace by gas chromatography (GC), 8 mL of headspace gas were extracted 102 into an empty syringe through a steel needle at ambient pressure (1 atm). Before extraction the plunger was pumped 10 times to ensure good equilibration between the headspace and the syringe. Contents of the syringe were analyzed on an SRI 8610C GC, equipped with a HayeSep D column (1.83 m, 3.2 mm ID) and flame ionization detector (FID). The limit of detection, LD, for CH4 analysis by GC-FID was 0.85 nM, calculated as: lD 3.3L  (3.5) where σ1 is the standard deviation (nM) of the low concentration standard using the GC-FID analysis (Magen et al., 2014). The relative precision of the instrument from repeated measurements of a CH4 standard gas was found to be 4.5%. 3.2.3 Theory According to Henry’s Law, the partial pressure of CH4 in equilibrium with the solution is proportional to the amount of CH4 dissolved in the solution, so the CH4 concentration in the water sample (mol/L) may be determined by measuring the CH4 partial pressure in the headspace (atm) with a gas chromatograph (GC). This partial pressure, 4CH HSF (ppmv), is determined from the area of the CH4 peak in the gas chromatogram by comparison with a gas standard of known CH4 partial pressure. The ideal gas law is used to convert 4CH HSF into the number of moles of CH4 in the headspace, HSCH4n , as follows: 4 4 CH HS HS HS CH HS F P Vn R(273.15 T)    (3.6) 103 where T is temperature (°C), R is the ideal gas constant (0.08206 L∙atm∙K-1∙mol-1), VHS is the headspace volume, and PHS is the headspace pressure (2 atm). The number of moles of CH4 remaining in the water sample after headspace equilibration, wCH4n , is calculated from the Bunsen coefficient, β, for CH4 at known temperature, pressure, and salinity (Yamamoto et al., 1976): 4 4 CH HS HS w CH w F P Vn ȕ R(273.15 T)    (3.7) where Vw is the sample volume (L). The moles of CH4 in the water, wCH4,n , may then be added to the moles of CH4 in the headspace, HSCH4,n , in order to determine the total moles of CH4 in the water sample CH4n at the time of sample collection. This is divided by the sample volume, Vw , to calculate the concentration of CH4 dissolved in the water, [CH4]: 9 w CH 4 10V n][CH 4  (3.8) where the factor 109 yields units of nmol/L (nM). Two of the triplicate samples taken at each time and site were analyzed for CH4 and the concentrations were averaged. The relative spread of sample duplicates was ≤ 8% for most measurements. 3.2.4 Calculation of stream CH4 saturation The saturation of stream waters with CH4 was determined as the saturation ratio S based on the dissolved CH4 concentrations. For each sample, S was determined from the equation: 104     4 4 Eq CHS = CH (3.9) where [CH4] is the measured CH4 concentration (nM) of the sample, and [CH4]Eq is the stream concentration (nM) of CH4 in equilibrium with the atmosphere. Stream CH4 concentrations in equilibrium with the atmosphere were determined using the Bunsen coefficient, β, for CH4 in water temperature 10°C, 1 atm pressure, and salinity 0 (Yamamoto et al., 1976), according to the following equation: 4 Atm CH ,Eq ȕ &n = 1000R(273.15+ T)   (3.10) where Catm is the atmospheric CH4 concentration, estimated to be 1.89 ppm (Blasing, 2014), T is the stream temperature (°C), and R is the ideal gas constant (0.08206 L∙atm∙K-1∙mol-1). Western MD streams were calculated using T=10°C, the average stream temperature across all sampling dates, and in Calvert County stream temperature was estimated as 25°C for the calculation. 3.2.5 Analysis of δ13C-CH4 After CH4 concentration analysis, a subset of samples was analyzed for δ 13C-CH4 in the isotope ratio mass spectrometry (IRMS) laboratory of Dr. Jeff Chanton at Florida State University, Department of Earth, Ocean, and Atmospheric Science. A total of 33 samples were analyzed for δ13C-CH4, all of which were >34 nM in CH4 concentration, which is sufficiently high for IRMS analysis. Two sites, YOUG105D and YOUG106D, which show consistently high CH4 concentrations relative to other sites, were analyzed for δ13C-CH4 across all six time points for assessment of 105 temporal variability of δ13C-CH4 (12 samples). As most sites reach the maximum CH4 concentration in summer and fall, samples from summer and fall time points of 15 other sites were also analyzed for δ13C-CH4 (21 additional samples). This allowed for determining spatial variability of δ13C-CH4 within the region. In the samples analyzed for carbon isotopic composition, 10 mL zero air was again added to each sample vial to allow the removal of headspace gas for δ13C-CH4 analysis. For water samples of CH4 concentration <1690 nM, (<100 ppmv CH4 in the headspace after the final zero-air addition), headspace gas was first injected into a preparative chromatographic column (stainless steel, 20 cm length, 1/8-inch OD, 1.5 mm ID; 80-100 mesh Hayesep D) cooled to –118°C to retain CH4 and CO2 and remove N2, O2, and Ar. After warming the preparation column, sample gas was transferred by He carrier gas to the analytical column, a porous-layer open tubular (PLOT) column (PoraPLOT Q, 25 m length, 0.32 mm ID; 10 µm particle layer) for gas separation. The gas stream flowed to a reactor containing NiO for combustion of CH4 to CO2, and effluent from the combustion was dried in a semi-permeable Nafion® membrane tube. Capillary tubing introduced effluent from the Nafion® tube to the IRMS. For samples with CH4 concentrations >1690 nM, the preparative column was not required, and the first 2 m of the PLOT column were cooled temporarily for trapping of CH4 and CO2 before separation (Merritt et al., 1995). Ion currents were measured for CO2 species at m/z 44, 45, and 46, with m/z 44 representing the abundance of 12CO2 and m/z 46 representing 12C18O16O. The known relative abundance of 17O to 18O was used to determine contributions of 13C16O2 and 12C16O17O to m/z 45. Relative abundance of 13C in the CH4 sample was compared to 106 relative 13C abundance in the standard material Pee Dee Belemnite (PDB) and was expressed in δ notation as defined in Equation (3.4), where RStandard, the known 13C/12C ratio of the PDB standard material, is equal to 1.124×10–2 (Craig, 1957). Values of δ13C-CH4 are reported in units of per mil (‰). Reported values of δ13C-CH4 are the mean of sample duplicates. Relative analytical precision based on repeated IRMS measurements of a standard was 0.3‰. The relative spread of sample duplicates was ≤ 2% for most δ13C-CH4 measurements. 3.3 Results The results of measured CH4 concentrations in all western MD sampled streams are shown in Appendix B, Table B1. Concentrations range from 1.4 nM to 5,100 nM and show both spatial and temporal variability. Results are discussed below within the context of the hypotheses posed in the introduction. 3.3.1 Temporal Variability Western MD streams of both the Sep 2012–Feb 2013 and the Apr 2013–Apr 2014 sets are variable with time (Table B1). Eight representative sites illustrate this in Figure 3.2, with four of relatively high concentration (Fig. 3.2, panels A and C) and four of relatively low concentration (Fig. 3.2, panel B and D) among western MD streams. Panels A and B of Fig. 3.2 show sites of the Sep 2012–Feb 2013 sampling set. Of samples collected from Tributary G8, CH4 concentrations reach as high as 5,100 nM in September and as low as 100 nM in February (Fig. 3.2, panel A). The highest concentration reached in the Reservoir is 4,000 nM in Oct (Fig. 3.2, panel A). 107 Among low concentration sites of this sample set (Sep 2012–Feb 2013), SRPS was found to reach a maximum concentration of 42 nM in November, with a minimum concentration of 6 nM, as measured in January (Fig. 3.2, panel B). In Tributary B7, the maximum concentration (142 nM) also occurs in November, although the minimum, 66 nM, was measured in September. Samples were not collected in the summer months at these sites. For the Apr 2013–Apr 2014 sample set, represented in Fig 3.2 panels C and D, all seasons were included in the sampling campaign. In panel C the highest concentrations for YOUG105D were found in October (3,820 nM), yet in June for YOUG106D (2,820 nM, Fig. 3.2, panel C). Concentrations were nearly an order of magnitude lower in the Casselman (CASS) catchment but showed similar patterns over time (Fig. 3.2, panel D). Concentrations in CASS101D were highest in June (373 nM) and lowest in December and April when concentrations were between 29–40 nM. Similarly, CASS103D shows highest concentrations in June (453 nM) and lowest in April (85 nM) (Fig. 3.2, panel D). In Calvert County, all sites vary in concentration between June and July, apart from GRAY02 and PARK01 which are stable between the two months (Fig. 3.3). Most sites show an increase in CH4, with concentrations increasing by 25–515% between June and July. Site THOM02 shows a decrease in CH4 concentration by 22% between June and July (Fig. 3.3, Table B3). 3.3.2 Spatial variability Spatial variability was assessed by determining average CH4 concentrations 2013 YOUG105D YOUG106D 0 100 200 300 400 500 CASS101D CASS103D Reservoir Tributary G8 0 100 200 300 400 500 SRPS Tributary B7 C H 4 C o n c e n t r a t i o n ( n M ) C H 4 C o n c e n t r a t i o n ( n M ) A M J J A S O N D J F M A A M J J A S O N D J F M A A M J J A S O N D J F M AA M J J A S O N D J F M A 2013 2012 2012 2014 2014 2013 2013 5,000 4,000 0 3,000 2,000 1,000 5,000 4,000 0 3,000 2,000 1,000 A B C D Figure 3.2 Temporal variability in western MD stream CH4 concentrations (nM) across 1 year. Top panels (A and B) show sites from the 2012–2013 set, with high CH4 concentration sites shown in A and low concentration sites in B. Bottom panels (C and D) show sites from the 2013–2014 set, with high concentration sites shown in C and intermediate concentration sites in D. SRPS indicates the Savage River Pumping Station. Note that the scale in panels B and D is an order of magnitude lower than in A and C. Bar color indicates site catchment, with yellow indicating Casselman River, blue Savage River, and green Youghiogheny River. Error bars indicate half of the difference between the sample duplicates, <8% of the concentration value in most cases. 109 G R A Y 0 2 T H O M 0 2 P A R K 0 2 G R A Y 0 1 T H O M 0 1 H E L L 0 1 P A R K 0 1 June 2014 July 2014 Wetland Anthropogenic C H 4 C o n c e n t r a t i o n ( n M ) Forested 35,000 5,000 25,000 15,000 0 Figure 3.3. Methane concentrations (nM) in Calvert County from June and July 2014 measurements, grouped by stream setting (forested, wetland, or anthropogenically influenced). among streams of the three different environmental settings: forested, wetland, and anthropogenically influenced. Averages were statistically compared to each other (Kruskal-Wallis test) and it was found that concentrations in wetland and anthropogenically influenced sites are not significantly different (p>0.05), but both of these environmental settings were found to be significantly greater than CH4 concentrations at forested sites (p<0.05). Among all wetland sites, concentrations range from 55−5,100 nM and among anthropogenically influenced sites from 4−4,047 nM. For all measurements at forested sites the concentration range is 4−373 nM. Spatial variability both within and among these stream settings is demonstrated in Fig. 3.4, which shows Aug/Sep concentrations in western MD. 110 While CH4 concentrations varied over all the sites visited, there was also small-scale variability within a 50-m stream reach where longitudinal measurements were made (Table B4, Fig. 3.5). In September 2013, CH4 concentrations were 170 nM at the center location, indicated as 0 m in Fig. 3.5. Twenty meters downstream, values reached 274 ± 40 nM and 10 m further downstream decreased again to 179 nM. Upstream, concentrations are slightly lower but do not vary between the +10 m and +20 m points, with concentrations of 125 ± 2 nM and 124 ± 1 nM, respectively. The average value of these 5 points in the stream reach is 174 ± 60 nM with a relative standard deviation of 34%. This average value is well within the range of concentrations occurring at this site over the 1-year study period, between 30–301 nM (Table B1). Methane concentrations in coastal plain streams of Calvert County, when compared with those in western MD, show that large-scale spatial variability is also observed. Concentrations are elevated relative to western MD and range between 2,100–33,000 nM (Table B3, Fig. 3.3). As in western MD, Calvert County CH4 concentrations vary across different environmental settings, as observed for July 2014 measurements which are shown in Fig. 3.4. On average the CH4 concentration in wetland sites is 23,904 ± 8,760 nM and that in anthropogenically influenced sites is 4,559 ± 1,940 nM. The average concentration in forested sites is 4,581 ± 4,696 nM. Although concentrations differ across the settings, concentration differences between the three setting types in Calvert County are not significant (p>0.05). C A S S 1 0 1 D G E O R 1 0 1 D S A V A 1 0 1 D S A V A 2 0 2 D S A V A 2 0 4 S S A V A 3 0 1 D Y O U G 1 0 2 D Y O U G 1 0 3 D Y O U G 3 0 1 D Y O U G 4 3 2 S Y O U G 1 0 4 D Y O U G 1 0 6 D T r i b . B 7 T r i b . G 8 R e s e r v o i r C A S S 1 0 3 D C A S S 1 0 5 D P R U N 3 0 1 D S A V A 3 0 2 D Y O U G 1 0 5 D Y O U G 2 0 1 D Y O U G 3 0 2 D 0 1000 2000 3000 4000 5000 Forested Wetland Anthropogenically influenced AnthropogenicWetlands Forested Appalachian Highlands G R A Y 0 2 T H O M 0 2 P A R K 0 2 G R A Y 0 1 T H O M 0 1 H E L L 0 1 P A R K 0 1 0 10000 20000 30000 40000 Wetland Anth. Coastal Plain Forested A B C H 4 c o n c e n t r a t i o n ( n M ) Figure 3.4 Spatial distribution of CH4 concentrations (nM) according to site setting, with sites characterized as forested, wetland, or anthropogenically influenced. Concentrations measured in Aug/Sep sampling in western MD (of both the 2012−2013 and 2013−2014 sets) are shown in (A) and those measured in July in Calvert County shown in (B). Concentrations shown in B are the same as those shown in Fig. 3.3 for July 2014. Note the order of magnitude higher concentrations in (B) which shows Calvert County measurements relative to (A) which shows those of western MD. Error bars indicate half of the difference between duplicate samples, <8% of the concentration values for most measurements. 112 CH 4 Longitudinal Variability: CASS101D –30 m –20 m 0 m +10 m +20 m C H 4 C o n c e n t r a t i o n ( n M ) 0 50 100 150 200 250 300 350 Figure 3.5 Longitudinal variability in dissolved CH4 (nM) along the CASS101D stream reach in Sep 2013. Sampling occurred at CASS101D (0 m), 2 upstream locations (+10 m, +20 m), and 2 downstream locations (−20 m, −30 m). Error bars indicate half of the difference between duplicate samples. Relatively high variability in dissolved CH4 is observed among streams within each of the three settings. The highest range in concentrations across both June and July is observed for wetland sites, with concentrations ranging from 11,820−32,748 nM. Anthropogenically influenced sites in Calvert County range between 2,168−6,220 nM and forested sites between 2,307−14,138 nM over the two months. Concentrations of each of the three settings in Calvert County are significantly greater than those of these stream settings in western MD (p<0.05). 113 3.3.3 Carbon isotopic composition of CH4 Across the Sep 2012–Feb 2013 and Apr 2013–Apr 2014 sampling sets, δ13C-CH4 values range from –60‰ to –44.3‰, with >80% of values distributed between –60‰ and –50‰ (Table B2). Due to the extensive nature of this dataset, the isotope time series is visualized by examining 2 sites (YOUG105D and YOUG106D) which were analyzed with highest temporal resolution for δ13C-CH4. The YOUG106D δ 13C-CH4 value was −54‰ in April 2012 (Fig. 3.6). This increased to −44‰ by October and then quickly decreased to −58‰ in December 2013 and April 2014 (Figure 3.6, Table B2). Values for YOUG105D showed less fluctuation and ranged between −59‰ and −52‰ over the same time period.    C - C H 4 ( ‰ ) A A M J J OS N D J F M A -62 -60 -58 -56 -54 -52 -50 -48 -46 -44 -42 YOUG105D YOUG106D A Figure 3.6 Measurements of δ13C-CH4 values (‰) for sites YOUG105D and YOUG106D across time (Apr 2013–Apr 2014). 114 Spatial variability in δ13C-CH4 is comparable to temporal variability. At one time point (e.g. Oct 2013; Fig. 3.7), δ13C-CH4 values span the range of –60‰ to –44.3‰, while at other time points (e.g. Apr 2013, Table B2), δ13C-CH4 values across the region range by only 5‰, with values between –52.1‰ and –57.0‰. Although CH4 concentrations in this region were found to vary based on setting of the stream site, δ13C-CH4 values do not vary according to stream setting (Fig 3.7) (Kruskal-Wallis test, p>0.05). Average δ13C-CH4 values in forested, wetland, and anthropogenically influenced settings are −54 ± 4‰, −52 ± 5‰, −54 ± 2‰, respectively. In Calvert County, spatial variability of δ13C-CH4 was assessed based on environmental settings of streams as in western MD. In forested, wetland, and anthropogenically influenced sites, average δ13C-CH4 values in this region are −53 ± 1‰, −44 ± 2‰, and −55 ± 2‰, respectively. Although wetland sites show the highest values, differences in average δ13C-CH4 values of the three settings are not significant (Kruskal-Wallis test, p>0.05). Measurements from the three settings in Calvert County do not differ significantly from those of any western MD stream settings (p>0.05). For the stream site PARK02, measurements of δ13C-CH4 were made in both June and July, and the two values were in agreement within the standard deviation of the measurements (Table B3). Overall, in Calvert County streams δ13C-CH4 values range between –56.7‰ and –43‰ (Table B3), within the range of values observed in western MD streams (–60‰ to –44.3‰), despite the significantly elevated CH4 concentrations in Calvert County    C - C H 4 ( ‰ ) C A S S 1 0 1 D S A V A 1 0 1 D S A V A 3 0 1 D T r i b T 1 Y O U G 1 0 6 D T r i b B 7 T r i b C 6 T r i b G 8 R e s e r v o i r C A S S 1 0 3 D C A S S 1 0 5 D Y O U G 1 0 5 D -80 -70 -60 -50 -40 -30 M a r c e l l u s C H 4 B i o g e n i c C H 4 Forested Wetland Anthropogenic Western MD   C-CH 4 values, Oct A T H O M 0 2 G R A Y S 0 2 P A R K 0 2 G R A Y S 0 1 T H O M 0 1 H E L L 0 1 P A R K 0 1 -80 -70 -60 -50 -40 -30 Forested Wetland Anthropogenic Calvert Co   C-CH 4 values, July B Figure 3.7 δ13C-CH4 values (‰) in Oct samples in western MD streams (A) and July samples in Calvert County streams (B), separated by stream setting (forested, wetland, or anthropogenically influenced). The range of δ13C-CH4 values characteristic of biogenic CH4, –110‰ to –50‰ (Whiticar, 1996), is indicated by the green bar. The range characteristic of Marcellus Shale CH4, from –43‰ to –27‰ (Hakala, 2014), is indicated by the gray bar. 116 relative to western MD. As in western MD, the majority of values are distributed between –60‰ and –50‰. 3.4 Discussion Dissolved CH4 in streams shows considerable spatial and temporal variability in the Marcellus Shale region of MD, within both the 2012–2013 and 2013–2014 sampling sets. Although variability was observed, CH4 concentration ranges (from 1.4–5,100 nM) are consistent with those reported for other North American temperate streams of relatively undisturbed catchments. In a long-term study in western Oregon streams, De Angelis and Lilley (1987) report spatial variability across stream environments, with CH4 concentrations ranging between 5–1,800 nM. In the northeastern U.S., naturally occurring dissolved CH4 was found at concentrations between 580–2,440 nM in predominantly forested or mixed forested and agricultural catchments (Sansone et al., 1999). De Angelis and Lilley (1993) found CH4 concentrations in the Upper Hudson River to range between 50–950 nM. Consistent CH4 oversaturation with respect to the atmosphere was found in streams of western Oregon as well as in those of the northeastern U.S. (De Angelis and Lilley, 1987; De Angelis and Scranton, 1993; Sansone et al., 1999). Similar to the previous studies, all western MD streams were found to be at saturation or oversaturated with CH4 with respect to the atmosphere, with saturation ratios S ranging from 1–1,075. All streams, therefore, provide a CH4 flux to the atmosphere. Western MD streams, ranging from 0−5,100 nM with 92% of measurements <1,000 nM, show a low concentration range relative to groundwater CH4 117 measurements in the northeastern U.S. Analyses from PA, West Virginia, MD, and NY show groundwater CH4 concentrations ranging from 0−4,400,000 nM, with 80% of measurements between 0−311,720 nM (Brantley et al., 2014; Vidic et al., 2013). Western MD stream CH4 is also very low in comparison to U.S. caution concentrations and action concentrations for CH4 in drinking water, 0.623 mM (10 mg/L) and 1.76 mM (28 mg/L), respectively (Vidic et al., 2013). 3.4.1 Temporal variability of stream CH4 Temporally, concentrations ranged by as much 5,000 nM at a single site (Table B1). All sites show lowest concentrations in winter and reach maximum concentrations in summer or fall (Fig. 3.2), consistent with my hypothesis that concentrations would be highest when temperature and microbial activity are highest. Previous works similarly report seasonal shifts in dissolved CH4, with elevated concentrations in warm months which is attributed to biogenic CH4 sources in anoxic bottom sediments and riparian soils near streams (De Angelis and Scranton, 1993; Fedorov et al., 2003; Jones et al., 1995). In lab experiments, anaerobic microbial CH4 production in peat was found to correlate with ambient temperature (Dunfield et al., 1993). Freshwater sediment incubation experiments have also shown a positive relation between temperature and rates of CH4 production (Kelly and Chynoweth, 1981; Zeikus and Winfrey, 1976). In the current field study, however, CH4 concentrations showed no direct correlation to water temperature (Fig. 3.8), which may be due to other factors affecting CH4 concentrations. Methanotrophy, a microbial CH4 consumption process, 118 7 H P S H U D W X U H  ƒ & ) CH 4 Concentration (nM) 0 1000 2000 3000 4000 5000 6000 0 5 10 15 20 25 Western MD, 2013–2014 Western MD, 2012–2013 Figure 3.8 Plot of CH4 concentration (nM) vs. water temperature (°C) for the 2012–2013 and 2013–2014 western MD sampling sets. may be one such factor as it responds differently to temperature than methanogenesis (Bastviken et al., 2008; Segers, 1998) hence, with both processes occurring simultaneously in streams, a correlation between CH4 concentration and temperature may not be observed. Inputs of leaf litter and organic debris also affect temporal variability in CH4 and may be another reason that a direct correlation between temperature and CH4 concentration is not observed. Leaf litter and organic debris in stream areas provide a labile carbon source in streambed and bank sediments, accelerating microbial CH4 production (Jones and Mulholland, 1998a). It has been shown that CH4 flux from freshwater lakes is positively related to organic matter input rates (Kelly and Chynoweth, 1981). Research in an agricultural stream catchment similarly indicated a positive link between the abundance of allochthonous organic matter inputs and stream CH4 production (Sanders et al., 2007). The 119 connection between organic matter deposits and CH4 production likely accounts for relatively high CH4 concentrations observed in late summer and fall when leaf litter and other plant debris has accumulated in and alongside streams. Findings of elevated dissolved CH4 in both summer and fall are consistent with temporal patterns observed in a study of hemiboreal streams which showed highest CH4 concentrations in summer and fall periods (Wallin et al., 2014). The temporal concentration patterns in this study support the hypothesis of biogenic production as the primary stream CH4 source. Biogenic production patterns are further supported by C isotopic measurements of CH4 in western MD. Values of δ13C-CH4 are between –60‰ and −44.3‰ in western MD streams (Table B2, Fig. 3.7), which are predominantly in the predicted range (–110‰ and −50‰) for a biogenic stream CH4 source. The δ 13C-CH4 range in western MD suggests that biogenic CH4 may be produced primarily by acetate fermentation, a CH4 production process which occurs in freshwater and is known to give δ13C-CH4 values between –65‰ and −50‰ (Hornibrook et al., 2000a; Whiticar et al., 1986). 3.4.2 Spatial variability of stream CH4 Stream CH4 concentrations are highly variable across western MD. Spatial variability appears to be driven largely by stream setting, as was hypothesized, with streams categorized as either forested, wetland, or anthropogenically influenced. Wetland and anthropogenically influenced streams were significantly elevated in CH4 relative to forested streams. Similarly, a principle component analysis (PCA) based on CH4 concentration, temperature, and flow shows that anthropogenic and wetland 120 sites cluster more closely in the diagram, with forested sites grouped separately (Fig. 3.9), indicating that stream concentrations, temperature, and flow were more similar in the anthropogenic and wetland sites. Figure 3.9 Diagram of Principle Component Analysis (PCA) based on variables CH4 concentration, temperature, and streamflow for western MD samples. Color separates samples by environmental setting. The relation of CH4 concentration to stream setting is reasonable since it has been found that streams typically receive a large portion of dissolved CH4 from the riparian zone through runoff and lateral diffusion (Jones and Mulholland, 1998b). In wetland settings in western MD, streams are found in lowland areas with significant plant matter and inundated soils. These conditions promote anoxia and methanogenesis in the zones surrounding streams and likely contribute to high CH4 concentrations. Also, in the wetland settings, stream beds have more fine-grained sediments with a predominance of organic matter, promoting in-stream 121 methanogenesis (Jones and Mulholland, 1998a). Further, streams of this type have relatively slow flow, contributing to lowers rates of CH4 evasion relative to streams with higher flow. Streamflow rates were not found to be directly related to CH4 concentration, however, (Fig 3.10), implying that a combination of stream characteristics influence concentrations. Flow (ft 3 /s) 0 50 100 150 200 250 300 C H 4 C o n c e n t r a t i o n ( n M ) 0 1000 2000 3000 4000 5000 Figure 3.10 Plot of CH4 concentration (nM) vs. streamflow for the 2013–2014 western MD sampling set. Streamflow values (ft3/s) obtained from each sampling trip were taken from the USGS gauge nearest to each site. Streams influenced by adjacent anthropogenic activities including major roads, residential or recreational developments, and coal extraction facilities show relatively high CH4 concentrations which were in general comparable to those at wetland sites (Fig. 3.4). These anthropogenic influences are expected to be associated with an excess of sediment and nutrients. Roads and highways are major sources of sediments and nutrients to runoff (Forman and Alexander, 1998), as are residential 122 areas (Waschbusch et al., 1993), which are found near some western MD stream sites. One anthropogenically influenced site (CASS103D) is adjacent to a golf course, which likely generates runoff with high levels of nutrients from grass fertilization. Another site (YOUG105D) is just downstream of existing oil and coal facilities, which may provide effluent with high amounts of sediment. In each of the anthropogenically influenced settings, excess sediments and nutrients may be delivered to stream riparian areas, where high rates of respiration and methanogenesis may occur, supporting an abundance of stream dissolved CH4 (Jones and Mulholland, 1998b). High amounts of sediment and nutrients may also be delivered directly to streams, supporting relatively high rates of respiration and methanogenesis in the stream bed (Sanders et al., 2007). Agriculture has been found to contribute to the occurrence of greenhouse gases CH4, CO2, and N2O in streams (Jacinthe et al., 2012); however, the percentage of upstream agricultural land use for western MD sites, as reported by MD DNR (StreamHealth, 2014), was not found to correlate with dissolved CH4 concentrations (Fig 3.11). Areas of agricultural land cover are found throughout the western MD study region (Appler, 2010), and the stream chemistry effects of this land use may be moderated by localized stream settings and lateral flow to streams from the riparian zone, where CH4 concentrations in these flowpaths are closely tied to riparian organic matter deposition and the availability of electron acceptors (Werner et al., 2012). 123 % Upstream Agricultural Land Use 0 10 20 30 40 50 60 70 C H 4 C o n c e n t r a t i o n ( n M ) 0 1000 2000 3000 4000 5000 6000 Apr Aug Oct Figure 3.11 Plot of CH4 concentration vs. Upstream Agricultural Land Use (%) for each site. April, August, and October samples are plotted, with each month including samples of the Sep 2012–Feb 2013 and the Apr 2013–Apr 2014 sample sets. Stream sites in forested settings, which show significantly low dissolved CH4 relative to wetland and anthropogenic sites, are typically well buffered by the forest from excessive sediment and nutrient inputs which contribute to methanogenesis at anthropogenic sites. Also, forested sites are characterized by primarily rocky substrate in and around the stream bed which are not conducive to organic matter deposition or methanogenesis (Jones and Mulholland, 1998a). Further, most stream sites in forested settings are fast flowing, with some turbulence due to rocky substrates. Methane at these sites is therefore removed at greater rates by evasion to the atmosphere (Anthony et al., 2012; Wallin et al., 2014), contributing to low concentrations. 124 Although CH4 concentrations vary spatially according to stream setting, some variability is observed among sites of the same setting (Fig. 3.4). This may be in part caused by groundwater contributions of CH4 to streams, as these contributions vary on both large and small scales in stream networks (Jones and Mulholland, 1998b; Werner et al., 2012). Also, the methanotrophic process of CH4 oxidation, a major CH4 sink in freshwater systems, affects CH4 concentrations (De Angelis and Scranton, 1993). Microbial oxidation is found to vary with several water chemistry parameters, including nutrient concentrations and fine particle size (De Angelis and Scranton, 1993), and this may allow greater variability in CH4 concentrations even among sites of the same setting type in western MD. Another source of concentration variability for sites within the same setting type is coarseness inherent in designating site settings. Some characteristics of wetland sites, such as relatively slow flow, for example, are found in certain forested and anthropogenically influenced sites, causing higher variability across sites in these settings. Similarly, some forest cover may be found nearby wetland and anthropogenically influenced sites, affecting the extent to which these sites are buffered from adjacent environments. Taking the average concentration for each stream setting also introduces high variability since the average combines all CH4 measurements, incorporating temporal shifts in concentrations. Yet another source of variability across sites is the small-scale variability within a longitudinal stream section which was shown in this study. The longitudinal measurements at five points surrounding site CASS101D shows a range of 150 nM within a 50-m stream reach, with four points between 120–180 nM CH4 and one point 125 (–20 m in the stream reach) of elevated CH4 concentration, 274 nM. This along-stream variability in concentration with localized points of particularly high CH4 is in agreement with “patchiness” in stream CH4 which has been observed in some streams (Jones et al., 1995; Wallin et al., 2014). In situ microbial CH4 production and microbial production in riparian zones are irregular along a stream reach, producing this small-scale concentration variability (Jones and Mulholland, 1998b). Groundwater inputs of CH4 are also a factor as these vary on the stream reach scale (Jones and Mulholland, 1998b). Variability inherent in sampling and GC analysis may affect some variability observed in the stream reach, as the relative precision of the sampling and analysis is up to 8% for stream samples based on analyses of sample duplicates. While small-scale variability on the stream reach scale is thought to contribute to spatial variability observed in western MD streams, larger- scale factors such as the stream morphology and surrounding organic matter deposition generate regional spatial variability which is up to two orders of magnitude greater than small-scale variability. Considering large-scale spatial variability across physiographic provinces of MD, CH4 concentrations show extreme variability. Stream CH4 is highly elevated in coastal plains of Calvert County relative to western MD, with Calvert County waters ranging between 2,100–33,000 nM (Table B3). Stream concentrations of each of the three environmental settings sampled in Calvert County—forested, wetland, and anthropogenically influenced—were found to be significantly greater in concentration than those of western MD (Fig. 3.4). All streams in Calvert County are oversaturated 126 with CH4 with respect to the atmosphere, with S ranging from 940–13,400, so each provides a CH4 flux to the atmosphere. Several factors may contribute to differences in CH4 concentrations between western MD and Calvert County. The first is that there are substantial differences in human development among the two regions, with greater population density and developed land in Calvert County relative to the western MD study area (U.S. Census Bureau, 2010). With greater urban and suburban development, higher nutrient loads in runoff allow greater microbial activity and depletion of electron acceptors in drainage areas (Jones et al., 1995). This eventually favors methanogenesis in and along streams (Jones et al., 1995) and may be a factor across stream settings of the region. Another reason for observed elevated CH4 concentrations in Calvert County relative to western MD may be the lower stream gradients in the Calvert County coastal plain. This allows slower hydrologic exchange with the riparian zone and subsurface and allows greater deposition of fine particles (Battin et al., 2008). These geophysical features may facilitate anoxic conditions within and around most streams in this region, contributing to high CH4 concentrations observed in Calvert County (Jones and Mulholland, 1998a). Most western MD streams, in contrast, are within a high stream gradient environment (Eshleman, 2013), where more rapid exchange with subsurface flow can occur, minimizing opportunities for anoxia (Jones and Mulholland, 1998a) and yielding comparatively low CH4 concentrations. Particles in the headwater streams are typically coarser and more permeable, which enhances exchange with subsurface flow and reduces the extent of anoxia and methanogenesis 127 within the streambed (Battin et al., 2008; Jones and Mulholland, 1998a). In a northeastern U.S. stream system, Bresney et al (2015) found lower-gradient sites had significantly higher concentrations of CH4 than higher-gradient sites. Finally, higher stream CH4 concentrations in Calvert County may be related to the warmer climate of the coastal plain region compared to the Appalachian region in western MD. The longer growing season and shorter dormant season of this climate likely fosters greater CH4 productivity and higher concentrations in streams. For hemiboreal streams studied across Sweden, Wallin et al. (2014) reported a positive relation between mean ambient air temperature and stream CH4 concentration, suggesting that stream CH4 concentrations are related to regional climate. Within the Calvert County region, average stream concentrations of environmental settings are not significantly different from each other, yet high variability in concentrations is observed across the region as well as among sites of the same setting. As in western MD, the concentration variability among sites may be associated in part with microbial oxidation of CH4, which may vary spatially based on various stream water characteristics including temperature and nutrient levels (De Angelis and Scranton, 1993). Also, as with western MD study streams, designation of stream setting is somewhat coarse, with characteristics of setting types overlapping in some cases. For example, the site PARK02 is designated a forested site as it is surrounding by significant tree cover, although it has certain characteristics of wetland sites such as inundated soils surrounding the stream and relatively slow flow. Although differences were not significant across settings, sites in the wetland settings (GRAY01 and THOM01) show a high average concentration relative to other 128 Calvert County settings, attributable to slow flow and substantial aquatic plant coverage at these sites. High rates of organic matter deposition and an abundance of inundated anoxic sediments likely facilitate high rates of CH4 production in these systems (Hornibrook et al., 2000b; Ortiz-Llorente and Alvarez-Cobelas, 2012). Methane concentrations at these sites, in the range of 11,820−32,750 nM (Table B3), are consistent with concentrations measured in North American boreal wetlands, where CH4 concentrations ranged between <1 µM and 23 µM, depending on season (Hamilton et al., 1994). At the wetland sites in both June and July, ebullition was observed in the surface waters, which was assumed to contain CH4. Ebullition of CH4 is observed in many wetland systems. It occurs when sediment porewater is oversaturated in CH4, with porewater concentrations >1.75 mM (Wilson, 2014). In systems where ebullition occurs, it constitutes a major CH4 emission source to the atmosphere, particularly during warm seasons (Hornibrook et al., 2000b). 3.4.3 Carbon Isotope Measurements Values of δ13C-CH4 in western MD, between –60‰ and –44.3‰, are consistent with a biogenic CH4 source, as mentioned, and are suggestive of the acetate fermentation pathway characterized by the δ13C-CH4 range of –65‰ to –50‰ (Hornibrook et al., 2000a). A biogenic CH4 source to streams is compatible with observations of high CH4 concentrations in summer and in areas of abundant organic matter. Further, western MD δ13C-CH4 stream values were compared with δ 13C-CH4 values of the Marcellus Shale and overlying formations (−43‰ to −27‰) (Hakala, 2014), and western MD values were found to be significantly lower than this range 129 representing deep formations (p<0.05). This indicates that such thermogenic reservoirs are not the primary CH4 source in western MD streams. Although in western MD δ13C-CH4 values and concentration patterns suggest microbial methanogenesis is the primary stream CH4 source, there is variability of δ13C-CH4 values with some values above the range expected for biogenic CH4 (–110‰ to –50‰). Higher values, though, are not associated with a specific season or site environment (p>0.05) (Figs. 3.6–3.7). This δ13C-CH4 range with no clear relation to specific environments or seasons may be associated with multiple factors influencing δ13C-CH4 of biogenic CH4 in the freshwater environment, as discussed below. One factor which likely affects δ13C-CH4 and may preclude clear spatial or temporal patterns is the contribution of another methanogenic pathway, CO2 reduction, in addition to acetate fermentation. It is expected that acetate fermentation is the primary methanogenic pathway at the study sites since this pathway is known to be predominant in freshwater; however, CO2 reduction is thought to also contribute to some extent in most settings (Hornibrook et al., 2000a). Methane produced through CO2 reduction is of lower δ 13C-CH4 values ( −110‰ to −60‰) than that produced by acetate fermentation (−65‰ to −50‰) (Whiticar et al., 1986), so areas with greater contributions from CO2 reduction pathways would likely have lower δ 13C-CH4 values in the biogenic CH4 pool. Although contributions from the acetate fermentation vs. CO2 reduction may vary based on the availability of labile organic matter, with acetate fermentation particularly dominant in areas with more organic matter (Hornibrook et al., 2000a), no spatial differences in δ13C-CH4 based on stream setting 130 were found (p>0.05). This may be related to alteration of the methanogenic isotopic signals by another factor, methanotrophy. Methanotrophic oxidation of CH4 is a factor that likely affects stream CH4 and causes positive shifts in δ13C-CH4 values. The microbial oxidation of CH4 fractionates δ13C-CH4 since microbes preferentially utilize 12CH4, leading to an increase in δ13C-CH4 (Whiticar, 1999). As a result, biogenic CH4—whether formed by CO2 reduction or acetate fermentation—shows more positive δ 13C-CH4 values which may be above expected ranges (Whiticar, 1999). Methanotrophy is known to be a significant sink of dissolved CH4 in aquatic systems (Hornibrook et al., 2000a). It occurs in both aerobic and anaerobic environments, although aerobic methanotrophy is more common in freshwater settings (Whiticar, 1999). In streams, methanotrophy is estimated to consume between 2% and 21% of CH4 inputs (Jones and Mulholland, 1998b). Laboratory studies indicate that, during aerobic methanotrophy, CH4 may be fractionated between 4–30‰ (Whiticar, 1999). This indicates that CH4 of a strongly biogenic isotopic signature, with δ 13C-CH4 ranging between −110‰ and −50‰, may be offset in δ13C-CH4 following methanotrophy, leading to values in the range observed here, −60‰ to −44.3‰. Rates of methanotrophy are known to vary with sediment and nutrient characteristics as well as temperature (De Angelis and Scranton, 1993). A clear methanotrophic effect on δ13C-CH4 values is not observable, though, in the current study when comparing values across site environments (Fig. 3.7). This is likely due to isotopic shifts from methanotrophy overlying variability of δ13C-CH4 associated with the two different methanogenic pathways. Additionally, neither the extent of methanogenic acetate 131 fermentation and CO2 reduction, nor the extent of methanotrophy was apparent from temporal trends in δ13C-CH4 (Fig. 3.6). The co-occurrence of these processes affecting dissolved CH4 may preclude the differentiation of the isotopic effects of the individual processes in a bimonthly time series. Previous measurements of C isotopic composition of freshwater CH4 in North America have shown similar ranges in δ13C-CH4 that were likewise attributed to concurrent methanogenic and methanotrophic processes. In measurements of δ13C-CH4 across a variety of freshwater rivers in the U.S., Sansone et al. (1999) showed a range of –58‰ to –36‰, similar to the range found in western MD and Calvert County streams. Sansone et al. argue that CH4 in the observed systems is biogenic and that it is affected by microbial oxidation. Across multiple temperate North American wetlands, porewater δ13C-CH4 values were reported to fall between –75.9‰ and –46.0‰, similar to the δ13C-CH4 range in western MD streams (Hornibrook et al., 2000a). Using isotopic characterization of porewater CH4 and CO2, Hornibrook et al. showed that the reported range of wetland δ 13C-CH4 values is associated with the combination of CO2 reduction and acetate fermentation pathways in biogenic CH4 production. Similarly, a combination of these pathways may contribute to CH4 pools in western MD, although with relatively low values (–75.9‰ to –61‰) occurring in wetlands, there may be greater influence of CO2 reduction in these environments than in western MD streams. In Calvert County streams, δ13C-CH4 values range from –56.7‰ to –43‰, in agreement with values of western MD streams and those previously reported for North American streams and wetlands (Hornibrook et al., 2000a; Sansone et al., 132 1999). Measurements of δ13C-CH4 in Calvert County may be most informative, since CH4 in Calvert County is known to be biogenic due to the lack of deep groundwater sources of thermogenic CH4. Calvert County July measurements of δ 13C-CH4 versus concentration show a positive relationship (Fig. 3.12, white points, R2 = 0.82), and this suggests that in this region methanotrophy, which increases δ13C-CH4 values, is a predominant factor affecting dissolved CH4. In a coastal plain river system in NY, De Angelis and Lilley (1993) show an inverse relationship between CH4 oxidation rate and CH4 concentration, with CH4 oxidation acting as a key control on riverine dissolved CH4 concentrations. Western MD measurements of δ 13C-CH4 vs. concentration in Fig. 3.12 (gray and black points) show no relationship, which may indicate that δ13C-CH4 values in this region are affected by greater variability of methanogenic pathways (acetate fermentation and CO2 reduction) compared to Calvert County waters, precluding an observable trend from methanotrophic fractionation. Fractionating processes, largely methanotrophy, are thus found to act upon biogenic CH4 in Calvert County, which may further support a biogenic CH4 source in western MD. Western MD stream CH4, which shows a δ 13C-CH4 range similar to that of Calvert County with some high values above the biogenic δ13C-CH4 range (Fig. 3.7), may also be predominantly biogenic, and stream values in the –50‰ to –40‰ range in western MD are likely attributable to fractionation from CH4 oxidation. 133  1 3 C - C H 4 CH 4 Concentration (nM) 0 5000 10000 15000 20000 25000 30000 35000 -60 -50 -40 -30 Western MD 2013–2014 Western MD 2012–2013 Calvert County 2014 Figure 3.12 Plot of δ13C-CH4 (‰) vs. CH4 concentration (nM) for each site’s highest measured CH4 concentration, apart from site THOM02, for which δ 13C-CH4 was measured in a sample with slightly lower CH4 concentration (July ’14). Still, while δ13C-CH4 values in the biogenically productive Calvert County region suggest that western MD CH4 is largely biogenic, the sources of stream CH4 are not confirmed by the isotopic analyses. It is possible that western MD δ13C-CH4 values are derived from physical mixing of biogenic and thermogenic CH4 associated with vertical or lateral migration of thermogenic gas. Values of –64‰ to –50‰ were found to represent mixing of biogenic and thermogenic CH4 (Schoell, 1980; Whiticar, 1999). With multiple processes likely affecting the C isotope composition of stream CH4, intermediate δ 13C-CH4 values found in the stream settings do not fully allow differentiation between biogenic and thermogenic CH4. Measurements of δ13C-CH4 across space and time from this study may instead be used as baseline δ13C-CH4 values in western MD, which could reveal future CH4 contamination with 134 deviation from the baseline values. To more clearly distinguish biogenic and thermogenic CH4 in streams, measurements of both δ 13C-CH4 and δ 2H-CH4 should be made. These two parameters together may better define particular types of microbial and thermogenic CH4 which are affected by unique fractionation paths, with visualization in a δ13C vs. δ2H plot (Whiticar, 1999). In addition, C isotopic measurements of C2H6 may be useful for comparison with stream δ 13C-CH4 values since C2H6 produced in the Marcellus Shale is characteristically less enriched in 13C relative to CH4, such that δ 13C-CH4 − δ 13C-C2H6 > 1 (Jackson et al., 2013a). Although C2H6 may not be detected in streams under baseline conditions, it could present as a result of fracking-related gas migration. With C2H6 detection in streams, C isotopic analysis of both CH4 and C2H6 could be carried out to differentiate Marcellus Shale gases, of the isotopic trend δ13C-CH4 − δ 13C-C2H6 > 1, from Upper Devonian gases which have C2H6 that is more enriched in 13C than CH4 (δ13C-CH4 − δ 13C-C2H6 < 1) (Jackson et al., 2013a). 3.4.4 Monitoring Implications and Conclusions Through the monitoring of stream CH4 concentrations, I found that dissolved CH4 concentrations vary significantly throughout space and time within and among two distinct MD regions, with concentrations between 1–5,100 nM in western MD and between 2,100–33,000 nM in Calvert County. Temporally, highest concentrations were found in summer and fall for all sites. Spatial variability showed overall higher concentrations in wetland and anthropogenically influenced sites and lower concentrations at forested sites, although variability was also observed among 135 sites of the same setting type. This may be in part attributable to longitudinal variability in CH4, as stream reaches may show a concentration range of as much as 150 nM. This variability is also likely attributable to the somewhat coarse site environment distinctions, with some sites having characteristics of multiple types of environments. In future studies, concentration variability across different stream environments may be better assessed with inclusion of more monitoring streams of wetland and anthropogenically influenced environments. Temporally, more frequent sampling may better resolve the high variability of stream CH4 concentrations. In this work, monthly monitoring showed higher resolution of concentration variability than bimonthly monitoring (Fig. 3.2, panels A and B vs. panels C and D), and biweekly monitoring may capture even better the temporal concentration variability. Still, sampling for CH4 analysis is an involved process requiring replicate sampling and field addition of preservative and headspace, so increasing the frequency of sampling across space and across time may be burdensome. A monthly sampling frequency with the stream sites assessed in this work may then be optimal for capturing variability in CH4 concentrations. To better understand sources of concentration variability, future studies of stream CH4 should incorporate measurements of stream nutrient levels and dissolved organic carbon (DOC) since these are likely linked to in situ biogenic methanogenesis, particularly in view of observed elevated CH4 concentrations at wetland and anthropogenically influenced sites. Although all western MD and Calvert County streams are saturated or oversaturated with CH4 with respect to the atmosphere, stream concentrations are 136 very low relative to groundwater concentration ranges (0–4.42 mM) and also low relative to harmful well water concentrations (>0.623 mM) (Vidic et al., 2013). Still, presence of CH4 above baseline stream concentrations in the future will be of concern as this may indicate contamination related to fracking. The variability observed in streams indicates that outlying CH4 concentration values which may occur will be site- and season-specific. In forested sites in winter and spring, CH4 concentrations may range from 1–250 nM. In summer, forested sites show concentrations of 5–400 nM. In fall, these sites may range from 5–1,200 nM. Greater ranges are found in anthropogenically influenced environments, with concentrations expected between 5–1,200 nM in winter and spring and concentrations ranging between 5–4,050 nM in summer and fall. At wetland sites, CH4 concentrations are found to range between 50–1,000 nM in winter and spring, and in summer and fall, wetland sites may range between 50–5,100 nM. A perturbation in stream CH4 related to future fracking activity may be identifiable from CH4 measurements above the ranges expected for specific sites and seasons. High CH4 concentrations indicating surface water contamination would also likely be indicative of CH4 contamination of groundwater from fracking wells and may indicate subsequent fracking fluid contamination by way of the same migration pathway (Heilweil et al., 2013; Jackson et al., 2013a). With understanding of baseline ranges, the CH4 concentration data set may be best used for identifying long term increases in stream CH4, which may occur with fracking-related CH4 contamination (Vengosh et al., 2014). A trend of increasing CH4 concentrations across all seasons, even if small, may be indicative of CH4 migration which could develop or intensify over well life times (Vengosh et al., 137 2014). Also, observation of concentration shifts in groups or clusters of sites may be indicative of CH4 migration and contamination in certain locations. Long term interannual CH4 monitoring across western MD region would allow observation of trends overlying natural variability of stream concentrations. Although CH4 concentrations vary considerably within and among the two MD regions, C isotope composition of CH4 is relatively invariable, ranging between –60‰ and –43‰. Measurements of carbon isotopic composition do not clearly distinguish a biogenic or thermogenic CH4 source, although they may be indicative of biogenic CH4 formed largely by acetate fermentation, particularly as biogenic CH4 is consistent with concentration increases in summer and in productive areas. Because multiple processes influence δ13C-CH4 in streams, δ 13C-CH4 values alone are not useful for confirming the CH4 sources. Measurements of δ 13C-CH4 which are constrained by corresponding δ2H-CH4 measurements may allow distinct CH4 types to be differentiated in streams (Whiticar, 1999). Knowledge of natural CH4 concentrations in streams, as characterized in this study, is important prior to fracking, particularly in view of the high variability of concentrations within and among stream catchments. Additionally, δ13C-CH4 measurements indicate values typical of western MD streams. These CH4 analyses also expand our overall understanding of CH4 biogeochemistry in headwater streams. Temporally and spatially resolved monitoring of CH4, as was carried out in this study, will allow easier recognition of potential changes in CH4 occurrence in streams and may help mitigate fracking impacts to these stream systems. 138 Chapter 4: Conclusions This study suggests that baseline monitoring of stream water is essential for a region slated for fracking. Background measurements of constituents associated with fracking contamination indicate the applicability of some of these constituents as tracers of contamination. Such baseline measurements also show the naturally occurring concentrations and ranges that are expected in streams not affected by fracking contamination. This allows identification of atypical concentrations which may represent stream contamination. 4.1 Use of Sr2+, Ba2+, and major ions as fracking fluid tracers Stream concentrations of Sr2+ and Ba2+ are variable across space and time in western MD and Calvert County. All concentrations are low compared to fracking fluids, which are enriched by factors of 25,000–50,000 relative to streams (Barbot et al., 2013; Warner et al., 2013a). Upon occurrence of fracking fluid contamination events, which should be initially detected through continuous monitoring of conductivity, stream measurements of Sr2+ and Ba2+ can be carried out, whereby concentrations exceeding normal variability may signal the presence of fracking fluids in stream waters. Values of log([Ba]/[Sr]) will also be useful as a tracer for fracking fluids. Most western MD sites show relatively constant log([Ba]/[Sr]) values across time, yet these are expected to shift in the presence of fracking fluids due to addition of Ba2+ and 24SO  with subsequent BaSO4 oversaturation. This will cause a significant increase in 139 Sr2+ relative to Ba2+ in cases of fracking fluid contamination, resulting in a substantial decrease of log([Ba]/[Sr]) from typical values at the affected site. Baseline measurements of Sr2+ and Ba2+, and log([Ba]/[Sr]) in a region slated for fracking are necessary in order to use these constituents as tracers so that normal spatial and temporal variability may be assessed. Monthly monitoring of Sr2+ and Ba2+ concentrations appears to be an advantageous frequency for resolving patterns in the metal concentrations and is logistically feasible. This monitoring frequency also allows capturing of the temporal variability of log([Ba]/[Sr]) in streams. These constituents should be monitored just downstream of well pads where streams are most susceptible to fracking fluid contamination and where signals of fluid contamination would be most readily detected. Regular monitoring of Sr2+ and Ba2+ will also be useful for observation of longer interannual trends in concentrations since there is potential with fracking for upward migration of shale fluids, and this phenomenon may occur over multi-year time scales (Vengosh et al., 2014). Further, sustained casing pressures may increase the potential for fluid leakage with well age (Vengosh et al., 2014). Monitoring of long-term trends in stream Sr2+ and Ba2+ across western MD allows surveillance of longer time scale fluid intrusion of streams which may be induced with fracking. The isotopic composition of Sr may be a valuable tracer for fracking fluids in streams. Ratios of 87Sr/86Sr in fracking fluids from the Marcellus Shale are distinct from ratios of groundwater and surface water in the region and are also distinct from ratios in overlying formation fluids (Chapman et al., 2012). Analysis of Sr isotopic 140 composition is more difficult and costly, however, making regular Sr isotopic measurements currently impractical for tracing fracking fluids. Baseline measurements of major ion concentrations, Na+, K+, Mg2+, Ca2+, Cl−, and 24SO  , were elevated in streams relative to Sr 2+ and Ba2+, as expected. Most major ion species show high variability, likely related to the variability of geologic and anthropogenic sources of these ions across the region. These findings confirm that stream sampling for major ion analyses is not useful for tracing fracking fluids. It was found through exploratory measurements in 2012 that the Br− ion may be optimal for indicating fracking fluid contamination of streams. In a subset of 2012 western MD samples, Br− was found to be below detection limit in stream waters. With enrichment of Br− in fracking fluids, a detection of this species in streams could thus signal the presence of fracking fluids. Continuous monitoring of Br− for at least 1 year, as was done for Sr2+, Ba2+, and major ions, should be carried out to fully characterize baseline stream levels of Br−, with monthly monitoring recommended to assess potential temporal variability. 4.2 Use of dissolved CH4 for tracing fracking contamination Stream dissolved CH4, another potential indicator of fracking contamination, was found to be significantly variable throughout space and time in both western and southern MD. Temporally, highest concentrations are observed in summer and fall at all sites. Spatially, wetland and anthropogenically influenced sites are on average elevated in concentration relative to forested sites. With this variability, atypical CH4 concentrations which may indicate contamination will vary based on site and season. 141 Although monitoring of stream CH4 in this study was mainly carried out on a bimonthly basis, with monthly monitoring occurring for a shorter period (Sep 2012–Feb 2013), the observed high temporal variability in concentrations suggests that monthly measurements of CH4 may better capture the characteristic baseline. Monthly concentration measurements more clearly show the temporal variability and ranges in CH4 (Fig. 3.2 panels A and B). Throughout the region prior to fracking, dissolved CH4 should be monitored in streams of different environmental settings, as was done in this study, since concentrations are highly variable between settings. As with Sr2+ and Ba2+ measurements, monitoring of dissolved CH4 should occur in close proximity to well pads. Upon fracking development, increased stream CH4 concentrations may indicate contamination of nearby groundwater reservoirs in addition to surface water contamination (Heilweil et al., 2013). The observed variability in dissolved CH4 suggests that monitoring its concentrations may be most useful over longer, perhaps interannual, time frames. Two important means of stimulated CH4 flow—upward migration through annuli and leakage through well casings—may increase in potential over the life time of the well (Brufatto et al., 2003; Jackson et al., 2013a), so monitoring of long-term trends (on the order of years to decades) throughout western MD may be a beneficial use of CH4 concentration data. It is particularly important to detect CH4 migration related to well casing leaks since these may be followed some time later by brine or fracking fluid migration through the same contaminant pathway (Jackson et al., 2013a). To determine whether stream CH4 is primarily biogenic or thermogenic, measurements of C isotope composition were made. The range of δ13C-CH4 values 142 (–60‰ to –43‰) across both western MD and Calvert County are not indicative of thermogenic CH4 but are predominantly in agreement with biogenic CH4 production by acetate fermentation (Whiticar, 1999). Biogenic production of CH4 is consistent with the concentration patterns observed, where the highest concentrations occur in summer and fall and in areas of high organic matter abundance. Nonetheless, δ13C-CH4 values in the range measured here do not rule out the presence of thermogenic CH4 in western MD streams. With multiple processes potentially affecting δ13C-CH4, measurements of δ 13C-CH4 alone are not able to differentiate biogenic and thermogenic CH4 in stream settings. Values of δ 13C-CH4 may be more informative if used in combination with measurements of δ2H-CH4. Together, these two measurements allow identification of specific types of biogenic as well as thermogenic CH4 in a 2-dimensional diagram of δ 13C-CH4 vs. δ 2H-CH4 (Whiticar, 1999). 4.3 Further implications for fracking regions For each of the tracer constituents of interest for stream water quality—Sr2+, Ba2+, Br−, and dissolved CH4—monitoring must be carried out before, during and after fracking activity such that potential changes in stream chemistry may be observed. This monitoring should thus be implemented as a Best Management Practice (BMP) for fracking in MD and other regions anticipating fracking development. Although not feasible within logistical constraints of the current study, monitoring of stream baselines should preferably be carried out for at least two years for all stream sites, as indicated in the Recommended Best Management Practices for 143 Marcellus Shale Gas Development in MD (Eshleman, 2013), such that interannual variability may be observed for stream constituents. It will be possible to continue baseline monitoring of streams in western MD as the statewide ban of fracking permitting was recently extended through October 2017 (Hicks, 2015). Once well pads have been constructed, it may be beneficial to add new monitoring sites upstream from well pads in order to assess unaffected stream reaches. Measurements from such sites may be compared to those of sites downstream of well pads, as has been recommended by MD Department of Natural Resources (MD DNR). This will help to control for natural shifts in Sr2+, Ba2+, and CH4 related to environmental factors other than fracking, which will be beneficial since natural variability in these constituents is considerable. Baseline measurements in Sr2+, Ba2+, CH4, and potentially Br − will be important reference data for comparison with measurements collected during and after fracking in the region, allowing recognition of contamination from fracking fluids or mobilized CH4. Recognition of such contamination is key to preserving stream quality. Recognition of fracking contamination may also indicate the effectiveness of other implemented BMPs, such as fluid containment practices or site setbacks, which are required distances separating well pads from vulnerable areas (Eshleman, 2013). Further, recognition of stream contamination informs fracking management in other regions of the U.S. and world and indicates the primary hazards to stream waters. Fracking will continue to spread in the U.S. and also globally, as only 11% of shale gas outside of North America has been recovered (Entrekin et al., 2011). Assessment 144 of stream impacts in MD through monitoring of stream water chemistry may thus influence the future development of fracking worldwide. 145 Appendices Appendix A Data tables of metal and major ion concentrations Table A1. Stream concentrations of Na+, K+, Mg2+, and Ca2+ (µM), Sr2+ and Ba2+ (nM), Cl−, 42SO  , 3NO , 3HCO , and DOC (µM), and pH. Values of 3HCO were approximated from charge balance, unless marked by * which indicates 3HCO measured by Gran Titration. Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH WTP Intake Feb '12 658 32.2 125 326 437 454 923 79.9 88.1 421 Mar '12 639 32.2 121 303 411 438 885 85.9 86.2 376 Apr '12 582 30.9 121 331 428 429 812 76.9 73.2 478 May '12 590 31.2 122 323 436 435 824 75.1 61.9 476 June '12 620 30.7 129 354 476 450 881 70.5 42.0 554 July '12 671 34.2 135 325 496 411 900 68.2 590 Aug '12 670 35.1 141 343 525 402 895 63.3 652 Sept '12 732 39.2 145 324 547 424 934 60.6 655 7.09 Oct '12 740 43.6 149 341 562 407 966 57.4 683 6.97 Nov '12 764 52.6 147 344 553 403 999 69.9 659 Dec '12 674 47.3 147 377 540 437 941 82.4 662 6.88 Jan '13 693 49.5 143 337 509 445 954 88.1 571 6.88 Feb '13 1180 40.0 139 351 511 519 1545 87.9 481 6.81 WTP Finished Feb '12 726 32.8 127 319 437 452 1071 88.3 87.8 317 Mar '12 918 32.1 126 322 441 449 1103 95.7 85.9 467 Apr '12 802 31.5 122 344 428 424 995 87.4 79.6 515 May '12 818 32.2 124 336 443 434 1010 84.5 60.4 531 June '12 819 29.4 126 360 472 459 1055 84.6 44.5 550 July '12 850 33.0 133 335 507 421 1112 79.3 550 Aug '12 902 33.9 140 359 541 419 1141 75.2 644 Sept '12 962 37.8 147 336 566 432 1189 72.0 634 7.02 Oct '12 1012 43.1 150 346 602 414 1269 69.1 639 6.84 Nov '12 998 48.9 149 355 576 379 1231 74.9 673 Dec '12 903 53.0 146 347 513 399 1135 96.1 614 6.81 Jan '13 912 49.6 144 338 501 415 1160 100 567 7.01 Feb ‘13 1322 42.2 136 334 488 492 1607 103 490 7.18 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH Reservoir Feb '12 673 54.0 111 190 346 386 776 94.1 125 239 Mar '12 690 41.0 106 185 349 374 789 93.6 120 225 Apr '12 679 39.2 110 189 364 369 776 91.3 90.9 267 May '12 591 36.8 109 190 378 370 681 87.6 51.5 318 June '12 603 38.7 114 204 412 323 680 83.7 34.4 396 July '12 641 37.8 123 219 479 325 732 79.4 472 Aug '12 714 37.7 133 235 500 335 836 72.7 506 Sept '12 809 43.4 143 253 550 384 945 65.8 569 7.53 Oct '12 830 50.5 147 264 570 449 980 61.7 599 7.33 Nov '12 880 64.3 143 260 542 423 1033 77.6 561 Dec '12 785 61.3 139 246 492 403 926 97.6 494 7.04 Jan '13 Feb '13 633 48.7 116 203 366 426 745 95.6 382 6.77 SRPS Feb '12 434 16.7 109 476 462 525 782 58.2 45.8 677 Mar '12 471 16.1 106 432 454 551 780 75.1 51.6 581 Apr '12 346 15.2 99.4 505 468 500 634 57.0 48.2 774 May '12 298 15.9 105 547 493 499 630 46.4 43.9 851 June '12 296 15.4 97.9 540 459 472 599 42.6 37.4 866 July '12 255 16.0 105 559 478 466 600 36.7 36.9 889 Aug '12 222 15.6 102 590 462 452 556 32.8 1001 Sept '12 241 16.7 116 620 456 450 620 32.2 1045 7.06 Oct '12 229 18.0 115 634 465 454 600 32.5 1081 6.92 Nov '12 299 17.0 125 646 573 537 709 43.6 1061 Dec '12 337 18.6 127 629 599 558 734 54.4 1025 6.67 Jan '13 515 17.1 153 638 658 676 1058 59.3 936 6.72 Feb ‘13 605 20.4 143 562 611 696 1115 69.7 782 6.65 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH Tributary T1 Feb '12 113 65.5 143 242 350 193 176 101 215 355 Mar '12 89.4 60.7 132 217 322 178 136 102 244 267 Apr '12 85.9 50.8 131 222 321 168 94.4 88.5 93.7 478 May '12 110 63.3 151 288 402 200 86.8 75.3 45.5 768 June '12 125 77.5 195 369 551 226 93.9 59.4 40.2 1078 July '12 187 129 254 503 706 332 161 37.0 1594 Aug '12 223 141 288 578 820 331 194 27.8 1848 Sept '12 185 147 271 509 724 418 138 17.3 1720 7 Oct '12 217 145 212 378 542 283 351 139 911 7.19 Nov '12 136 154 153 294 372 195 224 111 737 Dec '12 113 104 151 250 342 180 179 100 640 6.97 Jan '13 Feb '13 114 72.5 140 244 335 181 178 100 574 6.87 Tributary C6 Feb '12 839 40.1 126 193 369 968 96.6 114 241 Mar '12 314 33.8 103 154 285 381 96.4 119 168 Apr '12 293 29.9 99.4 149 288 362 89.0 67.6 211 May '12 317 31.1 94.2 155 314 348 85.7 49.0 278 June '12 538 40.3 138 217 430 708 82.0 62.0 354 July '12 651 68.4 178 295 609 858 67.2 673 Aug '12 759 72.9 197 329 681 1081 53.8 695 Sept '12 878 79.1 212 336 694 1299 49.4 654 Oct '12 649 84.5 202 331 656 910 77.6 734 7.1 Nov '12 305 58.0 133 210 384 373 109 457 Dec '12 246 42.4 118 177 329 288 99.3 392 7.26 Jan '13 Feb ‘13 603 40.1 121 196 367 738 97.1 345 6.65 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH CASS101D Apr '13 65.3 40.1 167 240 434 539 130 128 534 7.15 Jun '13 76.7 44.5 197 284 500 523 142 96.5 748 Aug '13 81.3 52.3 236 345 614 623 199 101 895 Oct '13 86.6 78.9 223 314 523 527 220 79.5 725* Dec '13 71.8 47.3 190 262 470 540 177 126 596 Jan '14 Apr '14 71.9 40.9 178 244 429 526 160 124 99.1 537 CASS103D Apr '13 140 86.6 114 216 333 644 170 119 479 7.18 Jun '13 135 62.0 124 228 358 502 147 87.4 580 Aug '13 143 88.2 121 239 372 484 163 79.9 627 Oct '13 141 82.3 148 244 389 491 200 44.0 634* Dec '13 135 87.2 124 242 359 565 167 97.7 592 Jan '14 Apr '14 143 60.7 110 218 324 592 168 119 166 434 CASS105D Apr '13 126 22.7 67.6 108 181 233 162 114 109 6.64 Jun '13 94.9 21.2 68.6 109 184 172 120 116 119 Aug '13 104 24.6 73.0 122 201 103 130 84.2 221 Oct '13 90.4 41.9 91.9 153 220 149 132 89.5 261* Dec '13 112 33.2 76.5 123 205 352 145 181 37.4 Jan '14 Apr '14 110 21.4 67.7 106 180 310 141 142 83.3 44.1 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH SAVA101D Apr '13 48.4 17.6 50.1 79.7 129 291 32.3 87.3 119 6.97 Jun '13 58.7 19.8 54.5 83.7 142 314 35.0 63.7 192 Aug '13 50.6 17.1 48.5 76.1 130 262 38.8 66.1 146 Oct '13 66.6 29.4 56.6 83.4 139 269 38.9 56.5 197* Dec '13 56.4 15.6 51.1 77.8 128 249 38.0 80.9 130 Jan '14 Apr '14 55.7 15.2 52.1 83.3 134 283 42.7 84.1 50.0 21.3 SAVA202D Apr '13 88.6 22.1 65.6 112 212 313 110 107 143 7.22 Jun '13 109 26.6 82.8 177 309 428 102 93.3 367 Aug '13 90.8 22.7 78.0 160 291 392 67.2 92.3 339 Oct '13 104 33.8 98.0 224 378 478 84.6 77.7 523* Dec '13 93.9 19.9 67.8 141 241 290 88.2 94.7 254 Jan '14 Apr '14 97.8 22.2 65.5 116 219 308 97.5 107 52.5 12.8 SAVA204S Apr '13 415 25.1 88.6 245 379 330 466 136 369 7.61 Jun '13 457 32.7 140 433 638 496 559 127 822 Aug '13 375 31.4 127 375 569 452 410 129 742 Oct '13 382 40.0 166 500 720 536 476 117 998* Dec '13 465 25.4 100 280 426 324 553 130 436 Jan '14 396 24.4 108 307 443 345 471 141 497 Apr '14 401 25.2 93.4 252 385 329 449 138 58.3 39.5 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH SAVA301D Apr '13 1034 26.2 95.9 192 329 333 1153 108 267 7.36 Jun '13 2241 38.0 164 361 609 499 2564 105 557 Aug '13 1593 43.3 166 346 587 406 1889 100 571 Oct '13 2097 76.0 255 495 874 606 2585 126 591* Dec '13 1549 30.4 125 252 439 361 1806 113 302 Jan '14 Apr '14 1370 28.1 109 209 378 334 1537 109 116 233 SAVA302D Apr '13 Jun '13 Aug '13 1863 39.5 207 416 751 490 2396 136 482 Oct '13 Dec '13 Jan '14 2235 31.6 194 390 639 513 2759 138 398 Apr '14 GEOR101D Apr '13 21.1 18.9 46.4 57.9 158 350 27.9 86.9 46.8 6.07 Jun '13 23.9 20.1 48.4 62.3 164 358 39.2 82.1 62.0 Aug '13 22.2 14.8 45.0 60.5 157 348 28.2 79.8 60.2 Oct '13 16.3 15.3 47.0 60.8 162 380 33.1 83.5 38.5* Dec '13 21.3 14.0 50.0 55.4 153 314 30.8 85.3 44.8 Jan '14 Apr '14 21.4 17.8 47.1 59.3 158 332 35.5 85.6 69.1 38.3 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH YOUG102D Apr '13 97.4 38.5 413 649 708 285 126 758 619 7.58 Jun '13 116 55.0 555 964 1087 404 126 1012 1058 Aug '13 97.7 54.6 602 997 1078 430 114 1174 888 Oct '13 108 75.0 815 1313 1380 525 146 1576 1035* Dec '13 112 40.4 317 549 597 216 159 531 664 Jan '14 Apr '14 79.8 34.9 323 531 579 234 96.3 549 64.9 620 YOUG103D Apr '13 79.7 33.4 306 528 626 270 103 536 609 7.60 Jun '13 102 47.2 411 776 1010 390 110 668 1078 Aug '13 85.4 42.5 451 796 955 400 104 832 854 Oct '13 99.6 62.8 597 1040 1260 476 128 1073 1078* Dec '13 91.1 34.8 250 465 553 214 128 402 623 Jan '14 Apr '14 68.2 28.3 253 456 536 232 82.2 410 69.1 604 YOUG104D Apr '13 69.2 23.0 193 737 859 326 58.3 269 1355 7.62 Jun '13 102 32.9 263 1137 1261 469 60.4 312 2252 Aug '13 86.1 34.0 257 1066 1241 474 65.7 302 2095 Oct '13 99.8 39.3 274 1233 1300 495 87.5 336 2334* Dec '13 76.5 22.8 177 701 790 277 94.9 230 1299 Jan '14 Apr '14 64.3 20.3 178 677 757 273 63.5 239 55.8 1247 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH YOUG105D Apr '13 17.3 11.8 48.6 94.7 162 215 18.5 150 -2.77 5.48 Jun '13 23.3 12.7 61.3 137 223 201 20.5 182 49.3 Aug '13 13.4 10.2 39.0 99.3 164 194 17.7 111 59.5 Oct '13 23.8 32.8 92.1 196 282 200 33.1 269 69.2* Dec '13 22.0 11.2 48.2 102 162 193 23.8 142 26.5 Jan '14 50.0 Apr '14 20.1 10.5 50.0 101 163 204 22.0 142 49.1 22.0 YOUG106D Apr '13 33.4 12.3 35.5 178 180 208 37.7 85.5 264 7.02 Jun '13 39.4 13.1 43.7 248 252 267 41.1 64.4 466 Aug '13 44.2 13.4 39.7 214 219 236 56.3 73.6 362 Oct '13 40.9 27.5 51.5 284 264 235 55.3 72.4 512* Dec '13 44.4 13.8 38.4 204 196 214 49.3 86.9 319 Jan '14 Apr '14 38.8 12.4 37.3 191 185 200 45.6 85.4 69.1 283 YOUG201D Apr '13 Jun '13 Aug '13 1699 32.6 148 348 532 442 2021 157 388 3 Oct '13 Dec '13 Jan '14 2579 32.9 163 383 604 514 3134 172 227 Apr '14 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH YOUG301D Apr '13 Jun '13 Aug '13 521 27.9 124 342 428 366 622 179 501 Oct '13 Dec '13 Jan '14 758 23.1 117 307 399 310 897 166 401 Apr '14 YOUG302D Apr '13 Jun '13 Aug '13 428 64.0 132 357 687 397 493 59.2 859 Oct '13 Dec '13 Jan '14 636 50.5 114 284 496 376 729 84.9 583 Apr '14 YOUG432S Apr '13 44.7 25.0 51.1 144 214 251 52.9 96.5 214 7.28 Jun '13 55.0 26.7 57.7 184 262 302 61.4 86.8 330 Aug '13 45.9 24.8 55.9 172 252 274 57.1 91.1 287 Oct '13 47.4 42.3 65.1 209 272 295 76.8 80.2 357* Dec '13 49.3 27.5 53.0 152 219 234 64.3 100 224 Jan '14 Apr '14 49.7 25.1 53.4 151 220 256 57.2 88.5 76.6 241 Table A1 (cont.) Sample Na+ K+ Mg2+ Ca2+ Sr2+ Ba2+ Cl− 24SO  3NO DOC 3HCO pH PRUN301D Apr '13 Jun '13 Aug '13 776 59.2 425 1203 3151 448 207 1545 794 Oct '13 Dec '13 Jan '14 405 33.1 237 596 1450 258 168 653 631 Apr '14 HELL01 June ‘14 1899 89.5 192 143 487 248 1821 134 569 July ‘14 1847 98.6 165 132 443 287 1698 113 616 GRAY01 June ‘14 364 21.8 105 87.7 302 222 417 52.9 248 July ‘14 514 20.4 105 87.5 292 211 498 36.3 350 GRAY02 June ‘14 524 32.9 106 85.6 275 287 643 73.7 150 July ‘14 608 34.2 114 91.0 296 304 732 69.6 182 PARK01 June ‘14 1384 37.0 147 539 1280 178 1721 76.6 918 July ‘14 1920 53.6 206 729 1749 202 2588 68.5 1119 PARK02 June ‘14 670 26.6 136 649 1516 142 765 63.6 1372 July ‘14 976 39.9 754 196 1814 198 1078 51.8 1735 THOM01 June ‘14 267 18.6 57.5 57.9 190 116 147 46.1 277 July ‘14 238 27.0 61.5 66.9 227 191 201 37.6 245 THOM02 June ‘14 195 30.6 49.9 22.1 190 109 163 63.4 79.2 July ‘14 214 29.3 47.7 20.4 110 149 187 58.4 75.4 156 Appendix B Data tables of CH4 concentrations and δ 13C-CH4 measurements. Table B1. Dissolved CH4 concentrations (nM) in western MD streams from the Sep 2012–Feb 2013 sample set and the Apr 2013–Apr 2014 sample set. The meaning of site names is described in Chapter 2. Concentration values are taken as the mean of sample duplicates, and relative errors, determined as half the difference between duplicate samples, is <8% for most values. The designated site environment is given in the column labeled ‘Env.’ with ‘A’ indicating anthropogenically influenced sites, ‘F’ indicating forested sites, and ‘W’ indicating wetland sites. Site Env. 2012 2013 2014 Sep Oct Nov Dec Jan Feb Apr Jun Aug Oct Dec Jan Apr WTP Intake 117 112 73 20 31 WTP Finished 128 89 76 50 22 31 Reservoir A 282 4050 164 137 405 SRPS 35 29 42 40 6 16 Tributary B7 W 66 87 142 125 76 Tributary G8 W 5100 1880 261 132 100 Tributary C6 W 904 401 245 132 Tributary T1 F 1130 254 225 242 CASS101D F 34 373 170 301 29 30 CASS103D A 127 453 242 381 94 85 CASS105D A 80 497 280 434 102 73 SAVA101D F 23 27 48 202 13 13 SAVA202D F 5.1 7.0 8.2 8.2 3 5.4 SAVA204S F 4.9 9 6.5 6.3 1.4 4.2 5.5 SAVA301D F 36 163 181 297 43 49 SAVA302D A 122 25 GEOR101D F 4 28 10 21 3.0 4.7 YOUG102D F 4.5 12 9.7 8.0 3.7 5.5 YOUG103D F 4.5 8.6 7.8 8.2 4.8 6.7 YOUG104D W 55 112 170 91 79 70 YOUG105D A 445 1490 1310 3820 757 483 YOUG106D W 727 2820 1030 497 1000 810 YOUG201D A 9.5 4.5 YOUG301D F 26 13 YOUG302D A 1560 1150 YOUG432S F 9 35 15 17 8.2 7.8 PRUN301D A 33 15 Table B2. Values of δ13C-CH4 (‰) measured in western MD streams from the Sep 2012–Feb 2013 sample set as well as the Apr 2013–Apr 2014 sample set. Values were determined from the mean δ13C-CH4 measurement of sample duplicates, and the relative error, determined as half the difference between duplicate samples, was <2% for most samples. Site 2012 2013 2014 Sep Oct Apr Jun Aug Oct Dec Jan Apr Reservoir −54.8 SRPS −57.4 Tributary B7 −50.6 Tributary G8 −46.7 Tributary C6 −53.1 Tributary T1 −56.1 CASS101D –54.8 −46.9 CASS103D –52.1 −51.4 CASS105D –57.0 −51.0 SAVA101D −60 SAVA301D –55.7 −50.2 SAVA302D −53 YOUG104D –55.2 −48.7 YOUG105D −54.5 −52.8 −55 −54.6 −56.1 −58.2 YOUG106D −54.3 −49.9 −48.8 −44.3 −58.4 −59.5 YOUG302D −54 −56.2 YOUG432S −56.5 159 Table B3. Dissolved CH4 concentrations (nM) and δ 13C-CH4 values (‰) in Calvert County streams in June and July of 2014. Site names were designated based on the stream in which sites were located, as defined in Chapter 2. Concentration values were determined from the mean of sample duplicates and showed relative error <8% for most measurements. δ13C-CH4 values were taken from the mean of sample duplicates and relative errors were <2% for most samples. Site [CH4] (nM) δ 13C-CH4 (‰) June ‘14 July ‘14 June ‘14 July ‘14 HELL01 2169 3793 −56.7 GRAY01 26130 32750 −45.7 GRAY02 2492 2576 −54.7 PARK01 6055 6219 −54.0 PARK02 2307 14140 −52.4 −52.1 THOM01 11820 24920 −43 THOM02 3363 2607 −53.2 Table B4 Longitudinal gradient of dissolved CH4 concentrations (nM) measured in Sep 2013. Duplicate samples were measured at the CASS101D site (0 m), as well as from locations downstream (−30 m and −20 m) and upstream (+10 m and +20 m). Errors were determined as half the difference between duplicate samples. distance from site (m) [CH4] (nM) –30 179 –20 274 ± 40 0 170 ± 1 +10 125 ± 2 +20 124 ± 1 160 Bibliography Adair, S.K.P., B.R.; Monast, J.; Vengosh, A.; Jackson, R.B., 2012. Considering Shale Gas Extraction in North Carolina: Lessons from Other States, Duke Environmental Law and Policy Forum. Ahearn, D.S., Sheibley, R.W., Dahlgren, R.A., Keller, K.E., 2004. Temporal dynamics of stream water chemistry in the last free-flowing river draining the western Sierra Nevada, California. J. Hydrol. 295, 47-63. Allen, D.T., Torres, V.M., Thomas, J., Sullivan, D.W., Harrison, M., Hendler, A., Herndon, S.C., Kolb, C.E., Fraser, M.P., Hill, A.D., Lamb, B.K., Miskimins, J., Sawyer, R.F., Seinfeld, J.H., 2013. Measurements of methane emissions at natural gas production sites in the United States. Proceedings of the National Academy of Sciences of the United States of America 110, 17768-17773. Alvarez, R.A., Pacala, S.W., Winebrake, J.J., Chameides, W.L., Hamburg, S.P., 2012. Greater focus needed on methane leakage from natural gas infrastructure. Proceedings of the National Academy of Sciences 109, 6435- 6440. Anning, D.W.F., M.E., 2014. Dissolved-solids sources, loads, yields, and concentrations in streams of the coterminous United States. U.S. Geological Survey, Reston, VA. Anthony, S.E., Prahl, F.G., Peterson, T.D., 2012. Methane dynamics in the Willamette River, Oregon. Limnol. Oceanogr. 57, 1517-1530. Appler, M., 2010. Garrett County Land Use/Land Cover, in: Planning, M.D.o. (Ed.), Baltimore, MD. Aravena, R., Wassenaar, L.I., 1993. Dissolved organic-carbon and methane in a regional confined aquifer, southern Ontario, Canada: Carbon-isotope evidence for associated subsurface sources. Appl. Geochem. 8, 483-493. Armor, J.N., 2014. Key questions, approaches, and challenges to energy today. Catalysis Today 236, 171-181. Arnaud, C.H., 2015. Figuring Out Fracking Wastewater. Chemical and Engineering News 93, 8-12. Barbot, E., Vidic, N.S., Gregory, K.B., Vidic, R.D., 2013. Spatial and Temporal Correlation of Water Quality Parameters of Produced Waters from Devonian- Age Shale following Hydraulic Fracturing. Environ. Sci. Technol. 47, 2562- 2569. 161 Bastviken, D., Cole, J.J., Pace, M.L., Van de Bogert, M.C., 2008. Fates of methane from different lake habitats: Connecting whole-lake budgets and CH4 emissions. Journal of Geophysical Research-Biogeosciences 113. Bastviken, D., Tranvik, L.J., Downing, J.A., Crill, P.M., Enrich-Prast, A., 2011. Freshwater Methane Emissions Offset the Continental Carbon Sink. Science 331, 50-50. Battin, T.J., Kaplan, L.A., Findlay, S., Hopkinson, C.S., Marti, E., Packman, A.I., Newbold, J.D., Sabater, F., 2008. Biophysical controls on organic carbon fluxes in fluvial networks. Nature Geoscience 1, 95-100. Benstead, J.P., Leigh, D.S., 2012. An expanded role for river networks. Nature Geoscience 5, 678-679. Blasing, T.J., 2014. Recent Greenhouse Gas Concentrations. U.S. Department of Energy, Carbon Dioxide Information Analysis Center. Brantley, S.L., Yoxtheimer, D., Arjmand, S., Grieve, P., Vidic, R., Pollak, J., Llewellyn, G.T., Abad, J., Simon, C., 2014. Water resource impacts during unconventional shale gas development: The Pennsylvania experience. Int. J. Coal Geol. 126, 140-156. Bresney, S.R., Moseman-Valtierra, S., Snyder, N.P., 2015. Observations of Greenhouse Gases and Nitrate Concentrations in a Maine River and Fringing Wetland. Northeast. Nat 22, 120-143. Brezinski, D.K., Conkwright, R.D., 2013. Geologic Map of Garrett, Allegany, and Western Washington Counties in Maryland. Maryland Department of Natural Resources. Brufatto, C., Cochran, J., Conn, L., Power, D., El-Zeghaty, S.Z.A.A., Fraboulet, B., Griffin, T., James, S., Munk, T., Justus, F., 2003. From mud to cement— building gas wells. Oilfield Review 15, 62-76. Burnham, A., Han, J., Clark, C.E., Wang, M., Dunn, J.B., Palou-Rivera, I., 2012. Life-Cycle Greenhouse Gas Emissions of Shale Gas, Natural Gas, Coal, and Petroleum. Environ. Sci. Technol. 46, 619-627. Canedo-Arguelles, M., Kefford, B.J., Piscart, C., Prat, N., Schafer, R.B., Schulz, C.J., 2013. Salinisation of rivers: An urgent ecological issue. Environmental Pollution 173, 157-167. Carnevale, G., Godfrey, S.J., Pietsch, T.W., 2011. Stargazer (Teleostei, Uranoscopidae) cranial remains from the Miocene Calvert Cliffs, Maryland, 162 U.S.A. (St. Marys Formation, Chesapeake Group). Journal of Vertebrate Paleontology 31, 1200-1209. Carpenter, N.E., 2014. Chemistry of sustainable energy. Taylor and Francis Group, Boca Raton, FL. Caulton, D.R., Shepson, P.B., Santoro, R.L., Sparks, J.P., Howarth, R.W., Ingraffea, A.R., Cambaliza, M.O.L., Sweeney, C., Karion, A., Davis, K.J., Stirm, B.H., Montzka, S.A., Miller, B.R., 2014. Toward a better understanding and quantification of methane emissions from shale gas development. Proceedings of the National Academy of Sciences of the United States of America 111, 6237-6242. Chapman, E.C., Capo, R.C., Stewart, B.W., Kirby, C.S., Hammack, R.W., Schroeder, K.T., Edenborn, H.M., 2012. Geochemical and Strontium Isotope Characterization of Produced Waters from Marcellus Shale Natural Gas Extraction. Environ. Sci. Technol. 46, 3545-3553. Clark, C.E., Horner, R.M., Harto, C.B., 2013. Life Cycle Water Consumption for Shale Gas and Conventional Natural Gas. Environ. Sci. Technol. 47, 11829- 11836. Considine, T., Watson, R., Considine, N., Martin, J., 2012. Environmental impacts during Marcellus shale gas drilling: Causes, impacts, and remedies. Report 2012-1 Shale Resources and Society Institute. Craig, H., 1957. Isotopic standards for carbon and oxygen and correction factors for mass-spectrometric analysis of carbon dioxide Geochim. Cosmochim. Acta 12, 133-149. Dale, A.T., Khanna, V., Vidic, R.D., Bilec, M.M., 2013. Process Based Life-Cycle Assessment of Natural Gas from the Marcellus Shale. Environ. Sci. Technol. 47, 5459-5466. Darrah, T.H., Vengosh, A., Jackson, R.B., Warner, N.R., Poreda, R.J., 2014. Noble gases identify the mechanisms of fugitive gas contamination in drinking-water wells overlying the Marcellus and Barnett Shales. Proceedings of the National Academy of Sciences. De Angelis, M.A., Lilley, M.D., 1987. Methane in surface waters of Oregon estuaries and rivers. Limnol. Oceanogr. 32, 716-722. De Angelis, M.A., Scranton, M.I., 1993. Fate of methane in the Hudson River and estuary. Glob. Biogeochem. Cycle 7, 509-523. 163 Downing, J.A., Cole, J.J., Duarte, C.M., Middelburg, J.J., Melack, J.M., Prairie, Y.T., Kortelainen, P., Striegl, R.G., McDowell, W.H., Tranvik, L.J., 2012. Global abundance and size distribution of streams and rivers. Inland Waters 2, 229-236. Dresel, P.E., Rose, A.W., 2010. Chemistry and origin of oil and gas well brines in western Pennsylvania. Open-File Report OFOG 10, 01.00. Dunfield, P., Knowles, R., Dumont, R., Moore, T.R., 1993. Methane production and consumption in temperate and sub-arctic peat soils - Response to temperature and pH Soil Biology & Biochemistry 25, 321-326. Elias, R.W., Hirao, Y., Patterson, C.C., 1982. The circumvention of the natural biopurification of calcium along nutrient pathways by atmospheric inputs of industrial lead. Geochim. Cosmochim. Acta 46, 2561-2580. Entrekin, S., Evans-White, M., Johnson, B., Hagenbuch, E., 2011. Rapid expansion of natural gas development poses a threat to surface waters. Frontiers in Ecology and the Environment 9, 503-511. Eshleman, K.N., Elmore, A., 2013. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland. University of Maryland Center for Environmental Science, Frostburg, MD. Etiope, G., Klusman, R.W., 2002. Geologic emissions of methane to the atmosphere. Chemosphere 49, 777-789. Fedorov, Y.A., Khoroshevskaya, V.O., Tambieva, N.S., 2003. Variations in methane concentrations in the water of the Don River and Taganrog Bay under the effect of natural factors. Water Resources 30, 81-85. Ferrar, K.J., Michanowicz, D.R., Christen, C.L., Mulcahy, N., Malone, S.L., Sharma, R.K., 2013. Assessment of Effluent Contaminants from Three Facilities Discharging Marcellus Shale Wastewater to Surface Waters in Pennsylvania. Environ. Sci. Technol. 47, 3472-3481. Fitzpatrick, M.L., Long, D.T., Pijanowski, B.C., 2007. Exploring the effects of urban and agricultural land use on surface water chemistry, across a regional watershed, using multivariate statistics. Appl. Geochem. 22, 1825-1840. Flewelling, S.A., Sharma, M., 2014. Constraints on Upward Migration of Hydraulic Fracturing Fluid and Brine. Ground Water 52, 9-19. Flewelling, S.A., Tymchak, M.P., Warpinski, N., 2013. Hydraulic fracture height limits and fault interactions in tight oil and gas formations. Geophysical Research Letters 40, 3602-3606. 164 Fontenot, B.E., Hunt, L.R., Hildenbrand, Z.L., Carlton, D.D., Oka, H., Walton, J.L., Hopkins, D., Osorio, A., Bjorndal, B., Hu, Q.H.H., Schug, K.A., 2013. An Evaluation of Water Quality in Private Drinking Water Wells Near Natural Gas Extraction Sites in the Barnett Shale Formation. Environ. Sci. Technol. 47, 10032-10040. Forman, R.T.T., Alexander, L.E., 1998. Roads and Their Major Ecological Effects. Annu. Rev. Ecol. Syst. 29, 207-C202. Fry, B., 2006. Stable Isotope Ecology. Springer, New York. Godwin, K.S., Hafner, S.D., Buff, M.F., 2003. Long-term trends in sodium and chloride in the Mohawk River, New York: the effect of fifty years of road-salt application. Environmental Pollution 124, 273-281. Grasby, S.E., Betcher, R.N., 2002. Regional hydrogeochemistry of the carbonate rock aquifer, southern Manitoba. Canadian Journal of Earth Sciences 39, 1053-1063. Gregory, K.B., Vidic, R.D., Dzombak, D.A., 2011. Water Management Challenges Associated with the Production of Shale Gas by Hydraulic Fracturing. Elements 7, 181-186. Gross, S.A., Avens, H.J., Banducci, A.M., Sahmel, J., Panko, J.M., Tvermoes, B.E., 2013. Analysis of BTEX groundwater concentrations from surface spills associated with hydraulic fracturing operations. Journal of the Air & Waste Management Association 63, 424-432. Hakala, J.A., 2014. Use of stable isotopes to identify sources of methane in Appalachian Basin shallow groundwaters: a review. Environmental Science- Processes & Impacts 16, 2080-2086. Haluszczak, L.O., Rose, A.W., Kump, L.R., 2013. Geochemical evaluation of flowback brine from Marcellus gas wells in Pennsylvania, USA. Appl. Geochem. 28, 55-61. Hamilton, J.D., Kelly, C.A., Rudd, J.W.M., Hesslein, R.H., Roulet, N.T., 1994. Flux to the atmosphere of CH4 and CO2 from wetland ponds on the Hudson Bay lowlands (HBLS). J. Geophys. Res.-Atmos. 99, 1495-1510. Hammer, R., VanBriesen, J., Levine, L., 2012. In fracking’s wake: new rules are needed to protect our health and environment from contaminated wastewater. Natural Resources Defense Council 11. 165 Harkness, J.S., Dwyer, G.S., Warner, N.R., Parker, K.M., Mitch, W.A., Vengosh, A., 2015. Iodide, Bromide, and Ammonium in Hydraulic Fracturing and Oil and Gas Wastewaters: Environmental Implications. Environ. Sci. Technol. Harrison, S.S., 1983. Evaluating system for groundwater contamination hazards due to gas-well drilling on the glaciated Appalachian Plateau. Ground Water 21, 689-700. Hayes, T., 2009. Sampling and Analysis of Water Streams Associated with the Development of Marcellus Shale Gas. Marcellus Shale Coalition. He, C., Li, M., Liu, W.S., Barbot, E., Vidic, R.D., 2014. Kinetics and Equilibrium of Barium and Strontium Sulfate Formation in Marcellus Shale Flowback Water. Journal of Environmental Engineering 140. Heilweil, V.M., Stolp, B.J., Kimball, B.A., Susong, D.D., Marston, T.M., Gardner, P.M., 2013. A Stream-Based Methane Monitoring Approach for Evaluating Groundwater Impacts Associated with Unconventional Gas Development. Ground Water 51, 511-524. Hicks, J., 2015. Md. fracking moratorium to become law without Hogan's signature, The Washington Post. Hitzman, M.W., 2012. Induced Seismicity Potential in Energy Technologies, in: Council, N.R. (Ed.). The National Academies Press, Washington, D.C. Hockstad, L.W., M., 2013. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2011, in: E.P.A., U.S. (Ed.), Washington, D.C. . Hoffmann, W., 2014. The Economic Competitiveness of Renewable Energy. Scrivener Publishing, Beverly, MA. Hogan, J.F., Blum, J.D., 2003. Tracing hydrologic flow paths in a small forested watershed using variations in (87)Sr(/8)6Sr, Ca / Sr , Ba / Sr and delta O-18. Water Resour. Res. 39. Hornibrook, E.R.C., Longstaffe, F.J., Fyfe, W.S., 2000a. Evolution of stable carbon isotope compositions for methane and carbon dioxide in freshwater wetlands and other anaerobic environments. Geochim. Cosmochim. Acta 64, 1013- 1027. Hornibrook, E.R.C., Longstaffe, F.J., Fyfe, W.S., 2000b. Factors influencing stable isotope ratios in CH4 and CO2 within subenvironments of freshwater wetlands: implications for delta-signatures of emissions (vol 36, pg 151, 2000). Isotopes in Environmental and Health Studies 36, U1-U1. 166 Jacinthe, P.A., Filippelli, G.M., Tedesco, L.P., Raftis, R., 2012. Carbon storage and greenhouse gases emission from a fluvial reservoir in an agricultural landscape. Catena 94, 53-63. Jackson, R.B., Vengosh, A., Darrah, T.H., Warner, N.R., Down, A., Poreda, R.J., Osborn, S.G., Zhao, K.G., Karr, J.D., 2013a. Increased stray gas abundance in a subset of drinking water wells near Marcellus shale gas extraction. Proceedings of the National Academy of Sciences of the United States of America 110, 11250-11255. Jackson, R.E., Gorody, A.W., Mayer, B., Roy, J.W., Ryan, M.C., Van Stempvoort, D.R., 2013b. Groundwater Protection and Unconventional Gas Extraction: The Critical Need for Field-Based Hydrogeological Research. Ground Water 51, 488-510. Johnson, C.E., Driscoll, C.T., Siccama, T.G., Likens, G.E., 2000. Element fluxes and landscape position in a northern hardwood forest watershed ecosystem. Ecosystems 3, 159-184. Jones, J.B., Holmes, R.M., Fisher, S.G., Grimm, N.B., Greene, D.M., 1995. Methanogenesis in Arizona, USA dryland streams. Biogeochemistry 31, 155- 173. Jones, J.B., Mulholland, P.J., 1998a. Influence of drainage basin topography and elevation on carbon dioxide and methane supersaturation of stream water. Biogeochemistry 40, 57-72. Jones, J.B., Mulholland, P.J., 1998b. Methane input and evasion in a hardwood forest stream: Effects of subsurface flow from shallow and deep pathways. Limnol. Oceanogr. 43, 1243-1250. Judson, S.K., M.E.; Leet, L.D., 1987. Physical Geology, 7th ed. Prentice-Hall, Inc., Englewood Cliffs, NJ. Kappel, W.M., Nystrom, E.A., 2012. Dissolved methane in New York groundwater, 1999–2011. US Geological Survey Open-File Report 1162. Kargbo, D.M., Wilhelm, R.G., Campbell, D.J., 2010. Natural Gas Plays in the Marcellus Shale: Challenges and Potential Opportunities. Environ. Sci. Technol. 44, 5679-5684. Kaushal, S.S., Groffman, P.M., Likens, G.E., Belt, K.T., Stack, W.P., Kelly, V.R., Band, L.E., Fisher, G.T., 2005. Increased salinization of fresh water in the northeastern United States. Proceedings of the National Academy of Sciences of the United States of America 102, 13517-13520. 167 Kell, S., 2009. Modern Shale Gas Development in the United States; A Primer, in: Energy, U.S.D.o. (Ed.). Ground Water Protection Council, Oklahoma City, OK. Keller, E.A., 1982. Environmental Geology, 3rd ed. Bell & Howell Company, Columbis, OH. Kelly, C.A., Chynoweth, D.P., 1981. The Contributions of Temperature and of the Input of Organic Matter in Controlling Rates of Sediment Methanogenesis. Limnol. Oceanogr. 26, 891-897. Key World Energy Statistics, 2014. Organisation for Economic Co-operation and Development, Paris, France. Retrieved from: http://www.iea.org/publications/freepublications/publication/keyworld2014.p df Kim, W.Y., 2013. Induced seismicity associated with fluid injection into a deep well in Youngstown, Ohio. Journal of Geophysical Research-Solid Earth 118, 3506-3518. Kolb, R.W., 2014. The Natural Gas Revolution. Pearson, Upper Saddle River, NJ. Kondash, A.J., Warner, N.R., Lahav, O., Vengosh, A., 2014. Radium and Barium Removal through Blending Hydraulic Fracturing Fluids with Acid Mine Drainage. Environ. Sci. Technol. 48, 1334-1342. Kravchenko, J., Darrah, T.H., Miller, R.K., Lyerly, H.K., Vengosh, A., 2014. A review of the health impacts of barium from natural and anthropogenic exposure. Environmental Geochemistry and Health 36, 797-814. Land, M., Ingri, J., Andersson, P.S., Ohlander, B., 2000. Ba/Sr, Ca/Sr and Sr-87/Sr- 86 ratios in soil water and groundwater: implications for relative contributions to stream water discharge. Appl. Geochem. 15, 311-325. Lautz, L.K., Hoke, G.D., Lu, Z.L., Siegel, D.I., Christian, K., Kessler, J.D., Teale, N.G., 2014. Using Discriminant Analysis to Determine Sources of Salinity in Shallow Groundwater Prior to Hydraulic Fracturing. Environ. Sci. Technol. 48, 9061-9069. Lester, Y., Ferrer, I., Thurman, E.M., Sitterley, K.A., Korak, J.A., Aiken, G., Linden, K.G., 2015. Characterization of hydraulic fracturing flowback water in Colorado: Implications for water treatment. Science of the Total Environment 512, 637-644. Letcher, T.M., 2014. Future Energy: Improved, Sustainable and Clean Options for Our Planet, 2nd ed. Elsevier, London. 168 Lewis, A., 2012. Wastewater Generation and Disposal from Natural Gas Wells in Pennsylvania, Nicholas School of the Environment. Duke University. Likens, G.E., Bormann, F.H., Johnson, N.M., Pierce, R.S., 1967. Calcium magnesium potassium and sodium budgets for a small forested ecosystem. Ecology 48, 772-&. Liu, Y.C., Whitman, W.B., 2008. Metabolic, phylogenetic, and ecological diversity of the methanogenic archaea, in: Wiegel, J., Maier, R.J., Adams, M.W.W. (Eds.), Incredible Anaerobes: From Physiology to Genomics to Fuels. Blackwell Publishing, Oxford, pp. 171-189. Llewellyn, G., 2014. Evidence and mechanisms for Appalachian Basin brine migration into shallow aquifers in NE Pennsylvania, USA. Hydrogeol J, 1-12. Long, D.T., Wilson, T.P., Takacs, M.J., Rezabek, D.H., 1988. Stable-isotope geochemistry of saline near-surface ground-water - East-central Michigan Basin. Geological Society of America Bulletin 100, 1568-1577. Magen, C., Lapham, L.L., Pohlman, J.W., Marshall, K., Bosman, S., Casso, M., Chanton, J.P., 2014. A simple headspace equilibration method for measuring dissolved methane. Limnology and Oceanography-Methods 12, 637-650. Mair, R., Bickle, M., Goodman, D., Koppelman, B., Roberts, J., Selley, R., Shipton, Z., Thomas, H., Walker, A., Woods, E., 2012. Shale gas extraction in the UK: a review of hydraulic fracturing. Malakoff, D., 2014. The gas surge. Science 344, 1464-1467. Marshak, S., 2005. Earth: Portrait of a Planet, 2 ed. W. W. Norton & Company. Martell, A.E., Smith, R.M., Motekaitis, R.J., 2004. NIST Critically Selected Stability Constants of Metal Complexes. Texas A&M University, NIST Standard Reference Database 46 Version 8.0. Maryland Dept of Planning, Land Use/Land Cover. 2010. Retrieved from: http://www.mdp.state.md.us/PDF/OurWork/LandUse/County/Garrett.pdf Mast, A., Turk, J.T., 1999. Environmental Characteristics and Water Quality of Hydrologic Benchmark Network Stations in the Eastern United States, 1963- 95. US Geological Survey. McKenzie, L.M., Witter, R.Z., Newman, L.S., Adgate, J.L., 2012. Human health risk assessment of air emissions from development of unconventional natural gas resources. Science of the Total Environment 424, 79-87. 169 Merritt, D.A., Hayes, J.M., Marias, D.J.D., 1995. Carbon isotopic analysis of atmospheric methane by isotope-ratio-monitoring gas-chromatography mass- spectrometry. J. Geophys. Res.-Atmos. 100, 1317-1326. Meyer, J.L., Strayer, D.L., Wallace, J.B., Eggert, S.L., Helfman, G.S., Leonard, N.E., 2007. The Contribution of Headwater Streams to Biodiversity in River Networks1. JAWRA Journal of the American Water Resources Association 43, 86-103. Miller, S.M., Wofsy, S.C., Michalak, A.M., Kort, E.A., Andrews, A.E., Biraud, S.C., Dlugokencky, E.J., Eluszkiewicz, J., Fischer, M.L., Janssens-Maenhout, G., Miller, B.R., Miller, J.B., Montzka, S.A., Nehrkorn, T., Sweeney, C., 2013. Anthropogenic emissions of methane in the United States. Proceedings of the National Academy of Sciences of the United States of America 110, 20018- 20022. Molofsky, L.J., Connor, J.A., Wylie, A.S., Wagner, T., Farhat, S.K., 2013. Evaluation of Methane Sources in Groundwater in Northeastern Pennsylvania. Ground Water 51, 333-349. Moniz, E.J.J., H.D.; Meggs, A.J.M, 2011. The Future of Natural Gas: An Interdisciplinary MIT Study. MIT Energy Initiative. Moore, M.E., Buckwalter, T.F., 1996. Ground-water resources data for Warren County, Pennsylvania. US Geological Survey; Branch of Information Services [distributor]. Negrel, P., Allegre, C.J., Dupre, B., Lewin, E., 1993. Erosion sources determined by inversion of major and trace-element ratios and strontium isotopic-ratios in river water - the Congo Basin case. Earth and Planetary Science Letters 120, 59-76. Nesbitt, H.W., Markovics, G., Price, R.C., 1980. Chemical processes affecting alkalis and alkaline-earths during continental weathering. Geochim. Cosmochim. Acta 44, 1659-1666. O’Malley, M., McDonough, J., 2011. Executive Order 01.01. 2011.11: The Marcellus Shale Safe Drilling Initiative. Oliver, B.G., Thurman, E.M., Malcolm, R.L., 1983. The contribution of humic substances to the acidity of colored natural waters. Geochim. Cosmochim. Acta 47, 2031-2035. Olmstead, S.M., Muehlenbachs, L.A., Shih, J.S., Chu, Z.Y., Krupnick, A.J., 2013. Shale gas development impacts on surface water quality in Pennsylvania. 170 Proceedings of the National Academy of Sciences of the United States of America 110, 4962-4967. Ortiz-Llorente, M.J., Alvarez-Cobelas, M., 2012. Comparison of biogenic methane emissions from unmanaged estuaries, lakes, oceans, rivers and wetlands. Atmospheric Environment 59, 328-337. Osborn, S.G., Vengosh, A., Warner, N.R., Jackson, R.B., 2011. Methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing. Proceedings of the National Academy of Sciences of the United States of America 108, 8172-8176. Parker, K.M., Zeng, T., Harkness, J., Vengosh, A., Mitch, W.A., 2014. Enhanced Formation of Disinfection Byproducts in Shale Gas Wastewater-Impacted Drinking Water Supplies. Environ. Sci. Technol. 48, 11161-11169. Patino, R., Dawson, D., VanLandeghem, M.M., 2014. Retrospective analysis of associations between water quality and toxic blooms of golden alga (Prymnesium parvum) in Texas reservoirs: Implications for understanding dispersal mechanisms and impacts of climate change. Harmful Algae 33, 1- 11. Paul, M.J., Stribling, J.B. Klauda, R.J., Kazyak, P.F., Southerland, M.T., Roth, N.E., 2002. A Physical Habitat Index for Freshwater Wadeable Streams in Maryland. Maryland Department of Natural Resources, Annapolis, MD. Peek, S., Clementz, M.T., 2012. Sr/Ca and Ba/Ca variations in environmental and biological sources: A survey of marine and terrestrial systems. Geochim. Cosmochim. Acta 95, 36-52. Perlman, H., 2014. Saline water use in the United States, U.S. Geological Survey. Petron, G., Frost, G., Miller, B.R., Hirsch, A.I., Montzka, S.A., Karion, A., Trainer, M., Sweeney, C., Andrews, A.E., Miller, L., Kofler, J., Bar-Ilan, A., Dlugokencky, E.J., Patrick, L., Moore, C.T., Ryerson, T.B., Siso, C., Kolodzey, W., Lang, P.M., Conway, T., Novelli, P., Masarie, K., Hall, B., Guenther, D., Kitzis, D., Miller, J., Welsh, D., Wolfe, D., Neff, W., Tans, P., 2012. Hydrocarbon emissions characterization in the Colorado Front Range: A pilot study. J. Geophys. Res.-Atmos. 117. Pond, G.J., 2012. Biodiversity loss in Appalachian headwater streams (Kentucky, USA): Plecoptera and Trichoptera communities. Hydrobiologia 679, 97-117. Prieto, M., FernandezGonzalez, A., Putnis, A., FernandezDiaz, L., 1997. Nucleation, growth, and zoning phenomena in crystallizing (Ba,Sr)CO3, Ba(SO4,CrO4), 171 (Ba,Sr)SO4, and (Cd,Ca)CO3 solid solutions from aqueous solutions. Geochim. Cosmochim. Acta 61, 3383-3397. Rahm, B.G., Bates, J.T., Bertoia, L.R., Galford, A.E., Yoxtheimer, D.A., Riha, S.J., 2013. Wastewater management and Marcellus Shale gas development: Trends, drivers, and planning implications. Journal of Environmental Management 120, 105-113. Rai, S., Iqbal, M.Z., 2015. Using fluorescein and bromide tracers to investigate the role of baseflow in a small suburban watershed in Iowa, USA. Hydrological Processes 29, 173-186. Revesz, K.M., Breen, K.J., Baldassare, A.J., Burruss, R.C., 2010. Carbon and hydrogen isotopic evidence for the origin of combustible gases in water- supply wells in north-central Pennsylvania. Appl. Geochem. 25, 1845-1859. Rhodes, C., 2014. Peak Oil is Not a Myth. Royal Society of Chemistry, Reading, UK. Roth, N.E., Southerland, M.T., Mercurio, G., Chaillou, J.C., Kazyak, P.F., Stranko, S.S., Prochaska, A.P., Heimbuch, D.G., Seibel, J.C., , 1999. State of the Streams: 1995-1997 Maryland Biological Stream Survey Results. Maryland Department of Natural Resources, Annapolis, MD. Rowe, D., Muehlenbachs, K., 1999. Isotopic fingerprints of shallow gases in the Western Canadian sedimentary basin: tools for remediation of leaking heavy oil wells. Organic Geochemistry 30, 861-871. Saba, T., Orzechowski, M., 2011. Lack of data to support a relationship between methane contamination of drinking water wells and hydraulic fracturing. Proceedings of the National Academy of Sciences of the United States of America 108, E663-E663. Sanders, I.A., Heppell, C.M., Cotton, J.A., Wharton, G., Hildrew, A.G., Flowers, E.J., Trimmer, M., 2007. Emission of methane from chalk streams has potential implications for agricultural practices. Freshw. Biol. 52, 1176-1186. Sansone, F.J., Holmes, M.E., Popp, B.N., 1999. Methane stable isotopic ratios and concentrations as indicators of methane dynamics in estuaries. Glob. Biogeochem. Cycle 13, 463-473. Saville, J., Kashiwagi, M.T., Becker, A.J., Graves, P.H., 2014. A Multi-Year Update (2011–2014) to Maryland Biological Stream Survey’s Sentinel Site Network. Schedl, A., McCabe, C., Montanez, I.P., Fullagar, P.D., Valley, J.W., 1992. Alleghenian regional diagenesis - A response to the migration of modified 172 metamorphic fluids derived from beneath the Blue Ridge-Piedmont Thrust Sheet. J. Geol. 100, 339-352. Schobert, H.H., 2013. Chemistry of Fossil Fuels and Biofuels. Cambridge University Press, New York. Schoell, M., 1980. The hydrogen and carbon isotopic composition of methane from natural gases of various origins Geochim. Cosmochim. Acta 44, 649-661. Schoell, M., 1988. Multiple origins of methane in the earth. Chem. Geol. 71, 1-10. Segers, R., 1998. Methane production and methane consumption: a review of processes underlying wetland methane fluxes. Biogeochemistry 41, 23-51. Shaffer, D.L., Arias Chavez, L.H., Ben-Sasson, M., Romero-Vargas Castrillón, S., Yip, N.Y., Elimelech, M., 2013. Desalination and Reuse of High-Salinity Shale Gas Produced Water: Drivers, Technologies, and Future Directions. Environ. Sci. Technol. 47, 9569-9583. Sharma, S., Mulder, M.L., Sack, A., Schroeder, K., Hammack, R., 2014. Isotope Approach to Assess Hydrologic Connections During Marcellus Shale Drilling. Ground Water 52, 424-433. Skalak, K.J., Engle, M.A., Rowan, E.L., Jolly, G.D., Conko, K.M., Benthem, A.J., Kraemer, T.F., 2014. Surface disposal of produced waters in western and southwestern Pennsylvania: Potential for accumulation of alkali-earth elements in sediments. Int. J. Coal Geol. 126, 162-170. Sloto, R.A., 2013. Baseline groundwater quality from 20 domestic wells in Sullivan County, Pennsylvania, 2012. US Department of the Interior, US Geological Survey. Soeder, D.J., 2012. Advances in Natural Gas Technology, in: Al-Megren, H. (Ed.), Shale Gas Development in the United States. InTech, pp. 3-28. Sovacool, B.K., 2014. Cornucopia or curse? Reviewing the costs and benefits of shale gas hydraulic fracturing (fracking). Renew. Sust. Energ. Rev. 37, 249- 264. Strahler, A.N., 1952. Hypsometric (Area-altitude) analysis of erosional topography. Geological Society of America Bulletin 63, 1117-&. Stranko, S., Bourquin, R., Zimmerman, J., Kashiwagi, M., McGinty, M., Klauda, R., 2013. Do Road Salts Cause Environmental Impacts? Maryland Department of Natural Resources. Retrieved from: http://www.dnr.state.md.us/streams/pdfs/RoadSalt2013.pdf 173 Suarez, A.A., 2012. The Expansion of Unconventional Production of Natural Gas (Tight Gas, Gas Shale and Coal Bed Methane), in: Al-Megren, H. (Ed.), Advances in Natural Gas Technology. InTech, pp. 123-146. Sugimura, Y., Suzuki, Y., 1988. A high-temperature catalytic-oxidation method for the determination of non-volatile dissolved organic carbon in seawater by direct injection of a liquid sample. Marine Chemistry 24, 105-131. Taylor, S.W., Lollar, B.S., Wassenaar, L.I., 2000. Bacteriogenic ethane in near- surface aquifers: Implications for leaking hydrocarbon well bores. Environ. Sci. Technol. 34, 4727-4732. Templeton, C.C., 1960. Solubility of Barium Sulfate in Sodium Chloride Solutions from 25° to 95° C. Journal of Chemical & Engineering Data 5, 514-516. Thompson, S.E., Basu, N.B., Lascurain, J., Aubeneau, A., Rao, P.S.C., 2011. Relative dominance of hydrologic versus biogeochemical factors on solute export across impact gradients. Water Resour. Res. 47. Timpano, A.J., Schoenholtz, S.H., Soucek, D.J., Zipper, C.E., 2015. Salinity as a Limiting Factor for Biological Condition in Mining-Influenced Central Appalachian Headwater Streams. JAWRA Journal of the American Water Resources Association 51, 240-250. Trexler, R., Solomon, C., Brislawn, C.J., Wright, J.R., Rosenberger, A., McClure, E.E., Grube, A.M., Peterson, M.P., Keddache, M., Mason, O.U., Hazen, T.C., Grant, C.J., Lamendella, R., 2014. Assessing impacts of unconventional natural gas extraction on microbial communities in headwater stream ecosystems in Northwestern Pennsylvania. Frontiers in Microbiology 5. Trimble, D.C., 2012. Unconventional Oil and Gas Development: Key Environmental and Public Health Requirements. Government Accountability Office, Washington, D.C. Turekian, K.K., Wedepohl, K.H., 1961. Distribution of the Elements in Some Major Units of the Earth's Crust. Geol Soc America Bull Geological Society of America Bulletin 72. U.S. Census Bureau, 2010, Maryland State Data Center. Retrieved from: http://census.maryland.gov/ van der Elst, N.J., Savage, H.M., Keranen, K.M., Abers, G.A., 2013. Enhanced Remote Earthquake Triggering at Fluid-Injection Sites in the Midwestern United States. Science 341, 164-167. 174 Van Stempvoort, D., Maathuis, H., Jaworski, E., Mayer, B., Rich, K., 2005. Oxidation of fugitive methane in ground water linked to bacterial sulfate reduction. Ground Water 43, 187-199. Vengosh, A., Jackson, R.B., Warner, N., Darrah, T.H., Kondash, A., 2014. A Critical Review of the Risks to Water Resources from Unconventional Shale Gas Development and Hydraulic Fracturing in the United States. Environ. Sci. Technol. 48, 8334-8348. Vidic, R.D., Brantley, S.L., Vandenbossche, J.M., Yoxtheimer, D., Abad, J.D., 2013. Impact of Shale Gas Development on Regional Water Quality. Science 340. Wallin, M.B., Lofgren, S., Erlandsson, M., Bishop, K., 2014. Representative regional sampling of carbon dioxide and methane concentrations in hemiboreal headwater streams reveal underestimates in less systematic approaches. Glob. Biogeochem. Cycle 28, 465-479. Warner, N.R., Christie, C.A., Jackson, R.B., Vengosh, A., 2013a. Impacts of Shale Gas Wastewater Disposal on Water Quality in Western Pennsylvania. Environ. Sci. Technol. 47, 11849-11857. Warner, N.R., Jackson, R.B., Darrah, T.H., Osborn, S.G., Down, A., Zhao, K.G., White, A., Vengosh, A., 2012. Geochemical evidence for possible natural migration of Marcellus Formation brine to shallow aquifers in Pennsylvania. Proceedings of the National Academy of Sciences of the United States of America 109, 11961-11966. Warner, N.R., Kresse, T.M., Hays, P.D., Down, A., Karr, J.D., Jackson, R.B., Vengosh, A., 2013b. Geochemical and isotopic variations in shallow groundwater in areas of the Fayetteville Shale development, north-central Arkansas. Appl. Geochem. 35, 207-220. Waschbusch, R.J., Selbig, W., Bannerman, R.T., 1993. Sources of phosphorus in stormwater and street dirt from two urban residential basins in Madison, Wisconsin, 1994-95, National Conference on Tools for Urban Water Resource Management and Protection proceedings, February 710, 2000, Chicago, IL. DIANE Publishing, p. 9. Watmough, S.A., 2014. Calcium, strontium and barium biogeochemistry in a forested catchment and insight into elemental discrimination. Biogeochemistry 118, 357-369. Waxman, H., Markey, E., DeGette, D., 2013. Chemicals Used in Hydraulic Fracturing. US House of Representatives Committee on Energy and Commerce Minority Staff. 175 Weaver, T.R., Frape, S.K., Cherry, J.A., 1995. Recent cross-formational fluid-flow and mixing in the shallow Michigan Basin. Geological Society of America Bulletin 107, 697-707. Werner, S.F., Browne, B.A., Driscoll, C.T., 2012. Three-dimensional spatial patterns of trace gas concentrations in baseflow-dominated agricultural streams: implications for surface-ground water interactions and biogeochemistry. Biogeochemistry 107, 319-338. Whiticar, M.J., 1996. Stable isotope geochemistry of coals, humic kerogens and related natural gases. Int. J. Coal Geol. 32, 191-215. Whiticar, M.J., 1999. Carbon and hydrogen isotope systematics of bacterial formation and oxidation of methane. Chem. Geol. 161, 291-314. Whiticar, M.J., Faber, E., 1986. Methane oxidation in sediment and water column environments - isotopic evidence. Organic Geochemistry 10, 759-768. 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. Wilson, B., 2014. Geologic and baseline groundwater evidence for naturally occurring, shallowly sourced, thermogenic gas in northeastern Pennsylvania. Aapg Bulletin 98, 373-394. Wilson, J.M., VanBriesen, J.M., 2012. Oil and gas produced water management and surface drinking water sources in Pennsylvania. Environmental Practice 14, 288-300. Wright, P.R.M., P.B.; Mueller, D.K.; Clark, M.L., 2012. Groundwater-Quality and Quality-Control Data for Two Monitoring Wells near Pavillion, Wyoming, April and May 2012, Reston, VA. Yamamoto, S., Alcauskas, J.B., Crozier, T.E., 1976. Solubility of methane and distilled water and seawater. Journal of Chemical and Engineering Data 21, 78-80. Zeikus, J.G., Winfrey, M.R., 1976. Temperature limitation of methanogenesis in aquatic sediments. Appl. Environ. Microbiol. 31, 99-107. Zhang, C.L., Grossman, E.L., Ammerman, J.W., 1998. Factors influencing methane distribution in Texas ground water. Ground Water 36, 58-66.