OR/17/049 Water
| Ward, R S1, Smedley, P L1, Allen, G2, Baptie, B J1, Daraktchieva, Z3, Horleston, A7, Jones, D G1, Jordan, C J1, Lewis, A4, Lowry, D5, Purvis, R M4, and Rivett, M O6. 2014. Environmental baseline monitoring project: phase II - final report. British Geological Survey Internal Report, OR/17/049.
(1) British Geological Survey, (2) University of Manchester, (3) Public Health England, (4) University of York (National Centres for Atmospheric Science), (5) Royal Holloway University of London, (6) University of Birmingham, (7) University of Bristol |
Introduction
Rationale for monitoring
Public concerns about shale-gas exploration and development have included the potential impacts on surface waters and groundwater and particularly the quality of drinking water. Concerns include contamination by hydrocarbon gases, produced water and hydraulic fracturing fluid. Potential pathways could include new or reactivated faultlines arising as a result of induced seismicity, well-integrity problems (poor well completion or long-term corrosion) or surface spills/leaks. Environmental safeguards need to be in place to ensure the risk of these is minimised and water-quality monitoring throughout the life cycle of the borehole provides assurance of those safeguards. The Operator has a remit to monitor water quality in the area around KM8. BGS’s monitoring activities have been additional and independent.
Although a number of studies in countries where shale gas is a developed industry have inferred a link between groundwater quality and hydrocarbon extraction (Osborn et al. 2011[1]; Llewellyn et al., 2015[2]; Jackson et al., 2013[3]), establishing a causal relationship is difficult without evidence of the pre-development baseline chemical composition. This can provide a more robust basis for establishing whether any water-quality changes are related to exploration and development. This chapter outlines the investigations carried out so far to establish the baseline chemistry of surface water and groundwater in the Vale of Pickering and covers key diagnostic analyses including pH and redox characteristics, major ions, trace elements, stable isotopes, dissolved methane and organic compounds, especially hydrocarbons. The analyses provide a detailed time series of water chemistry data, spanning up to 17 months.
Regional geology and hydrogeology
Background information on the geographical, geological and hydrogeological setting of the Vale of Pickering is given by Bearcock et al. (2015)[4], Ford et al. (2015)[5] and Newell et al. (2016)[6] but a brief overview is provided here. The vale constitutes a broad flat-lying plain bordered by higher ground extending towards the North York Moors in the north, Howardian and Hambleton Hills in the west and Yorkshire Wolds in the south-east (Figure 1). The vale forms the catchment of the River Derwent. Landuse is rural and dominated by arable farming with some livestock: clay soils support dominantly cereal crops in the western part of the vale and root crops and oil seed rape in the east.
The shallow rock formations of the Vale of Pickering contain two main aquifers. A covering layer of superficial Quaternary sediments, of mainly glaciolacustrine origin with marginal alluvial fan deposits, forms a shallow aquifer across the valley floor. Much of this material was the product of a former proglacial lake, Lake Pickering, which occupied the vale during the Devensian until ice retreat some 10 000 years ago (Evans et al. 2017[7]). Borehole records show that these sediments are of variable thickness, typically less than 40 m thick, but are relatively thin or absent in the north and western part of the vale, and close to Kirby Misperton (Figure 2). In the vicinity of Kirby Misperton and on the northern flanks of the vale, discontinuous caps of Devensian glacial till occur which form local topographic features. Ford et al. (2015)[5] also identified relatively permeable sand and gravel horizons derived from coalesced fans along the margins of the vale. Numerous discontinuous lenses of sand/gravel also appear elsewhere throughout the Quaternary profile.
The Quaternary deposits overlie a thick sequence of Jurassic Ampthill & Kimmeridge Clay which, though clay-dominant and poorly permeable, contains some sandy horizons and weathered upper sections which can contain locally significant amounts of groundwater. Together with the overlying Quaternary deposits, these sustain small-scale private supplies locally. Abstraction is predominantly for agricultural and domestic use; none is used for drinking water. As depths and construction of some water-supply boreholes are poorly understood, the geological formation contributing to water supply is often uncertain. Contribution of the Kimmeridge deposits is likely to be proportionately larger in the west where the Quaternary deposits are thinner and less well-developed. These water-bearing units are hereinafter referred to collectively as the ‘Superficial’ aquifer.
Groundwater levels in the Superficial aquifer are shallow and artesian in parts; flow broadly follows the topographic gradient of the valley bottom. The small inliers of till influence shallow groundwater flow locally but are unlikely to contain groundwater themselves. Flow from the central part of the vale near Kirby Misperton and the proposed shale gas site appears to be southwards towards the lower part of the Derwent catchment. Groundwater levels in the Superficial aquifer across the vale are shown in Figure 3.
Limestones of the Jurassic Corallian Group constitute a second important aquifer in the region. This is defined as a Principal Aquifer and provides a source of water for both public and private supply, though is only exploited along the margins of the vale (Figure 1). Within the vale itself, the Corallian Limestone is not used as a water resource because of its depth (typically 200 m+ in the central area) and unfavourable quality.
The Corallian aquifer is recharged in the outcrop area along the margins of the Vale of Pickering and groundwater flows downgradient into the confined section. The upper reaches of the River Derwent provide recharge to the Corallian aquifer in the north-east, although further downstream, groundwater feeds the river. Groundwater is abstracted from the Corallian for public supply, agriculture and industry (Bearcock et al., 2015[4]).
Investigations of the deeper geology of the Vale of Pickering, focussed on the post-Permian structure (Newell et al., 2016[6]), have provided support for a strong structural control on the disposition of the Corallian aquifer and adjacent strata. Relatively abundant faults have dissected the Corallian into numerous blocks, with a predominant east-west orientation. This strong fracturing presumably disrupts the flow of groundwater in the confined sections of the aquifer, leading to likely compartmentalisation of the aquifer.
A further Principal aquifer, the Chalk, occurs to the south-east of the Vale of Pickering at some 10 km or more from the KM8 site (Figure 1). The aquifer constitutes a major resource in the region further south (Reeves et al., 1978[8]; Tattersall and Wilkinson, 1974[9]).The northern escarpment of the Chalk forms an unconfined aquifer which sustains a number of private supplies. As the Chalk is more distant from the KM8 site, it does not form a component of the current investigation.
Water monitoring
Site selection
A monitoring network of streams and pre-existing groundwater (borehole, well) sites across the Vale of Pickering has been established and is described in detail by Smedley et al. (2015)[10]. The network is depicted in Figure 4 and comprises 25 groundwater sites from the Superficial and Corallian aquifers (the latter along the basin margins in the unconfined or feather edge of confined sections), as well as 10 low-order stream sites. Sites have been monitored for 12 months at monthly intervals since September 2015 and quarterly thereafter. Following discussion with local residents, a further two stream monitoring sites are being investigated for augmentation of the stream network. These are close to KM8 and if safe and accessible, will be added to the network in future sampling campaigns.
Sampling and analysis
Streams have been monitored by, where possible, emplacement of sensors (pH, temperature, redox potential (Eh), specific electrical conductance (SEC), dissolved oxygen (DO)) in the streamline. Where not possible, a more limited suite of analytes have been measured in a container. Water samples for major-ion and trace-element analysis in the laboratory were filtered (0.22 µm) into factory-new LDPE bottles. A separate aliquot was collected for Hg (mercury) analysis. Samples for major and trace-element analysis were acidifed (1% v/v) with pure HNO3 (nitric acid) and stored on ice and/or in a refrigerator until analysis. These were further acidified with 0.5% HCl (hydrochloric acid) in the laboratory before analysis. Following standard procedure, samples for anion analysis and ammonium (NH4) were not acidified. Filtered (0.45 µm) samples were also collected for measurement of non-purgeable organic carbon (NPOC) using a silver-impregnated filter and glass syringe. Unfiltered samples were collected for analysis of dissolved radon as well as for polyaromatic hydrocarbons (PAHs), volatile and semi-volatile organic compounds (VOCs, SVOCs) and total petroleum hydrocarbons (TPH).
Sampling of groundwater was carried out as close as possible to the wellhead (point of emergence of groundwater at the surface) to enable representative samples to be taken. Boreholes were purged and discharging groundwater monitored for DO, temperature, Eh, pH and SEC until stable readings were obtained, at which point, samples were collected for laboratory analysis. pH, Eh and DO were measured in a flow cell to avoid groundwater aeration. Samples for dissolved gases (CH4, CO2 and C2H6) were collected under pressure in tightly-sealed steel sampled containers. Samples for major-ion and trace-element analysis and NPOC were collected as for stream waters. Inline filters were used where possible to minimise sample aeration.
Analysis of major cations and trace elements was carried out by ICP-MS. Analysis of anions was by ion chromatography. ICP-MS and IC methods are certified to ISO 17025:2005. Aliquots for Hg analysis were measured by a commercial laboratory by ICP-MS using an ISO 17025-certified method. NH4 was determined by colorimetry, dissolved gases by gas chromatography and stable isotopes by mass spectrometry. Dissolved radon has been measured by three different methods in experiments designed to obtain the most robust methodology. The analysis methods were by Alphaguard, portable liquid scintillation and gamma spectrometry. Results reported here are by the latter method.
Samples for Total Petroleum Hydrocarbons (TPH) were solvent-extracted and hydrocarbons with carbon banding in the range C8-C40 determined by GC-FID, while Gasoline Range Organics (GRO) in the carbon chain range of C4-12 were determined by headspace GC-FID. Samples for SVOC were solvent-extracted and analysed by GC-MS. The testing laboratory holds ISO 17025 accreditation for TPH CWG and SVOC in surface water.
VOC determination was by direct aqueous injection purge-and-trap GC-MS; PAH determination by solvent extraction followed by GC-MS. The testing laboratory for VOC and PAH determinations holds ISO 17025:2005 accreditation for these methods in water.
In addition, a non-accredited (semi-quantitative) analysis of water samples by target-based screening has also been determined using GC-MS and LC-MS. This method provides determinations of analytes including some pesticides, plasticisers, perfluorinated compounds, pharmaceutical drugs and chlorinated solvents.
Real-time monitoring
Sensors for real-time monitoring of groundwater quality have been installed in four of the new BGS boreholes. These monitor physico-chemical variables (water level, pH, temperature, barometric pressure, SEC) in the groundwater at hourly intervals and data are telemetered to BGS for display on the BGS website (www.bgs.ac.uk/valeofpickering). Sondes are retrieved periodically (approximately every other month) for inspection, maintenance and recalibration.
Results
Spatial variability
Groundwater from the Superficial aquifer shows a clear dominance of Na-HCO3 compositions (Figure 5) while that from the unconfined Corallian is of Ca-HCO3 type as expected for groundwater derived from limestone. Some confined Corallian groundwaters have mixed-ion compositions, trending towards Na-Cl type, likely in response to ion exchange under confined conditions. Streams have dominantly Ca-HCO3 compositions but Na-HCO3 types are also represented, these dependent on the lithology of the underlying rock types (Figure 5).
Box plots also highlight the differences between groundwater compositions from the two aquifers. Figure 6 shows some of the redox characteristics of groundwaters from the monitoring network. Groundwater from the Superficial aquifer is overwhelmingly anoxic, with DO <1 mg/L, NO3 usually below or close to detection limit and often high concentrations of NH4, Fe and Mn. Dissolved CH4 shows a large range, with a high maximum of around 50 mg/L. For the Corallian aquifer, Figure 6 shows unconfined and confined groundwaters plotted together and shows that compositions have a range of redox conditions, with DO varying from saturated to close to detection limit. Nitrate concentrations show a correspondingly large range. Concentrations of Fe, Mn and NH4 are usually low (e.g. mostly below drinking-water limits as a reference guide). Concentrations of CH4 in the Corallian groundwaters (confined and unconfined around the margins of the vale) have relatively low dissolved CH4 concentrations.
Figure 7 shows the spatial variability in dissolved CH4 concentrations for groundwaters from both the monitoring network and the new BGS and TE boreholes. For most of the area, the distributions of CH4 show no clear relation with the location of current and abandoned hydrocarbon wells in the vale, though an association cannot be ruled out for the immediate vicinity of KM8. No relation is seen either with the path of the gas pipeline connecting the conventional well sites. Indeed, many sampled wells close to this line have low CH4 concentrations. The pump depths of many of the sampled groundwater boreholes are larger than would be expected for the buried pipeline, such that a connection between the two is unlikely.
Stable carbon-isotopic data for the dissolved methane in the Superficial groundwaters appear to indicate a predominantly biogenic signature (δ13C-CH4 in the range -44‰ to -87‰ VPDB, but almost entirely in the -68 to -87‰ range). This is supported by dissolved molar C1/C2 ratios of ≥100. The origin of the CH4 is still under investigation but a provenance in the shallowest parts of the Kimmeridge Clay is considered most likely. One of the TE boreholes completed in the confined Corallian also has a high CH4 concentration (ca.50 mg/L). This too has a biogenic carbon-isotopic composition (δ 13C-CH4 -69‰ VPDB, C1/C2>10 000) such that a deep hydrocarbon source is unlikely.
Distributions of radon in the Superficial and Corallian groundwaters from one sampling round are shown in Figure 8. Activities were found to be universally low (<20 Bq/L) with slightly higher values represented in the Corallian group. This is consistent with higher Rn values inferred from Rn in air maps (Section 5).
Temporal variability
Time series for selected physico-chemical analytes and inorganic constituents are shown in Figure 9–Figure 20 for the groundwaters from the Superficial and Corallian aquifers, both from the original monitoring network and the new Superficial BGS and TE boreholes, together with the streams. Variability in Eh and temperature is relatively large for some sites, though pH and SEC show much more consistency. For the major ions, Superficial and Corallian groundwaters show much more uniformity than the streams which are likely to vary in chemical composition as a result of variations in rainfall contribution to stream discharge.
Of the trace elements shown, Fe appears to show one of the largest variations in concentrations. Large variations in concentrations of dissolved CH4 are also apparent, though in the Superficial aquifer, concentrations are consistently high, in the mg/L range (Figure 17, Figure 18, Figure 19, Figure 20). The variability may reflect difficulties of sampling and degassing, despite considerable efforts taken to minimise this. The trends emphasise the commonly high CH4 concentrations in the groundwaters of the Superficial aquifer.
Data from the monitoring of the new BGS and TE boreholes in the Superficial aquifer (Figure 10, Figure 14, Figure 18) show similarities with those in the monitoring network for the same aquifer, with a couple of sites showing consistently high CH4 concentrations (10–40 mg/L). The TE borehole completed in the confined Corallian aquifer showed CH4 concentrations varying in the range 50–57 mg/L. Data from the BGS boreholes recently installed in the same geological formation are not yet available.
A subset of 10 samples of Superficial groundwater for Hg analysis showed all to be less than the 0.1 µg/L detection limit.
Of the organic compounds that have been monitored over the course of the project, almost none has shown concentrations above detection limits. Values for PAHs, TPH, phenols, semi-volatile compounds are almost universally non-detects.
From the target-based screening, some organic compounds have been detected but these are overwhelmingly of low concentration (usually <1 µg/L, typically orders of magnitude less). Variations in concentrations of selected compounds for one stream site, Site 9, show some consistent detections over the 1 year interval of monitoring (Figure 21). Although a semi-quantitative technique for the measurement of trace organics, this constitutes a potentially useful indicator of new organic contaminants that if identified and found likely to be significant, could be pursued using quantitative techniques.
Real-time monitoring
Measurements from the downhole real-time monitoring have suffered periodically from loss of battery power and other technical difficulties, but a discontinuous record of downhole physical and chemical measurements has been recorded since August 2016 and reported since November 2016. Data are displayed in Figure 22. Periods of instability have also caused problems. These may have been due to problems including coating of the membranes with iron oxides. Figure 22 also shows the intervals with large sudden changes (vertical spikes) when sondes were retrieved for maintenance and calibration. Some other variations have arisen from instrument drift. Some cleaning of the database is required to remove spurious values.
Short-term temporal variations in pH and SEC in EBM8 appear to be real and are being investigated further.
Statistical handling of time-series data
Monitoring data for dissolved CH4 and NH4 in groundwater from the Superficial aquifer are depicted Figure 23. The concentrations were transformed to natural logs before analysis. The analysis was done using a ‘hybrid’ estimator in which the spatial mean of the target variable across the domain (aquifer) is analysed treating the samples as random, and the correlation between the spatial means over time is modelled statistically (Brus, 2014[11]). Treating the spatial sampling as random is not justified by the sampling scheme in this case, but this assumption is made for simplicity and because of the relatively small sample size.
The target variable is the space-time mean log CH4 or NH4 concentration, i.e. the mean value over the spatial domain of interest over a 12-month period. The variance of the mean values, i.e. the mean square error of the estimate is a quality measure and could be used, for example, to compute a confidence interval for the estimate. This variance was computed for a range of notional numbers of monitoring sites, and for two to 12 sampling times (i.e. sampling intensity from 6-monthly to monthly). Figure 23 shows how increasing the number of sampling times and sites reduces the variance of the estimated space-time mean for both analytes. This allows an evaluation of a practically feasible combination of spatial and temporal sampling intensities for which the variance has an acceptable value. Figure 23 shows that for each analyte, the spatial variation dominates the uncertainty in the estimate and that there is a much larger reduction in variance from increasing the number of sites for a given number of times than vice versa. The response for NH4 is even less than for CH4.
There is clearly a rather small gain in precision of the estimate from increasing the frequency of temporal sampling beyond quarterly in this system. In practice however, a higher frequency of monitoring may be warranted, at least in the early stages of a monitoring scheme. There are two general reasons. First, information on the space-time variability of the properties of interest is initially limited, and it would be valuable to have more detailed information. Second, estimates of the space-time mean are not the only objective for a monitoring scheme. For example, more frequent sampling may be desirable to improve the detection of transient events or to allow more precise prediction of the spatial mean at unsampled times.
Summary and conclusions
Monitoring of streams and groundwater in the Superficial and Corallian aquifers for a period up to 17 months has established a baseline for a comprehensive suite of physical, organic and inorganic chemical analytes that provide the basis for assessing any future perturbations. The data indicate that water quality in the Superficial and Corallian aquifers is distinct as a result of distinctive geology and hydrogeology and that redox processes play a key role in defining the regional groundwater chemical signatures. From the analysed samples, major-ion composition of streamwater varies according to the underlying rock types and may have closer affinity to either Superficial or Corallian groundwater compositions, depending on stream course.
Chemical compositions show some consistency over time, in both surface water and groundwater, though the greatest variability occurs in stream samples. This is likely in response to seasonal variations in rainfall.
One of the key findings from the baseline monitoring has been the observations of high dissolved CH4 concentrations in a number of groundwaters from the Superficial aquifer. Concentrations in the Corallian around the margins of the vale are low (µg/L range). One borehole in the confined Corallian at KM8 has a higher CH4 concentration, in line with the highest values observed for this dissolved gas in the Superficial groundwaters. Carbon-isotopic evidence provided to date indicates that the CH4 in the shallow groundwaters of the Vale of Pickering has a biogenic rather than thermogenic signature, although further work is needed to support this conclusion more robustly. Isotopic analysis of the methane in the TE deep confined Corallian borehole is not available nor are the data (concentration and/or isotopes) from the recently installed BGS boreholes in the deep Corallian in the centre of the Vale of Pickering. Statistical evaluation of the monitoring data for CH4 produced to date indicates that spatial variation is a much greater source of uncertainty in derivation of the space-time mean than temporal variability and that reducing uncertainty can be achieved more effectively by increasing numbers of sites than increasing sampling frequency. Similar analysis for other parameters is on-going.
References
- ↑ Osborn, S G, Vengosh, A, Warner, N R, and Jackson, R B. 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. 2011;108(20):8172–6.
- ↑ Llewellyn, G T, Dorman, F, Westland, J L, Yoxtheimer, D, Grieve, P, and Sowers, T, et al. Evaluating a groundwater supply contamination incident attributed to Marcellus Shale gas development. Proceedings of the National Academy of Sciences of the United States of America. 2015;112(20):6325–30.
- ↑ Jackson, R B, Vengosh, A, Darrah, T H, Warner, N R, Down, A, and Poreda, R J, et al. 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. 2013;110(28):11250–5.
- ↑ 4.0 4.1 Bearcock, J B, Smedley, P L, and Milne, C J. 2015. Baseline groundwater chemistry: the Corallian of the Vale of Pickering, Yorkshire. BGS Open Report, OR/15/048.
- ↑ 5.0 5.1 5.2 Ford, J R, et al. 2015. The Vale of Pickering: an initial summary of the Quaternary/superficial geology and data holdings. BGS Open Report, OR/15/064.
- ↑ 6.0 6.1 Newell, A J, Ward, R S, and Fellgett, M W. 2016. A 3D geological model of post-Permian aquifers and aquitards in the Vale of Pickering, North Yorkshire, UK. Open Report OR/15/068 (in press).
- ↑ Evans, D J A, et al., 2017. Glacial Lake Pickering: stratigraphy and chronology of a proglacial lake dammed by the North Sea Lobe of the British–Irish Ice Sheet. Journal of Quaternary Science, 32(2), pp.295–310. Available at: http://dx.doi.org/10.1002/jqs.2833.
- ↑ Reeves, M J, Parry, E L, and Richardson, G. 1978. Preliminary investigation of the groundwater resources of the western part of the Vale of Pickering. Quarterly Journal of Engineering Geology, Vol.11, 253–262.
- ↑ Tattersall, K H, and Wilkinson, W B. 1974. Groundwater Investigation in the Vale of Pickering: Final report on preliminary investigation. Yorkshire River Authority and Water Resources Board.
- ↑ Smedley, P L, Ward, R S, Allen, G, Baptie, B, Daraktchieva, Z, Jones, D G, Jordan, C J, Purvis, R M and Cigna, F. 2015. Site selection strategy for environmental monitoring in connection with shale-gas exploration: Vale of Pickering, Yorkshire and Fylde, Lancashire. British Geological Survey Open Report OR/15/067.
- ↑ Brus, D J. 2014. Statistical sampling approaches for soil monitoring. European Journal of Soil Science 65, 779–791.