Distribution and origin of dissolved methane, ethane and propane in shallow groundwater of Lower Saxony, Germany

Applied Geochemistry 67 (2016) 118e132

Contents lists available at ScienceDirect

Applied Geochemistry journal homepage: www.elsevier.com/locate/apgeochem

Distribution and origin of dissolved methane, ethane and propane in shallow groundwater of Lower Saxony, Germany S. Schloemer a, *, J. Elbracht b, M. Blumenberg a, C.J. Illing a a b

Federal Institute for Geosciences and Natural Resources (BGR), 30655 Hannover, Germany State Authority of Mining, Energy & Geology (LBEG), 30655 Hannover, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 November 2015 Received in revised form 2 February 2016 Accepted 5 February 2016 Available online 11 February 2016

More than 90% of Germany's domestic natural gas production and reserves are located in Lower Saxony, North Germany. Recently, research has been intensified with respect to unconventional shale gas, revealing a large additional resource potential in northern Germany. However, many concerns arise within the general public and government/political institutions over potential groundwater contamination from additional gas wells through hydraulic fracturing operations. In order to determine the naturally occurring background methane concentrations, ~1000 groundwater wells, covering ~48 000 km2, have been sampled and subsequently analyzed for dissolved methane, ethane and propane and the isotopic composition of methane (d13C). Dissolved methane concentrations cover a range of ~7 orders of magnitude between the limit of quantification at ~20 nl/l and 60 ml/l. The majority of groundwater wells exhibit low concentrations (<1 ml/l), a small number of samples (65) reveal concentration in the range >10 ml/l. In 27% of all samples ethane and in 8% ethane and propane was detected. The median concentration of both components is generally very low (ethane 50 nl/l, propane 23 nl/l). Concentrations reveal a bimodal distribution of the dissolved gas, which might mirror a regional trend due to different hydrogeological settings. The isotopic composition of methane is normally distributed (mean ~
70‰ vs PDB), but shows a large variation between
110‰ and þ20‰. Samples with d13C values lower than
55‰ vs PDB (66%) are indicative for methanogenic biogenic processes. 5% of the samples are unusually enriched in 13C (25‰ vs PDB) and can best be explained by microbial methane oxidation. According to a standard diagnostic diagram based on methane d13C values and the ratio of methane over the sum over ethane plus propane (“Bernard”-diagram) less than 4% of the samples plot into the diagnostic field of typical thermogenic natural gases. However, in most cases only ethane has been detected and the remaining less than 15 samples demonstrate an uncommon ratio of ethane to propane compared to typical thermogenic hydrocarbons. These data do not suggest a migration of deeper sourced gases, but a thermogenic source cannot be excluded entirely for some samples. However, ethane and propane can also be generated by microbial processes and might therefore represent ubiquitous background groundwater abundances of these gases. Nevertheless, our data suggest that due to the exceedingly low concentration of ethane and propane, respective concentration changes might prove to be a more sensitive parameter than methane to detect possible migration of deeper sourced (thermally generated) hydrocarbons into a groundwater aquifer. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Dissolved hydrocarbon concentration Groundwater Baseline Lower Saxony (Germany) Carbon isotopes

1. Introduction More than 90% of Germany's domestic natural gas production and reserves are located in Lower Saxony, North Germany. Most of

* Corresponding author. E-mail address: [email protected] (S. Schloemer). http://dx.doi.org/10.1016/j.apgeochem.2016.02.005 0883-2927/© 2016 Elsevier Ltd. All rights reserved.

these gases are generated from Carboniferous source rocks, while the also prominent and still produced petroleum accumulations in Lower Saxony derive from Jurassic Posidonia or Cretaceous Wealden source rocks (Binot et al., 1993; Kockel et al., 1994). Recently, research has been intensified with respect to unconventional shale gas, revealing a large additional resource potential in northern Germany (Pierau et al., 2013). Simultaneously, concerns arose within the general public and government/political institutions

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

over potential environmental impacts from increased exploration and production activities. Many concerns have been raised regarding the potential for contamination of shallow groundwater, i.e. potable or otherwise useable water with a total dissolved solids (TDS) content of less than 1000 mg/L (Jackson et al., 2013a), associated with hydraulic fracturing operations. Potential risks for groundwater, surface water and/or soil by hydraulic fracturing and/or production of shale gas include contamination by fugitive natural gas (i.e. stray gas contamination), hydraulic fracturing fluids and/or formation (brines) and waste waters from the deep formations (Vengosh et al., 2014). However, groundwater gas contamination is not necessarily associated with shale gas production and increased hydraulic fracturing operations. Methane can be generated biologically under methanogenic aquifer conditions. Any hydrocarbon accumulation in the subsurface (oil/gas reservoirs, coal bed methane) might lose methane and higher homologues through natural and induced fractures (Myers, 2012; Nakata et al., 2012) or producing and abandoned wells (Kang et al., 2014; Vengosh et al., 2014). Likewise hydrocarbon gases leaking from underground gas storage sites
ve
sz et al., 2010) and pipeline leaks (Jackson et al., (Laier, 2012; Re 2014) have been observed. Numerous studies with respect to potential groundwater contamination by stray or fugitive gases related to shale gas production have been published in recent years. Some of them disclaim a relationship of observed dissolved hydrocarbon gases with hydraulic fracturing or drilling activities (Cheung et al., 2010; Davies, 2011; Li and Carlson, 2014; McMahon et al., 2015; Molofsky et al., 2011, 2013), whereas others provide objective evidence for shale gas production-induced stray gas contamination (Darrah et al., 2014; Dyck and Dunn, 1986; Jackson et al., 2011, 2013b;
ve
sz et al., 2010; Warner et al., 2012). Osborn et al., 2011a; Re Whether there is enough evidence that the intensive fracturing operations in the main shale gas plays in the US is affecting the groundwater quality is still under debate (e.g. Jackson et al., 2011; Osborn et al., 2011b; Schon, 2011; Saba and Orzechowski, 2011). Modeling and monitoring of vertical fracture heights induced by hydraulic fracturing shows that these do not exceed several hundred meters (Davies et al., 2012; Fisher and Warpinski, 2012; Flewelling et al., 2013), hence suggesting they will usually not extend to the typically shallow (<200 m) groundwater bodies (Vengosh et al., 2014) and the highest risk is gas release due to a loss of well integrity. Background measurements of hydrocarbon concentrations in groundwater wells are mandatory in many states of the US (Jackson et al., 2013a) and in Canada (e.g. baseline water well testing program for coal-bed methane development according to ERCB, 2006). Several national (e.g. UBA, 2014) and international recommendations (EU, 2014; IEA, 2012) call for intensive pre- and postproduction groundwater monitoring programs particularly mentioning dissolved hydrocarbon gases in groundwater. Additionally, the analysis of the isotopic composition (carbon and hydrogen) of methane is mandatory at high concentrations, e.g. >2 mg/l (~2.8 ml/l) according to the Colorado Oil and Gas Conservation Commission (COGCC orders 112e156 and 112e157), >5 mg/l (~7.1 ml/l) according to the Wyoming OGCC or in Pennsylvania above 7 mg/l (~9.9 ml/l) (PA Code, Oil and Gas Act Regulations Chapter 78) to pinpoint the source of the dissolved gas. So far only sparse data sets are available for German groundwater concerning the composition and isotopic signatures of dissolved hydrocarbons. Scherer (2000) published the methane concentration of 68 groundwater extraction wells in northern Germany ranging from 0.15 to 52 ml/l (0.1e37 mg/l). Melchers (2008) published data from the Cretaceous Münsterland basin (bordering southeast to Lower Saxony) where elevated methane

119

concentrations up to 67 ml/l have been observed. Intensive mining operations (coal, strontianite) induced mine subsidence thus potentially creating effective migration pathways for deeper sourced methane from underlying coal seams. Gruendger et al. (2015) published a small data set (10 wells) of deep aquifers close to an open pit coal mine in North Rhine-Westphalia. Dissolved methane concentrations in this data set range from 0.3 to 2.2 ml/l. The aim of this study is to gain a first comprehensive data set on present day background hydrocarbon concentrations in groundwater of Lower Saxony, Germany as well as the methane stable carbon isotope ratios. Due to the limited number of hydraulic fracturing operations (327 Frac stages in Lower Saxony, depth generally >1100 m; LBEG, 2014) results of this study represent a virtual “pre-fracking” baseline. For our study approximately 1000 groundwater wells covering ~48 000 km2 have been sampled and subsequently analyzed for dissolved methane (CH4), ethane (C2H6) and propane (C3H8). The isotopic composition of methane (d13C) as diagnostic tool to characterize the origin of methane and as sensitive parameter to any additional methane input to the groundwater, in particular at initially low CH4 concentrations, was also analyzed. Hence, special emphasis has been devoted to develop a sample preparation line for routine GC-irMS (Gas Chromatographic System coupled to isotope-ratio Mass Spectrometry) analysis at low dissolved gas concentrations (lower limit for d13C measurements of methane is ~0.7 ml/l), which is considerably lower than for other baseline studies e.g. > 0.5 ml/l (Darrah et al., 2014). 2. Regional geological and hydrogeological conditions The two main structural geological units of the area under investigation comprise the Lower Saxony Sedimentary Basin (LSB) in the north and the Mesozoic/Paleozoic mountainous region in the south (Weser, Osnabrück and Leine hills, Harz mountains). The most common rocks in the Harz Mountains in the southeast are Devonian to early Carboniferous argillaceous shales and greywackes, the Weser, Osnabrück and Leine highlands are dominated by Mesozoic rocks. In the northwestern part of the LSB the principal source rocks for natural gas are Upper Carboniferous coals, while the marine Jurassic Posidonia and Cretaceous Wealden shale represent the primary source rock for oil. The Posidonia Shale is present throughout the LSB while the Wealden Shale exists primarily in the western portion of the basin. In the northeastern part of the LSB Upper Permian (Zechstein) marine shales, Kupferschiefer and Stinkschiefer, are the primary sources of oil. Above Mesozoic rocks of the LSB thick Cenozoic sediments have been deposited. The early Paleogene sediments have been deposited under marine conditions that changed during time into a continental accumulation environment. These common unconsolidated Paleogene sediments can reach thicknesses of several hundred meters in synclines around the widespread salt structures of Lower Saxony. Quaternary glaciation cycles resulted in sedimentation of thick glacial sands and debris in large areas in the northern part of the LSB. In addition major areas are dominated by fluvial deposits and the northeastern coastal area by shore sediments. The groundwater bodies of Lower Saxony are composed of 11 major terranes (Fig. 1) which can be further divided into 81 subunits (Elbracht et al., 2010). Briefly, the southern/southeastern part of Lower Saxony is dominated by aquifers in fractured consolidated sediments and the majority of the potable groundwater is related to porous aquifers in unconsolidated sands and gravels of Quaternary age. The northwest of Lower Saxony, bordering the North Sea, is dominated by intertidal flats, near-shore estuarine flat und bog lands of the rivers Elbe and Weser. Marine saltwater intrusion is

120

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

Fig. 1. Locations of sampled groundwater wells and the major hydrogeological units of Lower Saxony. Black dots ¼ NLWKN data set (acidified), squares ¼ LBEG sample set (see Section 4).

common in near shore groundwater bodies which are mostly unconfined. The confined or semi-confined groundwater reservoirs in unconsolidated rocks of the marsh areas are enriched in dissolved organic carbon (DOC, Elbracht et al., 2010). The central part of Lower Saxony is dominated by two hydrological terranes, the Geest areas and the lowlands of the rivers Elbe, Weser and Ems with numerous low-lying bogs. The Geest areas comprise the moraine sandy soils of the North German Plain which are topographically slightly higher than the eastern near-shore estuarine flat. Two main groundwater bodies occur in unconsolidated high permeable meltwater deposits. Groundwater chemistry is variable in these hydrogeological units, and occasionally dissolved organic carbon is high (e.g. groundwater in Miocene lignitebearing sands). The broad valleys of the Pleistocene watercourses formed the lowlands of the main rivers Elbe, Weser and Ems. Lowland bogs and alluvial meadow soils occur frequently. Two main aquifers are present in quaternary sands and gravels, Pliocene or Miocene sands or deep glacial channel casts. The upper aquifer is usually unconfined. The hardness of groundwater is typically low, Feconcentration and DOC are usually high (Elbracht et al., 2010). The transition from unconsolidated to consolidated sediments depicts the border to the southern/southeastern hydrological terranes of the Northwest-German Highlands, the Lower Triassic Bunter Sandstones and basement, the Subhercynian and Thuringian depression. The heterogeneous geological setting of the

Northwest-German Highlands leads to a complex hydrogeological situation. Groundwater is located in varying depths with highly variable maximum flow rates. The major groundwater aquifers are located in Lower Cretaceous sandstones (Wealden Formation, Bentheim Sandstones), Upper Jurassic limestones, Upper Triassic sandstones and carstified limestones and the sandstones of the Bunter formation. In areas with Middle to Late Permian Zechstein sediments, elevated sulfate concentrations are a common observation accompanied with a high water hardness. Groundwater in sandstones and some Tertiary aquifers can be rich in iron with a low pH (Elbracht et al., 2010). 3. Geochemistry of hydrocarbons Natural gas is a combustible gas with the primary component methane and a variable amount of the higher hydrocarbons ethane, propane, butane and pentane. The exact composition primarily depends on the formation process. Natural gas in the shallow subsurface is usually formed by anaerobic microbial reactions (Whiticar, 1999; Whiticar et al., 1986), while deeper hydrocarbons are generated by thermocatalytic (thermogenic) cracking (Hunt, 1996; Tissot and Welte, 1984) of macromolecular organic matter. Rarely observed are hydrocarbon gases derived from deep abiogenic sources (e.g. through FischereTropsch-type reactions; see for overview Faber et al., 1999). Biogenic gas is created by methanogenic archaea under anoxic

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

conditions once the electron acceptors (free oxygen, nitrate, manganese, iron and sulfate) are consumed. Methanogenic archaea require an absence of oxygen, which is uncommon in upper parts of unconfined aquifers. Anaerobic conditions and methane production at low rates might occur in micro niches e.g. around organic matter (Darling and Gooddy, 2006; Murphy et al., 1992) or when the consumption of O2 in deeper parts of stratified groundwater exceeds the rate of its diffusivity and nitrate and sulfate are completely reduced. Methane is the ultimate product of microbial reactions by two different metabolic pathways, fermentation of acetate and reduction of CO2 (Whiticar et al., 1986). Both processes occur in marine and freshwater, but acetate formation is the major pathway in nearsurface freshwater environments while CO2 reduction is dominant in the sulfate-free zones of marine sediments (Whiticar et al., 1986). Methane from both microbial pathways can be distinguished isotopically. The d13C value of methane from CO2 reduction in marine sediments is more negative (
110‰ to
60‰) than methane from acetate fermentation (mostly in freshwater environments;
65‰ to
50‰; Whiticar et al., 1986). Dissolved ethane (C2) and propane (C3) have been commonly observed as traces in dissolved gases in oxic and anoxic seawater (Hinrichs et al., 2006; Schobert and Elstner, 1980; Taylor et al., 2000; Xie et al., 2013) and lake sediments (Oremland, 1981; Oremland et al., 1988) but the order of magnitude is 3e4 lower than corresponding methane concentrations. Hence, with a gas dryness ratio C1/(C2 þ C3) (Bernard et al., 1977; Faber and Stahl, 1984; Schoell, 1983) or the simple methane-to-ethane ratio (Kang et al., 2014; Molofsky et al., 2013; Taylor et al., 2000) higher than 100 the gas under consideration is commonly interpreted as gas of biogenic origin. Methane formed by microbial processes is virtually ubiquitous in groundwater (Baldassare et al., 2014; Carroll, 2015; Clark and Fritz, 1997; Coleman et al., 1988; Molofsky et al., 2013). The opposite process, microbial methane oxidation, is a major sink of methane in the geosphere and water columns (Whiticar, 1999; Whiticar and Faber, 1986). Methane oxidation (methanotrophy) can either be accomplished aerobically or anaerobically. Methanotrophy is known to shift the isotopic composition of methane (Alperin and Hoehler, 2010; Whiticar, 1999; Alperin et al., 1988). Under aerobic conditions and the final product is carbon dioxide. Anaerobic methane oxidation (AOM) is a common and well understood process in anoxic marine and freshwater systems. Methane is oxidized with sulfate, nitrate, nitrite and metals as terminal electron acceptor (e.g. Alperin and Hoehler, 2010; Van Stempvoort et al., 2005). Both aerobic and anaerobic methane oxidation are associated with kinetic isotope effects. However the degree of isotope fractionation between educt (methane) and product (carbon dioxide), which leads to an enrichment of 13C and Deuterium in the educt, is usually much lower than for the reverse process of methanogenesis (Alperin and Hoehler, 2010; Whiticar, 1999). Thermogenic gas is a result of thermocatalytic cracking of different bonds in organic matter (kerogen) with increasing temperature (i.e. burial depth). The resulting composition and isotope ratios of the generated hydrocarbons depend on the source rock type (Schoell, 1983; Whiticar, 1996). The different source rocks (lacustrine, marine, or terrestrial) as well as the maturity (i.e. degree of alteration) can be distinguished by combined compositional and compound-specific isotope measurements (“fingerprinting”, Tilley and Muehlenbachs, 2011; Whiticar, 1994). Oil associated gas is enriched in C2þ components (>5%) as compared to later generated dry gases of both marine and terrestrial source rocks (C2þ <5%, mostly lower than 1%, Schoell, 1983; Whiticar, 1996). Throughout the last 40 years by means of analysis of gas samples

121

from well-defined different petroleum systems (e.g. Faber and Stahl, 1984; Lorant et al., 1998; Stahl, 1979; Schoell, 1983; Whiticar, 1990), as well as experimental data (e.g. Chung et al., 1988; Clayton, 1991; Galimov, 1988; Rooney et al., 1995; Sackett, 1984) practical relationships between various easily derived parameters (chemical, isotopic composition) and the known maturity and origin (source rock-type) have been established. Typically methane carbon isotopes range from
50‰ to
25‰, ethane and propane range from
40‰ to
20‰ for thermogenic gases, even in low temperature settings (Rowe and Muehlenbachs, 1999). Gas accumulations in the North German Basin exhibit high molecular nitrogen (50%e90% N2) concentrations. These gas occurrences are often sourced from Carboniferous coals and dispersed organic matter. It is proposed that nitrogen is released from the terrestrial organic material, which contains significant amounts of fixed nitrogen, and probably also from thermal alteration of ammonium bearing clay minerals at high temperatures (Krooss et al., 1995; Littke et al., 1995). 4. Material and methods Different methods for methane monitoring have been proposed in the literature ranging from application of pressure-sealed in-situ samplers, pressure measurements in sealed wells (Roy and Ryan, 2010, 2013) or even a stream-based approach (Heilweil et al., 2013). This baseline survey was performed in combination with regular sampling of groundwater wells with respect to water quality monitoring by means of submersible pumps. All wells sampled are regular groundwater monitoring wells, no private landowner wells have been probed. This study includes two sample sets which have been processed differently. The larger sample set (n ¼ 879, Fig. 1 black dots) was sampled by the NLWKN, the Lower Saxony authority in charge for regular groundwater sampling and analysis. Sampling took place between August and December 2014. A second smaller sample set (n ¼ 164, Fig. 1 white squares) was taken by LBEG (Lower Saxony State Authority for Mining, Energy and Geology). Samples have been taken between summer 2014 and winter 2015. In both cases, according to sampling regulations, groundwater is pumped through small hoses until field parameters (electric conductivity, oxygen content, pH, temperature) are constant prior to sampling. Water samples for dissolved hydrocarbon analysis were collected in 122 ml glass bottles with Teflon coated butyl rubber seals and aluminum crimp caps. Small tubes were inserted to the base of the bottle and completely purging the volume with low flow rates to avoid disturbances. Due to logistical constraints (transport, storage and laboratory capacity) the sample set from NLWKN was immediately acidified to pH < 2 with hydrochloric acid (2 ml, 37% HCl) to stabilize the samples until further analysis whereas the second sample set was not acidified (immediately analyzed after arrival in the laboratory). Samples were stored at 4  C until further analysis. Due to the large number of samples and small volumes, the dissolved gas concentrations were determined applying a headspace equilibration method (Capasso and Inguaggiato, 1998; Kampbell and Vandegrift, 1998) and not with the more time consuming total vacuum degassing method (e.g. Schmitt et al., 1991). 25 ml of the water was replaced by laboratory grade He (5.9) and the samples equilibrated at a temperature of 20  C for at least 2 h on a laboratory shaker. Due to the abovementioned acidification of the larger sample set, the samples were equilibrated under different conditions. The non-acidified samples were kept at atmospheric pressure whereas the needle for the other set was withdrawn immediately. As the pH is below 4.3 the equilibrium of the dissolved carbonate species is completely shifted to the CO2, the

122

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

resulting pressure increase in the headspace thus being a function of the alkalinity. For these samples the total pressure in the headspace was measured prior the following analysis. Hydrocarbon concentration of methane, ethane, ethene, propane and propene were determined using an Agilent 6890 GC equipped with a manual split/splitless injector and flame ionisation detector. Hydrocarbons are separated on a 50 m Al2O3 column in combination with 4 m HP1 pre-column. Due to the high variability in the gas concentrations two different calibration curves with varying injection amounts have been applied. The quantification limit is ~0.2 ppm hydrocarbon concentration in the sample gas. The concentrations of the dissolved gases were calculated using the partial pressure of the gases in equilibrium with the water (derived from fractional concentration and total headspace pressure), temperature of the sample, volume of headspace gas (25 ml) and remaining water (97 ml) and the Henry's Law constants of the compounds. Electrical conductivity is very low and hence total salinity negligible small, thus no correction for salting-out effects (Setchenow constants) has been applied. The relative error of the GC analysis is around ±3%. Throughout the text we refer to volume concentrations of the dissolved gases (nl/l or ml/l) and refer to 293 K and 101325 Pa. Conversion factors for methane are 1 ml/l ¼ 0.71 mg/l and 1 ml/l ¼ 44.6 mmol/l, respectively. Isotopic composition of methane at higher concentrations (>200 ppm headspace concentration) and the presented d13C data of CO2 have been determined on a standard GC-irMS system, an Agilent 6890 with compound separation on a 30 m Porapak column coupled to a MAT253. Methane has been combusted to CO2 at 980  C. Low concentration samples have been measured applying a cryo-focusing with liquid nitrogen of methane on a 1 m 1/16 packed column. This system was coupled to a Delta Plus XL. Using a sample loop of 8 ml the lower quantification limit for the d13C value of methane is ~2 ppm headspace concentration, which

4e5000 ppm methane concentration in the headspace (Fig. 2). Identical samples have been measured in duplicate (n ¼ 2) and the difference of the determined d13C value is plotted against absolute methane concentration in the headspace of the sample vial. The overall average deviation on identical samples is ±0.3‰ with no significant difference between the cryo-focusing and standard method. d13C of methane in air is measured daily as performance test, the average standard deviation being ±0.7‰. The precision for low concentration data is more than sufficient for the questions being asked of the data. 5. Results and discussion 5.1. Uncertainty estimation on split sample analysis During the initial phase of the study 23 sites have been randomly chosen for split sampling with respect to absolute concentration measurements and 15 sites for paired determination of the isotopic composition. All samples have been treated identically with respect to sample storage, headspace introduction, headspace pressure measurement and GC/CF-irMS analysis. The calculated relative percent difference (RPD1) of dissolved methane concentration of two samples is ~ ±10% and the average difference of two complementary isotope measurements is around ±1.5‰ (Fig. 3). The factual error of dissolved methane concentration is slightly higher than expected for duplicate analyses (e.g. 6%, Gorody, 2012) but far less than any error of reported concentrations if different laboratories are involved (40% between minimum and maximum values, Gorody, 2012). Our results on duplicate measurements meet the requirements RPD <25% (US EPA, 2002) for all concentration ranges. The errors of both concentration and isotope ratio are randomly distributed to the positive and negative with no apparent

Fig. 2. Results of control measurements of the isotopic compositions of methane. Identical samples were measured in duplicate (n ¼ 2). Vertical axis depicts the absolute difference of the two determined d13C values. Ideally the difference would be zero where the horizontal axis is intersecting the y-axis. Daily control sample was air (1.7 ppm concentration).

relates to a dissolved methane concentration of 700 nl/l (0.5 mg/l). Isotope ratios are given in the standard ‰ notation vs. PDB (Coplen, 2011). Precision and repeatability of isotope measurements has been tested on some samples with a concentration range of

1 Absolute(C1
C2)/(C1 þ C2)*100 with C1 and C2 results of sample pairs (e.g. US EPA, 2002).

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

123

Fig. 3. Results of split sample analysis. Blue dots represent measurements of methane d13C isotopic composition (left/bottom axis), red dots represent results of the methane concentration (right/top axis). Concentration values are given at a log-scale leading to some compression of the relative differences in the plot (see text). Grey line represents the ideal 1:1 correlation. No outliers have been removed. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

correlation to the total concentration (Fig. 3). These nonsystematic variations in the resulting gas concentrations can only partly be explained by normal errors during sample preparation (e.g. varying headspace/water volume ratios) and analysis. This suggests that an additional nonsystematic error occurs during field sampling procedures, e.g. filling and capping of sample vials will affect the gas concentration. Variations in the isotopic composition are much higher than the typical analytical error (see above) and should not be affected by sampling and sample preparation in the laboratory. Hence, the observed differences might reflect small heterogeneities in the groundwater reservoir which are successively pumped to the surface. Storage of the acidified samples over very long times might have some influence on the results. Nine split samples (no headspace introduced) have been analyzed for concentration and d13C values after storage of 110 days in between (data not shown). The second samples showed generally lower concentrations, 0.4%e45% less. However, the sample with the highest difference (45%) had an initially low concentration in the range of 200 nl/l and might therefore represent an outlier. The average difference (disregarding the outlier) is 7% between the two measurements, hence within the error range of split samples described above. Corresponding d13C values are generally slightly enriched in 13C in the repetitive measurement by þ0.2‰ to þ0.5‰ with one exception being depleted in 13 C by 2.2‰. No correlation between isotope shift and concentration decrease was observed suggesting that this is a nonsystematic error/variation. 50% of the acidified stored samples have been processed within 42 days, 80% within 70 days after sampling.

5.2. Concentrations of dissolved hydrocarbons Concentrations of dissolved methane cover a range of ~7 orders of magnitude between the limit of quantification at ~20 nl/l and maximum values of ~60 ml/l. There is no apparent correlation with the depth of the aquifer (Fig. 4). Similar to other studies we observed an average maximum concentration of ~35 ml/l (log10 (nl/l) ¼ 7.55). This phenomenon is well documented in many other studies (e.g. Molofsky et al., 2013; Roy and Ryan, 2010), and this value corresponds to the solubility of methane at 1 bar partial pressure (10  C ~42 ml/l, 15  C ~38 ml/l, Duan et al., 1992; Wiesenburg and Guinasso, 1979). This effect can be attributed to partial degassing during the sampling process and therefore methane concentrations in samples exceeding ~35 ml/l are most likely underestimated. However, more than 95% of our samples are not affected by this phenomenon. The carbon isotope fractionation factor a between CH4(dissolved) and CH4(gas) is between 1.00033 and 1.0005 (Bacsik et al., 2002; Fuex, 1980; Harting et al., 1976; Jansco, 2002). This accounts for a solubility difference of 0.03e0.05% or a difference of 0.3‰e0.5‰ between d13CH4(dissolved) and d13CH4(gas). Therefore a partial degassing during sampling will only marginally affect the isotopic composition of methane as there is no significant isotope fractionation effect between dissolved methane in water and the gas phase. The U.S. Department of the Interior (Office of Surface Mining Reclamation and Enforcement; Eltschlager et al., 2001) recommends monitoring if water contains more than 10 mg/l (~14 ml/l) of methane and immediate action if concentrations rise above 28 mg/l

124

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

Fig. 4. Methane concentration of Lower Saxony groundwater wells plotted against well depth. Methane concentrations are given as log10-values of the base unit [nl/l]. Hence 3 ¼ 1 ml/l and 6 ¼ 1 ml/l.

(39 ml/l) although sometimes lower values for household utilization are recommended (e.g. 7 mg/l ¼ 10 ml/l; Vengosh et al., 2014). For our data set of Lower Saxony 6% (n ¼ 60) of the groundwater samples exceed 14 ml/l and 1.3% (n ¼ 14) exceed the threshold of 39 ml/l. The frequency plot of methane concentration clearly shows an uneven distribution (Fig. 5). A first analysis indicates a bimodal distribution of the dissolved gas which can be fitted by two lognormal distributions and some arbitrary noise. The first narrow population has a median value of ~350 nl/l (log10 ¼ 2.5), the second, much broader distribution, a value of ~4 ml/l (log10 ¼ 3.6), respectively. There is an additional large variation especially at high concentrations which cannot be adequately matched by a similar log-normal distribution. In 27% of all samples (n ¼ 268) ethane and in 8% ethane and propane (n ¼ 82) was detected. The median concentration of ethane is 50 nl/l and for propane 23 nl/l, respectively (Fig. 6). There is no apparent correlation between ethane and propane concentrations or between between methane and the C2þ compounds (Figs. 5 and 6). Unfortunately the concentrations of ethane and propane were too low for determination of their d13C values, which restricts a more detailed interpretation. 5.3. Methane isotopic composition The d13C values of dissolved methane in groundwater from Lower Saxony is normally distributed (mean ~
70‰ vs. PDB), but shows a large variation between
110‰ and þ20‰ (Fig. 7A). By far the majority of the samples is within the range of biogenic methanogenic gas (d13C typically 50‰ for freshwater environments, Whiticar et al., 1986). Natural thermogenic gas isotopic composition is usually in the range of
50‰ to
25‰ (Tissot and Welte, 1984; Hunt, 1996) and depends on source rock type and maturity. However, this range overlaps with the range of isotopic composition resulting

from methanotrophic biogenic processes depending on the isotope value of original methane pool. A more precise distinction for methane in this overlapping range (partly oxidized or untainted thermogenic gas) can be based on dD value of methane. According to empirical data (e.g. Schoell, 1980, 1983) dD of thermogenic gas is enriched in Deuterium (
170‰ to
130‰) and correlates linearly with d13C of methane. A bacterial methane oxidation would result in enrichment of both Deuterium and 13C. However, as the fractionation factors are different (deuterium > carbon) the relationship would deviate from the typical relationship of a thermogenic gas. Further, if the gas under consideration contains higher alkanes and their isotopic composition d13C values are known a natural gas plot (Chung et al., 1988) could be applied to identify partial methane oxidation. The rarely encountered heavily 13C enriched samples (10‰) are certainly a specific anomaly. The highest carbon isotope fractionation factors are observed for aerobic methane oxidation (up to 1.031; Chanton and Liptay, 2000). Most likely these samples are a result of extensive methane oxidation in closed confined environments with limited occurrence of the methane substrate, i.e. comparable to a Rayleigh fractionation process. Comparing the isotopic composition with the absolute methane concentration, two populations become apparent in our sample set (Fig. 7B). At high concentrations, above 10 ml/l (log10(nl/l) ¼ 7), the isotope values are consistently clustering in the range of typical biogenic origin (
60‰ to
80‰). Below a methane concentration of ~0.1 ml/l (log10(nl/l) ¼ 5) we observe a large scattering of d13C(CH4) (Fig. 7B). At high concentrations the basic process of methanogenesis determines obviously the isotopic composition. The scatter in d13C of samples showing low methane concentrations cannot be pinpointed to a single process. Since these samples are highly sensitive to any variation (e.g. mixing/diluting with freshwater, oxidation, microbial activity) the statistical spread cannot be fully explained based on the existing dataset.

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

125

Fig. 5. Frequency plot of methane concentrations in potable groundwater aquifers of Lower Saxony (n ¼ 1042). Colored lines (red and green) indicate fitted single distributions, solid black line the summation (see text for explanation). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 6. Frequency plots of ethane (red) and propane (blue) concentrations. 11 samples with ethane concentration above 750 n/l were omitted from the plot for better resolution. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

126

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

Fig. 7. Frequency distribution (n ¼ 693) of d13C isotopic composition of methane (A) and cross plot of methane d13C isotopic composition versus concentration (B).

5.4. Potential sources of dissolved hydrocarbons As mentioned above, the majority of the samples has been acidified to pH <2 with HCl without forgoing filtering (to avoid degassing). For these samples d13C measurement of coexisting carbon dioxide are excluded as the dissolution of particulate carbonate matter would tamper with the isotopic composition of the dissolved carbon dioxide. Hence, the d13C of carbon dioxide has only been determined for a small number of non-treated samples. The isotopic composition d13C(CO2) is slightly varying around
20‰. The prevailing pathway of microbial methane

formation (CO2 reduction vs. acetate fermentation) or effect of microbial methane oxidation can be estimated using the cross plot of d13C of CO2 versus CH4 after Whiticar (1999, Fig. 8). In this diagram we included published data of Melchers (2008) from the Cretaceous Basin of Münster bordering Lower Saxony in the southwest and data from Gruendger et al. (2015). The data indicate that very high methane concentrations are associated with CO2 reduction rather than near-surface freshwater acetate fermentation. CH4 and CO2 isotopic compositions of a subsets of samples clearly indicate secondary methane oxidation as most plausible explanation. However, one needs to consider that this empirical

Fig. 8. Cross plot of d13C isotopic composition of carbon dioxide and methane in a diagram modified after Whiticar (1999) classed by dissolved methane concentration. Squares: this study, Lower Saxony (n ¼ 80). Triangles: Cretaceous Basin of Münster (Melchers, 2008; n ¼ 21), Dots: Open pit lignin mine (Gruendger et al., 2015; n ¼ 10).

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

diagram was originally derived from pore water data of marine and lake sediments which resemble a closed system with limited amounts of organic matter, dissolved inorganic carbon and nutrients after deposition. Groundwater aquifers are rather an open system with varying recharge and flow rates. The most commonly used interpretative diagram to differentiate between microbial and thermogenic gas is based on the gas composition and d13C of methane (Bernard et al., 1977; Faber and Stahl, 1984). In this diagram (Fig. 9) the gas dryness, i.e. the ratio of methane to the sum of ethane and propane (C1/(C2 þ C3)), is plotted against the methane carbon isotope value. The two domains representing the 12C enriched, almost pure methane “biogenic gases” and the isotopically 12C depleted “typical thermogenic natural gases” with high abundance of C2þ components are based on empirical data. As mentioned above, in 268 samples ethane and in 82 of these also propane was quantifiable. The resulting data scatter largely in the Bernard-Diagram (Fig. 9). It has to be noted that due to the Yaxis logarithmic scale, the majority of studied samples (~74%), which lack ethane and/or propane and share generally high methane abundances with mostly very negative d13C(CH4) vales are not included in the Bernard-Diagram. The first and largest sample group among samples with ethane and/or propane (Group I) plots in the empirical region of biogenic gases. Samples with ethane only are scattering whereas samples containing both ethane and propane cluster at high dryness ratios with an average methane d13C value of
60‰ to
70‰. Samples with ethane fall into two additional groups, one in the region with methane d13C values of less than
60‰ and dryness ratios between 10 and 200 (Group II). The other one (group III) is characterized by isotopically 12C depleted methane (
20‰ to even positive values) with varying but relatively high dryness (>100). A certain number of these samples (~40) plot into the diagnostic field of typical thermogenic natural gases (Group IV) with the

127

majority of these having a dryness ratio >> 100 and a methane d13C value between
50‰ and
30‰. A general interpretation of these data in this diagram is complicated taking into account that our data set represents the whole variability of geological/hydrological conditions, potential variations in terms of hydrochemistry, degree of microbial degradation, and, last but not least, our low detection limit of ethane/ propane which makes the presentation of the data in the Bernard diagram possible in the first place. The occurrence of ethane in general does not follow a regional pattern. The majority of the samples containing both ethane and propane (Fig. 9) can be reasonably classified as biogenic gas due to the mostly negative d13C(CH4) values. The above mentioned samples of Group I with a narrow range of d13C(CH4), including traces of propane and high dryness ratio are, to a certain extent, concentrating in the western part of Lower Saxony where marsh and bog areas are common. Usually the samples exhibiting propane are positively correlated with very high absolute concentrations of dissolved methane, indicating that propane (as well as ethane) is a byproduct of methanogenesis which has been previously described as a common characteristic of sea and lake sediments (Hinrichs et al., 2006; Schobert and Elstner, 1980; Oremland, 1981; Oremland et al., 1988; Taylor et al., 2000; Xie et al., 2013). However, a wide-spread occurrence of biogenic ethane and/or propane in many different aquifers in the same region have not been described previously. The essentially propane-free samples (Group II and Group III) are non-specific in this diagram. We suppose that these samples are representing a pool of biogenic gas (biogenic background) exhibiting methane and ethane which has been altered by secondary processes. It is well known that shallow groundwater aquifers are a very dynamic system with high spatial and temporal variability due to changing recharge/discharge conditions, groundwater flow rates, local inhomogeneities of the aquifer, vertical layering (oxygenated

Fig. 9. Cross plot of gas dryness versus d13C isotopic composition of methane (after Bernard et al., 1977). Red dots: samples with ethane only, blue dots: with ethane and propane. 774 samples with no ethane/propane do not show in this plot. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

128

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

above reduced groundwater) to name a few examples. Hence microbially mediated processes (degradation of organic matter, CO2 reduction, methane oxidation and others) are highly variable as well resulting in complex isotopical and compositional patterns. The C2/C3 ratios of NW-German natural gases are typically in the range 2, but depend on source rock type and maturity. Less than fifteen samples out of ~40 samples from Group IV reveal quantifiable amounts of both ethane and propane which would be a prerequisite to classify them as thermogenic gas. However these samples do not show the typical ethane/propane ratio. The most remarkable sample of this group has the highest measured amount of ethane (290 ml/l), but the propane concentration is four orders of magnitude lower (25 nl/l). In all other samples of this group propane is completely missing although the analytical sensitivity for propane is better (more carbon atoms and much better peak shape/ signal noise ratio). These unusual compositional data leave two possible explanations. These gases can be of thermogenic origin. But, most of these sample gases reveal a very high dryness ratio (>1000) which would point to a gas with a high thermal maturity (vitrinite reflectance > 2.0, e.g. Berner and Faber, 1988). Although source rocks with high maturities do exist in Lower Saxony (high maturity Posidonia Shale at the southern margin, Carboniferous shales to the North) current depths are > 4 km, hence making migration to the shallow surface unlikely but not impossible. In addition, high maturity thermal gases should exhibit relatively high d13C values for methane (
30‰ to
10‰) which is not observed for the samples under consideration. Alternatively these gases derived from more shallow sources. Rowe and Muehlenbachs (1999) described a low-temperature generation of methane through butane in shallow shales (200e600 m). However, the isotopic composition of this methane is more depleted in 12C (<<
60‰) as compared to the samples from our study. If these gases are of non-biogenic origin, there must be an additional process active which removes (partly or completely) propane selectively from the dissolved gas to account for the atypical compositional ratio. Recent studies have identified strains of bacteria which can biodegrade short-chain hydrocarbons under aerobic conditions where the results indicate that propane and butane are oxidized more rapid (Kinnaman et al., 2007). Similarly sulfate reducing bacteria have been identified which selectively oxidize propane and butane over methane and ethane (e.g. Jaekel et al., 2013; Kniemeyer et al., 2007). Even in the absence of oxygen or other oxygen donors (nitrate or sulfate) anaerobic methanogenesis from long-chain alkanes is possible (Zengler et al., 1999). Whether the absence of higher homologues is a result of these potential secondary processes cannot be inferred from the available data. The process of exclusion leaves another possible explanation: The gases are not of thermogenic origin, but biogenic. As mentioned above dissolved ethane and propane have been described as traces in seawater (Hinrichs et al., 2006; Schobert and Elstner, 1980; Taylor et al., 2000; Xie et al., 2013) or lake sediments (Oremland, 1981; Oremland et al., 1988). Thus it is possible that our data represent microbially mediated background concentrations in the potable groundwater aquifers. The exact origin of these gases cannot be determined without further detailed investigations. Possible sources include microbial degradation of organic matter in interbedded clay rich aquitards or extraction from shallow Cretaceous mudrocks which underlie the main aquifers and/or in-situ formation from dissolved and bioavailable organic matter (humic and fulvic acids), directly in the aquifer. A combination of both scenarios mentioned above (mixture of biogenic and thermogenic gas) is also principally possible. However,

samples from Group II could only be explained if the “biogenic end member” would be strongly enriched in 12C (
100‰ or lower). 5.5. Comparison to other published data sets When comparing our results to other recently published studies, these eventually thermogenic gases in our data set are generally low concentrated with a median value of 60 nl(C2 þ C3)/l, only 10 samples exceeding 1 ml(C2 þ C3)/l (Fig. 6). In their study of Northeastern Pennsylvania Molofsky et al. (2013) concluded that dissolved methane in groundwater on a regional scale is not correlated with shale gas extraction. Although propane concentrations are not reported in this study, there are some remarkable similarities between the two sample sets in terms of methane and ethane concentrations when comparing the statistic parameters mode and median (Table 1). In our study the values for methane are ~400/2500 nl/l and in Susquehanna County ~260/900 nl/l, respectively. Similarly the ethane statistical parameters mode/median match quite closely (Table 1) suggesting that both sample sets describe a geological/microbiological background of ethane dissolved in groundwater. In our data set is one potential outlier with an ethane concentration of ~300 ml/l (Table 1), all other samples are < 20 ml/l. Support for this interpretation comes from the fact that this particular sample shows a propane concentration of barely quantifiable 25 nl/l, thus 4 orders of magnitude lower. In contrast, in a published data set with assumed stray gas from shale gas production the median/mode concentrations of dissolved hydrocarbons are by orders of magnitude higher (Darrah et al., 2014; Reese et al., 2014, Table 1) than results from our study. Above the Barnett Shale the minimum determined dissolved methane concentration is > 1 ml/l and >3 ml/l above the Marcellus shale. In both areas the median values are by orders of magnitude higher than in our study. Reported ethane concentration above the Barnett Shale are generally higher than 2000 nl/l (mode 450 ml/l) and 100 nl/l (mode 3.4 ml/l) overlying the Marcellus Shale (Darrah et al., 2014). In the compiled data set of Reese et al. (2014), ethane and propane are only listed as lower than the very high quantification limit. Moreover, the studies in Pennsylvania have routinely involved different laboratories, which is demonstrated by the high range of quantification limits (Table 1), thus making a simple comparison of data sets difficult. 5.6. Spatial distribution of dissolved methane The regional spatial variability is depicted in Fig. 10. An overall trend of increasing values from SE to NW is readily identifiable and can be confirmed by an analysis of variance when comparing different regions. Potential explanations for this general tendency are related to regional geological and/or hydrogeological characteristics of Lower Saxony. The northwestern part of Lower Saxony is an estuarine low-lying area with higher soil organic carbon and to a certain extent higher alkalinity and generally lower sulfate concentrations (corresponding maps can be accessed through the BGR and LBEG WMS servers2). A similar general observation that a cooccurrence of high bi-carbonate and low sulfate concentrations can be an indication for high methane concentrations has also been supposed for groundwater in England (Darling and Gooddy, 2006). However, very large variations (several orders of magnitude) within small distances between wells can be observed. No apparent

2 BGR Groundwater data: http://www.bgr.bund.de/EN/Themen/Wasser/ Produkte/produkte_node_en.html LBEG Soil and Hydrogeology data: http://nibis. lbeg.de/cardomap3/?TH¼636&lang¼en.

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

129

Table 1 Comparison of dissolved gases of different study areas. All values are given in [nl/l], bdl: below detection limit, a) detection limit cannot be re-calculated from given information, b) propane part of the analysis/study, but no values reported, c) shape of frequency distribution irregular, mode value would be meaningless, evalue/statistical parameter not computable.

Our study Lower Saxony

Molofsky et al., 2013 Susquehanna C.

Darrah et al., 2014 Marcellus Shale

Darrah et al., 2014 Barnett Shale

Reese et al., 2014 Sullivan County

Detection limit

# values

# bdl

Median

Mode

Minimum

Maximum

CH4 C2H6 C3H8 CH4 C2H6 C3H8

25 20 20 140e36,400 13e4500

1,043 268 82 1,331 217

0 746 960 371 1,485

2,579 57 23 925 149

402 21 24 168 c)

25 25 25 140 13

61,998,278 292,350 2,568 60,237,625 89,447

CH4 C2H6 C3H8 CH4 C2H6 C3H8 CH4 C2H6 C3H8

a) a) b) a) a) b) 7,000e36,400 2,500e17,000 3,700e19,400

111 64 e 59 59 e 221 0 0

3 50 e 0 0 e 1,662 1,883 1,883

980,000 3,400 e 8,400,000 450,000 e 222,739 e e

10,000 200 e c) c) e 151,295 e e

3,000 100 e 1,090,000 2,000 e 7,257 e e

155,500,000 1,738,900 e 71,480,000 7,870,000 e 40,205,113 e e

correlation between hydrocarbon concentration or isotopic composition of methane with measured field parameters (electric conductivity, oxygen or pH) was observed. 6. Conclusions and outlook This study provides “baseline” concentrations of dissolved hydrocarbon gases (methane-propane) in shallow groundwater aquifers of Lower Saxony in Northern Germany. These data will allow to better assess results of future monitoring demands

required by any deep drilling regulations. Apparently naturally occurring methane concentrations cover more than 6 orders magnitude although the majority of shallow groundwater wells exhibit low concentrations <1 ml/l. However, a small number of samples reveals concentrations in the range >10 ml/l or even exceeding the solubility limit at atmospheric pressure conditions (partial pressure of methane 1 bar). To correctly determine the dissolved gas concentrations of these wells modified sampling procedures have to be applied. Carbon stable isotope ratios of methane point to a biogenic

Fig. 10. Base map of Lower Saxony (geographical coordinates, WGS84) with superimposed concentration data. Green dots <1 ml/l, yellow dots 1 ml/l e 1 ml/l, red dots >1 ml/l, HB City of Bremen, HH City of Hamburg. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

130

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

origin of methane. At low concentrations the variability is high due to different processes involved (methane oxidation, mixing of different gases) or rapidly changing boundary conditions. However, our data set combines isotope ratios with measurements of dissolved methane at very low concentrations and the combination of both will allow a better interpretation if an addition of hydrocarbons with different origin to groundwater is suspected. Comparable to other studies we detected higher homologues (ethane and propane) in a subset of our samples. According to the standard Bernard-Diagram less than 15 samples would classify as thermogenic gas. However, the combination of methane d13C values and compositional ratios (dryness, ethane/propane) of these samples does not suggest a migration of deeper sourced gases. Ethane and propane could more likely be generated by microbial processes and simply represent a ubiquitous background. Consequently we recommend to apply the Bernard plot, which is only based on d13C(CH4) and one basic compositional ratio, solely to identify samples which would warrant further investigation. Interpretations which go a step further need to take into account occurrences and compositional ratios of higher homologues (e.g. butane is a common constituent of thermal gases) as well as the isotope composition of C2þ. Especially because the concentration of higher homologues (ethane and propane) is exceedingly low in our sample set potential concentration changes might proof as a more sensitive parameter than methane to detect possible migration of deeper sourced (thermally generated) hydrocarbons into a groundwater aquifer. Migrated hydrocarbons are occasionally associated with brine migration (salinity increase of the groundwater) or identified by inert noble gases and their stable isotope ratios (Myers, 2012; Darrah et al., 2014). Considering the special geological setting of Lower Saxony (Germany) and the major gas plays, molecular nitrogen and d15N2 (which is either depleted or enriched compared to atmospheric d15N2 depending on the source rock of nitrogen) might be used as proxy as it can be more easily determined as noble gas isotope ratios and concentrations. This ongoing project will include a second complete sampling of all sites at a different time of the year (spring vs. autumn) as well as sampling at frequent intervals of selected wells. Both will be necessary to evaluate potential short or long-term changes in methane, ethane and propane concentrations and isotopic compositions of methane. This would be mandatory to estimate threshold values (including sampling uncertainties) beyond which a significant influence based on anthropogenic activities (drilling, shale gas or geothermal energy hydraulic fracturing, gas production) could be confirmed. A major challenge will be the interpretation of the results across different hydrogeological units or smaller regional scales where strong variations have been observed. Acknowledgments We gratefully acknowledge the contribution of C. Reisener and her colleagues from the NLWKN to this study, who organized the sampling of 850 groundwater wells in Lower Saxony and T. Meyer and P. Thorhauer (LBEG) for providing additional samples. We wish €ger-Trampe and A. Larm for constructive discussions to thank J. Gro and J. Poggenburg, D. Laszinski and D. Graskamp for their assistance in analysis. We thank two anonymous reviewers for their thorough reviews which significantly contributed to improving the quality of the manuscript. Supplementary information The data published in this study will be made available to the public via the LBEG WMS data server.

LBEG Soil and Hydrogeology cardomap3/?TH¼636&lang¼en

data:

http://nibis.lbeg.de/

References Alperin, M.J., Reeburgh, W.S., Whiticar, M.J., 1988. Carbon and hydrogen isotope fractionation resulting from anaerobic methane oxidation. Glob. Biogeochem. Cycles 2, 279e288. Alperin, M.J., Hoehler, T.M., 2010. Anaerobic methane oxidation by archaea/sulfatereducing bacteria aggregates: 2. Isotopic constraints. Am. J. Sci. 309 (10), 958e984.
, G., Mink, J., Pa
dua, A.A.H., 2002. Bacsik, Z., Lopes, J.N.C., Gomes, M.F.C., Jancso Solubility isotope effects in aqueous solutions of methane. J. Chem. Phys. 116 (24), 10816e10824. Baldassare, F.J., McCaffrey, M.A., Harper, J.A., 2014. A geochemical context for stray gas investigations in the northern Appalachian Basin: implications of analyses of natural gases from Neogene-through Devonian-age strata. AAPG Bull. 98 (2), 341e372. Bernard, B., Brooks, J.M., Sackett, W.M., 1977. A geochemical model for characterization of hydrocarbon sources in marine sediments. Ninth Annu. Offshore Technol. Conf. Prepr. 3, 435e438. Berner, U., Faber, E., 1988. Maturity related mixing model for methane, ethane and propane, based on carbon isotopes. Org. Geochem. 13, 67e72. Binot, F., Gerling, P., Hiltmann, W., Kockel, F., Wehner, H., 1993. The petroleum system in the Lower Saxony Basin. In: Spencer, A.M. (Ed.), Generation, Accumulation, and Production of Europe's Hydrocarbons. Springer, Berlin, pp. 121e139. Capasso, G., Inguaggiato, S., 1998. A simple method for the determination of dissolved gases in natural waters. An application to thermal waters from Vulcano Island. Appl. Geochem. 13 (5), 631e642. Carroll, A.R., 2015. Geofuels - Energy and the Earth. Cambridge University Press. Chanton, J., Liptay, K., 2000. Seasonal variation in methane oxidation in a landfill cover soil as determined by an in situ stable isotope technique. Glob. Biogeochem. Cycles 14 (1), 51e60. Cheung, K., Klassen, P., Mayer, B., Goodarzi, F., Aravena, R., 2010. Major ion and isotope geochemistry of fluids and gases from coal-bed methane and shallow groundwater wells in Alberta, Canada. Appl. Geochem. 25 (9), 1307e1329. Chung, H.M., Gormly, J.R., Squires, R.M., 1988. Origin of gaseous hydrocarbons in subsurface environments: theoretical considerations of carbon isotope distributions. Chem. Geol. 71, 97e103. Clark, I., Fritz, P., 1997. Environmental Isotopes in Hydrogeology. CRC Press LLC. Clayton, C., 1991. Carbon isotope fractionation during natural gas generation from kerogen. Mar. Pet. Geol. 8, 232e240. Coleman, D.D., Liu, C.-L., Riley, K.M., 1988. Origins of methane in the Earth microbial methane in the shallow Paleozoic sediments and glacial deposits of Illinois, U.S.A. Chem. Geol. 71 (1), 23e40. COGCC Colorado Oil and Gas Conservation Commission Orders: https://cogcc.state. co.us/orders/orders/112/156.html, https://cogcc.state.co.us/orders/orders/112/ 157.html, last assessed November 03rd, 2015. Coplen, T.B., 2011. Guidelines and recommended terms for expression of stableisotope-ratio and gas-ratio measurement results. Rapid Commun. Mass Spectrom. 25 (17), 2538e2560. Darling, W.G., Gooddy, D.C., 2006. The hydrogeochemistry of methane: evidence from English groundwaters. Chem. Geol. 229 (4), 293e312. 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. Proc. Natl. Acad. Sci. U. S. A. 111 (39), 14076e14081. Davies, R.J., 2011. Methane contamination of drinking water caused by hydraulic fracturing remains unproven. Proc. Natl. Acad. Sci. U. S. A. 108 (43). E871, author reply E872. Davies, R.J., Mathias, S.A., Moss, J., Hustoft, S., Newport, L., 2012. Hydraulic fractures: how far can they go? Mar. Pet. Geol. 37 (1), 1e6. Duan, Z., Møller, N., Greenberg, J., Weare, J.H., 1992. The prediction of methane solubility in natural waters to high ionic strength from 0 to 250 C and from 0 to 1600 bar. Geochim. Cosmochim. Acta 56 (4), 1451e1460. Dyck, W., Dunn, C.E., 1986. Helium and methane anomalies in domestic well waters in southwestern Saskatchewan, Canada, and their relationship to other dissolved constituents, oil and gas fields, and tectonic patterns. J. Geophys. Res. Solid Earth 91 (B12), 12343e12353. €ume in Elbracht, J., Meyer, R., Reutter, E., 2010. Hydrogeologische R€ aume und Teilra Niedersachsen. Geoberichte 3, 3-117, Landesamt für Bergbau, Energie und Geologie, ISSN 1864e7529. Eltschlager, K.K., Hawkins, J.W., Ehler, W.C., Baldassare, Fred, 2001. Technical measures for the investigation and mitigation of fugitive methane hazards in areas of coal mining. Off. Surf. Min. Reclam. Enforc. 125. ERCB (Energy Resources Conservation Board), 2006. Directive 35: Baseline Water Well Testing Requirement for Coalbed Methane Wells Completed above the Base of Protection. ERCB, Calgary, Alberta, Canada. May 8, 2006. EU, 2014. Commission Recommendation of 22 January 2014 on minimum principles for the exploration and production of hydrocarbons (such as shale gas) using high-volume hydraulic fracturing (2014/70/EU). Off. J. Eur. Union 39, 72e78. Faber, E., Stahl, W., 1984. Geochemical surface exploration for hydrocarbons in

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132 North Sea. AAPG Bull. 68, 363e368. Faber, E., Whiticar, M.J., Gerling, P., 1999. Comparison of hydrocarbons from unconventional sources: KTB, EPR and bit metamorphism. Geol. Jahrb. D 107, 175e194. Fisher, K., Warpinski, N., 2012. Hydraulic-fracture-height growth: real data. SPE Prod. Oper. 27 (1), 8e19. Richardson, Texas: Society of Petroleum Engineers, SPE 145949. Flewelling, S.A., Tymchak, M.P., Warpinski, N., 2013. Hydraulic fracture height limits and fault interactions in tight oil and gas formations. Geophys. Res. Lett. 40 (14), 3602e3606. Fuex, A.N., 1980. Experimental evidence against an appreciable isotopic fractionation of methane during migration. Adv. Org. Geochem. 12, 725e732. Galimov, E.M., 1988. Sources and mechanisms of formation of gaseous hydrocarbons in sedimentary rocks. Chem. Geol. 71, 77e95. Gorody, A.W., 2012. Factors affecting the variability of stray gas concentration and composition in groundwater. Environ. Geosci. 19 (1), 17e31.
nez, N., Thielemann, T., Straaten, N., Lüders, T., Richnow, H.H., Gruendger, F., Jime Krüger, M., 2015. Microbial methane formation in deep aquifers of a coalbearing sedimentary basin. Ger. Front. Microbiol. 6, 200. http://dx.doi.org/ 10.3389/fmicb.2015.00200. Harting, P., Schütze, H., Christoph, G., 1976. Der thermodynamische Kohlenstoffisotopieeffekt im System CH4eH2O. Isot. Isot. Environ. Health Stud. 12 (6), 232e234. 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 (4), 511e524. Hinrichs, K.-U., Hayes, J.M., Bach, W., Spivack, A.J., Hmelo, L.R., Holm, N.G., Johnson, C.G., Sylva, S.P., 2006. Biological formation of ethane and propane in the deep marine subsurface. Proc. Natl. Acad. Sci. U. S. A. 103 (40), 14684e14689. Hunt, J.M., 1996. Petroleum Geochemistry and Geology. W.H. Freeman. IEA, 2012. Golden Rule for a Golden Age of Gas. World Energy Outlook Special Report on Unconventional Gas. IEA Publications, Paris, p. 143. Jackson, R.B., Down, A., Phillips, N.G., Ackley, R.C., Cook, C.W., Plata, D.L., Zhao, K., 2014. Natural gas pipeline leaks across Washington. DC. Environ. Sci. Technol. 48 (3), 2051e2058. Jackson, R.B., Osborn, S.G., Vengosh, A., Warner, N.R., 2011. Reply to Davies: hydraulic fracturing remains a possible mechanism for observed methane contamination of drinking water. Proc. Natl. Acad. Sci. 108 (43), E872. Jackson, R.B., Vengosh, A., Darrah, T.H., Warner, N.R., Down, A., Poreda, R.J., Osborn, S.G., Zhao, K., Karr, J.D., 2013b. Increased stray gas abundance in a subset of drinking water wells near Marcellus shale gas extraction. Proc. Natl. Acad. Sci. 110 (28), 11250e11255. Jackson, R.E., Gorody, A.W., Mayer, B., Roy, J.W., Ryan, M.C., Van Stempvoort, D.R., 2013a. Groundwater protection and unconventional gas extraction: the critical need for field-based hydrogeological research. Groundwater 51 (4), 488e510. Jaekel, U., Musat, N., Adam, B., Kuypers, M., Grundmann, O., Musat, F., 2013. Anaerobic degradation of propane and butane by sulfate-reducing bacteria enriched from marine hydrocarbon cold seeps. ISME J. 7 (5), 885e895.
, G., 2002. Interpretation of isotope effects on the solubility of gases. Jancso Nukleonika 47, 53e57. Kampbell, D.H., Vandegrift, S.A., 1998. Analysis of dissolved methane, ethane, and ethylene in ground water by a standard gas chromatographic technique. J. Chromatogr. Sci. 36 (5), 253e256. Kang, M., Kanno, C.M., Reid, M.C., Zhang, X., Mauzerall, D.L., Celia, M.A., Chen, Y., Onstott, T.C., 2014. Direct measurements of methane emissions from abandoned oil and gas wells in Pennsylvania. Proc. Natl. Acad. Sci. 111 (51), 18173e18177. Kinnaman, F.S., Valentine, D.L., Tyler, S.C., 2007. Carbon and hydrogen isotope fractionation associated with the aerobic microbial oxidation of methane, ethane, propane and butane. Geochim. Cosmochim. Acta 71 (2), 271e283. Kniemeyer, O., Musat, F., Sievert, S.M., Knittel, K., Wilkes, H., Blumenberg, M., Michaelis, W., Classen, A., Bolm, C., Joye, S.B., Widdel, F., 2007. Anaerobic oxidation of short-chain hydrocarbons by marine sulphate-reducing bacteria. Nature 449, 898e902. Kockel, F., Wehner, H., Gerling, P., 1994. Petroleum systems of the Lower Saxony Basin, Germany. AAPG Mem. 60, 573e586. Krooss, B.M., Littke, R., Mueller, B., Frielingsdorf, J., Schwochau, K., Idiz, E.F., 1995. Generation of nitrogen and methane from sedimentary organic matter: implications on the dynamics of natural gas formation. Chem. Geol. 126, 291e318. Laier, T., 2012. Results of monitoring groundwater above the natural gas underground storage at Stenlille, Denmark. Geol. Surv. Den. Greenl. Bull. 26, 45e48. LBEG 2014. http://www.lbeg.niedersachsen.de/download/82740/Liste_Hydraulic fracturing-Massnahmen.pdf (last accessed 15/10/2015). Li, H., Carlson, K.H., 2014. Distribution and origin of groundwater methane in the Wattenberg oil and gas field of Northern Colorado. Environ. Sci. Technol. 48 (3), 1484e1491. Littke, R., Krooss, B., Frielingsdorf, J., Idiz, E., 1995. Molecular nitrogen in natural gas accumulations: generation from sedimentary organic matter at high temperatures. AAPG Bull. 79 (3), 410e430. Lorant, F., Prinzhofer, A., Behar, F., Huc, A.Y., 1998. Carbon isotopic and molecular constraints on the formation and the expulsion of thermogenic hydrocarbon gases. Chem. Geol. 147, 249e264. McMahon, P.B., Caldwell, R.R., Galloway, J.M., Valder, J.F., Hunt, A.G., 2015. Quality

131

and age of shallow groundwater in the Bakken Formation production area, Williston Basin, Montana and North Dakota. Ground Water 53 (Suppl. 1), 81e94. Melchers, C., 2008. Methan im südlichen Münsterland e Genese, Migration und €lische Wilhelms-Universita €t, Münster, Gefahrenpotential. Ph.D. Thesis. Westfa p. 153. 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 (3), 333e349. Molofsky, L.J., Connor, J.A., Farhat, S.K., Wylie, A.S., Wagner, T., 2011. Methane in Pennsylvania water wells unrelated to Marcellus shale fracturing. Oil Gas J. 109 (49), 54e93. Murphy, E.M., Schramke, J.A., Fredrickson, J.K., Bledsoe, H.W., Francis, A.J., Sklarew, D.S., Linehan, J.C., 1992. The influence of microbial activity and sedimentary organic carbon on the isotope geochemistry of the Middendorf Aquifer. Water Resour. Res. 28 (3), 723e740. Myers, T., 2012. Potential contaminant pathways from hydraulically fractured shale to aquifers. Ground Water 50 (6), 872e882. Nakata, E., Tanaka, S., Suenaga, H., Mahara, Y., Nakamura, T., Nakagawa, K., Ohta, T., 2012. Stable carbon isotope behavior of natural seepage of deep underground (super 13) C-rich methane detected along a fault zone and adsorbed in mudstone; Tokyo Bay area. Jpn. Appl. Geochem. 27 (9), 1710e1723. Oremland, R.S., 1981. Microbial formation of ethane in anoxic estuarine sediments. Appl. Environ. Microbiol. 42 (1), 122e129. Oremland, R.S., Whiticar, M.J., Strohmaier, F.E., Kiene, R.P., 1988. Bacterial ethane formation from reduced, ethylated sulfur compounds in anoxic sediments. Geochim. Cosmochim. Acta 52 (7), 1895e1904. Osborn, S.G., Vengosh, A., Warner, N.R., Jackson, R.B., 2011a. Methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing. Proc. Natl. Acad. Sci. U. S. A. 108 (20), 8172e8176. Osborn, S.G., Vengosh, A., Warner, N.R., Jackson, R.B., 2011b. Reply to Saba and Orzechowski and Schon: methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing. Proc. Natl. Acad. Sci. 108 (37), E665eE666. PA code: http://www.pacode.com/secure/data/025/chapter78/chap78toc.html, last assessed Nov 03rd, 2015. Pierau, R., Ladage, S., Franke, D., Andruleit, H., Rogalla, U., 2013. Schiefergas; potenzial in Deutschland. Shale gas; potential for Germany. Geowiss. Mittl. 51, 6e14. Reese, S.O., Neboga, V.V., Pelepko, S., Kosmer, W.J., Beattie, S., 2014. Groundwater and Petroleum Resources of Sullivan County, Pennsylvania. Pennsylvania Geological Survey, Harrisburg.
v e
sz, K.M., Breen, K.J., Baldassare, A.J., Burruss, R.C., 2010. Carbon and hydrogen Re isotopic evidence for the origin of combustible gases in water-supply wells in north-central Pennsylvania. Appl. Geochem. 25 (12), 1845e1859. Rooney, M.A., Claypool, G.E., Chung, H.M., 1995. Modelling thermogenic gas generation using carbon isotope rations of natural gas hydrocarbons. Chem. Geol. 126, 219e232. Rowe, D., Muehlenbachs, A., 1999. Low-Temperature thermal generation of hydrocarbon gases in shallow shales. Nature 398, 61e63. Roy, J.W., Ryan, M.C., 2010. In-well degassing issues for measurements of dissolved gases in groundwater. Ground Water 48 (6), 869e877. Roy, J.W., Ryan, M.C., 2013. Effects of unconventional gas development on groundwater; a call for total dissolved gas pressure field measurements. Ground Water 51 (4), 480e482. Saba, T., Orzechowski, M., 2011. Lack of data to support a relationship between methane contamination of drinking water wells and hydraulic fracturing. Proc. Natl. Acad. Sci. 108 (37), E663. Sackett, W.M., 1984. Determination of kerogen maturity by the pyrolysis-carbon isotope method. Org. Geochem. 6, 359e363. €sser: Charakterisierung und Scherer, E., 2000. . Methanhaltige reduzierte Grundwa Aufbereitung mittels Desorption und biologischer Oxidation in Schnellfiltern. €ster, Berlin, Wissenschaftliche Schriftenreihe Umwelttechnik 9. Verlag Dr, Ko ISBN 3-89574-362-3. Schmitt, M., Faber, E., Botz, R., Stoffers, P., 1991. Extraction of methane from seawater using ultrasonic vacuum degassing. Anal. Chem. 63 (5), 529e532. Schobert, B., Elstner, E.F., 1980. Production of hexanal and ethane by phaedactylum triconutum and its correlation to fatty acid oxidation and bleaching of photosynthetic pigments. Plant Physiol. 66, 215e219. Schoell, M., 1980. The hydrogen and carbon isotopic composition of methane from natural gases of various origins. Geochim. Cosmochim. Acta 44, 649e661. Schoell, M., 1983. Genetic characterization of natural gas. AAPG Bull. 67, 2225e2238. Schon, S.C., 2011. Hydraulic fracturing not responsible for methane migration. Proc. Natl. Acad. Sci. 108 (37), E664. Stahl, W.J., 1979. Carbon isotopes in petroleum geochemistry. Lect. Isot. Geol. 274e282. Taylor, S.W., Sherwood Lollar, B., Wassenaar, I., 2000. Bacteriogenic ethane in nearsurface aquifers: implications for leaking hydrocarbon well bores. Environ. Sci. Technol. 34 (22), 4727e4732. Tilley, B.J., Muehlenbachs, K., 2011. Fingerprinting of gas contaminating groundwater and soil in a Petroliferous Region, Alberta, Canada. In: Environmental Forensics: Proceedings of the 211 INEF Conference, pp. 115e125. Tissot, B.P., Welte, D.H., 1984. Petroleum Formation and Occurrence. Springer-Verlag, Berlin; New York.

132

S. Schloemer et al. / Applied Geochemistry 67 (2016) 118e132

UBA, 2014. Umweltauswirkungen von Fracking bei der Aufsuchung und Gewinnung €tten Teil 2-Grundwassermovon Erdgas insbesondere aus Schiefergaslagersta nitoringkonzept, Frackingchemikalienkataster, Entsorgung von Flowback, For€t, schungsstand zur Emissions- und Klimabilanz, induzierte Seismizita Naturhaushalt, Landschaftsbild und biologische Vielfalt. Umweltbundesamt Dessau, Report No. UBA-FB 00. ISSN: 1862-4804, p. 619. US EPA, 2002. Technical Guidance for the Natural Attenuation Indicators: Methane, Ethane and Ethene. Revision 1, p. 18. http://www.epa.gov/region1/measure/ Natatten.pdf. 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 (2), 187e199. 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 (15), 8334e8348. Warner, N.R., Jackson, R.B., Darrah, T.H., Osborn, S.G., Down, A., Zhao, K., White, A., Vengosh, A., 2012. Geochemical evidence for possible natural migration of Marcellus Formation brine to shallow aquifers in Pennsylvania. Proc. Natl. Acad. Sci. U. S. A. 109 (30), 11961e11966. Whiticar, M.J., Faber, E., 1986. Methane oxidation in sediment and water column environmentsdisotope evidence. Org. Geochem. 10 (4), 759e768.

Whiticar, M.J., 1990. A geochemical perspective of natural gas and atmospheric methane. Org. Geochem. 16, 531e547. Whiticar, M.J., 1994. Correlation of natural gases with their sources. In: Magoon, L.B., Dow, W.G. (Eds.), The Petroleum System from Source to Trap, AAPG Memoir, 60, pp. 261e283. Whiticar, M.J., 1996. Stable isotope geochemistry of coals, humic kerogens and related natural gases. Int. J. Coal Geol. 32, 191e215. Whiticar, M.J., 1999. Carbon and hydrogen isotope systematics of bacterial formation and oxidation of methane. Chem. Geol. 161, 291e314. Whiticar, M.J., Faber, E., Schoell, M., 1986. Biogenic methane formation in marine and freshwater environments: CO2 reduction versus acetate fermentation isotope evidence. Geochim. Cosmochim. Acta 50, 693e709. Wiesenburg, D.A., Guinasso, N.L., 1979. Equilibrium solubilities of methane, carbon monoxide, and hydrogen in water and sea water. J. Chem. Eng. Data 24 (4), 356e360. Xie, S., Lazar, C.S., Lin, Y.-S., Teske, A., Hinrichs, K.-U., 2013. Ethane- and propaneproducing potential and molecular characterization of an ethanogenic enrichment in an anoxic estuarine sediment. Org. Geochem. 59, 37e48.
-Mora, R., Michaelis, W., Widdel, F., 1999. Zengler, K., Richnow, H.H., Rossello Methane formation from long-chain alkanes by anaerobic microorganisms. Nature 401, 266e269.