Cover Note

EPP6
APPEAL BY ISLAND GAS LTD, PORTSIDE ELLESMERE PORT

APPEAL REFERENCE APP/A0665/W/18/3207952

Inadequacies in the geological and geophysical investigations by IGas at Ellesmere Port, Cheshire

PROOF OF EVIDENCE by DAVID SMYTHE BSc PhD
(Emeritus Professor of Geophysics, University of Glasgow)

Contents

1 Introduction
1.1 Relevant personal details from my CV
1.2 Declaration of interest, independence and non-liability

2 Nature of the development
2.1 The definition of unconventional
2.2 Chert

3 Potential conduits for contamination
3.1 Fault modelling studies
3.2 Hydrocarbon seepages
3.3 Empirical evidence of faults as conduits
3.4 Wellbore failure as a conduit for contamination
3.5 Hydrogen sulphide
3.6 Discussion of potential conduits for contamination

4 Structure and stratigraphy at the wellsite
4.1 Introduction
4.2 The IGas cross-section
4.3 Velocity survey
4.4 Quality of seismic data in the vicinity of the well
4.5 The nature of the Appellant’s proposed acidisation
4.6 Regional faulting
4.7 Faulting in the neighbourhood of the wellsite
4.8 Shallow water supply boreholes
4.9 Discussion of the geology at the wellsite

5 Conclusions and recommendations
5.1 Summary of findings
5.2 Conclusion and recommendations

6 Appendix

Figures (within text)
Figure 2.1. Permeability of various reservoir rocks.
Figure 3.1. Mersey Special Protection Area.
Figure 4.1. North-south cross-section allegedly through EP-1.
Figure 4.2. Shotpoint map of the IG-14 2D survey lines.
Figure 4.3. Migrated version of part of seismic line IG-14-03.
Figure 4.4. Extract from BGS regional cross-section through Ince Marshes-1.
Figure 4.5. 1:50,000 solid geology map sheet compilation.
Figure 4.6. Unmigrated version of IG-14-03 with fault interpretation.
Figure 4.7. BGS borehole viewer map of the area around EP-1.
Figure 4.8. Comparison of stacked and migrated versions of part of seismic line IG-14-03.

Table 1. Six numerical modelling studies of flow up faults

1. Introduction

1.1 Relevant personal details from my CV

1.1.1 I am Emeritus Professor of Geophysics in the University of Glasgow. Although I am now a French resident I remain a British citizen, and take an active interest in UK, French and foreign affairs, as well as in various facets of scientific research.

1.1.2 My professional qualifications are: BSc Geology (Glasgow 1970), PhD Geophysics (Glasgow 1987); I was made a Chartered Geologist in 1991 but am no longer registered as such.

1.1.3 Prior to my taking up the Chair of Geophysics at the University of Glasgow in 1988 I was employed by the British Geological Survey (BGS) in Edinburgh, from 1973 to 1987. I was a research scientist, rising to the post of Principal Scientific Officer. In the 1990s I was closely involved in the search for a UK underground nuclear waste repository. I served on the BNFL Geological Review Panel from 1990 to 1991, to support BNFL’s case for a Sellafield site for a Potential Repository Zone (PRZ), at the time when Nirex was investigating both Dounreay and Sellafield.

1.1.4 I was closely involved with Nirex at this epoch, and conducted for Nirex an experimental 3D seismic reflection survey, which took place in 1994. The survey encompassed the volume of the proposed rock characterisation facility (RCF) – a deep underground laboratory planned as a precursor to actual waste disposal. This was a double world ‘first’ – the first ever 3D seismic survey of such a site, and the first academic group to use this method, which at the time was just emerging as an essential tool of the oil exploration industry.

1.1.5 I have published around 70 technical and scientific papers and reports (44 papers in the peer-reviewed literature). Since my retirement from the university in 1998 I have carried out private research, acted as a consultant to the oil industry, and maintained a professional interest in the geological problems raised by nuclear waste disposal, unconventional hydrocarbon exploration and coal-bed methane exploration.

1.1.6 While at the BGS I worked closely with the Department of Energy (DEn) and sometimes with the Foreign and Commonwealth Office. I was once invited to join the panel at which the DEn (predecessor in hydrocarbon regulation to the DTI, DECC, BEIS and the OGA) interviewed BP for a licence west of the UK. I sat on the regulatory side with the Chief Geologist and the Chief Geophysicist of DEn. Some 25 years later, during the period when I worked as an oil industry consultant, I sat at the other side of the table representing an Applicant for an onshore licence in the south of England. I am probably the only person who has ever sat on both sides of the table at PEDL award interviews.

1.2 Declaration of interest, independence and non-liability

1.2.1 I have no interests to declare. This proof of evidence was requested by Frack Free Ellesmere Port and Upton, which may pay me a modest honorarium. I am not connected to, nor am I a member of, any activist group, political party, or other organisation. I am solely responsible for the contents of this submission. It is supplied in good faith, but I can accept no liability resulting from any errors or omissions. The evidence in this proof of evidence is true to the best of my knowledge and belief and I confirm that the opinions expressed are my truly held professional opinions.

1.2.2 I was previously involved in a legal dispute with the University of Glasgow (2016-2018), but this has been settled amicably, and the Secretary of the University has stated (5 January 2018):
“I have no reason to doubt your integrity as a scientific researcher, and hope that you will continue to be as productive in your research as you have been since your retirement in 1998.”

1.2.3 He has also confirmed that I am free to continue to use the title of Emeritus Professor of Geophysics without hindrance. I remain a member of the College of Science and Engineering, but not attached to any specific school or group within the University, and the views expressed are my own.

2 Nature of the development

2.1 The definition of unconventional

2.1.1 The Appellant initially asserted that the chert rock to be targeted is a conventional target (EP23 letter of 5 November 2018). The Appellant stated:
“With regard to your request for confirmation that the proposed development is not for conventional oil and gas extraction, we can confirm that the Pentre Chert is a conventional resource. The Appellant will not be carrying out any drilling, hydraulic fracturing or acid fracturing activities as part of the proposed development.”

2.1.2 However, the Appellant has now accepted that the Pentre Chert is an unconventional target and in a letter on 22 November 2018 (EP23) stated as follows:
“With regard to the permeability of the Pentre Chert, this is of course a matter to be explored as part of the proposed development. However, as set out in Section 2.4 of the Planning Statement submitted with the planning application, the Pentre Chert consists of fractured cherts and siliceous shales. It is also confirmed in Section 5.2.1 of the Site Condition Report, submitted in support of the environmental permit application, that the Pentre Chert comprises interbedded chert, siliceous shale and occasional sand deposits. The Appellant does not disagree with your description of the Pentre Chert being a tough, brittle rock, though we would say that it has low permeability as opposed to not being permeable. However, the Pentre Chert encountered at Ellesmere Port is considered to be naturally fractured.”

2.1.3 There is no universally agreed definition of the difference between conventional and unconventional hydrocarbon mineral extraction; various versions in the scientific and technical literature emphasize different aspects. However, all reasonable definitions that I am aware of include, either implicitly or explicitly, the permeability of the host rock. The figure of 0.1 mD (milliDarcies) for the host rock is generally agreed to differentiate between the two extraction procedures. This is illustrated in Figure 2.1.

2.1.4 Next in importance to a quantitative definition using permeability comes the geological setting in which the hydrocarbon-bearing rock occurs. Thus conventional resources are found in finite and well-defined traps, whereas unconventional gas or oil is distributed throughout a widespread layer with no clear-cut boundaries.

2.2 Chert

2.2.1 Chert is a siliceous rock type which, like limestone, straddles the 0.1 mD dividing line of permeability in Figure 2.1. It has been targeted both as a conventional reservoir rock, e.g. the Mississippian ‘chat’ (local name for chert) of Kansas, which has a permeability of greater than 5 mD (Watney et al. 2001) [1], but also as an unconventional resource. One of the five or six main shale plays of North America, the Woodford Shale, is also known as the Woodford Chert.

fig 2.1
Figure 2.1. Permeability of various reservoir rocks. Chert permeability lies in the same range as limestone, straddling the conventional/unconventional divide.

2.2.2 Despite it being requested, the Appellant has failed to provide data on the permeability of the chert, or how the hydrocarbons have migrated into this otherwise unproductive material. However, based upon the discussion in the previous section, it can be concluded that the permeability of the chert in question is below 0.1 mD.

3 Potential conduits for contamination

3.1 Fault modelling studies

3.1.1 Leakage up faults on land is particularly important since it has the potential to contaminate the groundwater that we depend on for drinking, crops and livestock. Leakage also occurs offshore, e.g. in the North Sea, but is far from people and the natural mechanisms of the sea tend to deal with small leakages. The problem of leakage is particularly important in the UK, since the UK is 500 times more faulted than the US where groundwater contamination has been experienced. It follows that the UK is 500 times more likely, statistically speaking, to experience groundwater contamination from this source.

3.1.2 Table 1 summarises the principal results of the six model studies published to date that predict a time for contamination to ascend from a shale to a shallow aquifer. Note that it does not matter for the present purpose whether or not the shale has been, or is being fractured or acidised; it is the prediction of time of passage up a fault of either liquid or gas from a producing layer to the biosphere that is of relevance to Ellesmere Port.

3.1.3 Myers (2012) was the first to attempt quantitative hydrogeological modelling of a fault, and have it published in the peer reviewed literature. However, the German study discussed below, initially non-peer reviewed, is contemporary with Myers, and more comprehensive.
table 1
Table 1. Six numerical modelling studies of flow up faults

3.1.4 Myers[2] used the Marcellus Shale as a basis for simulation. He assumed a 30 m thick layer of shale and an overburden of 1500 m thickness, based on typical values for southern New York state. He assumed a homogeneous vertical fault 6 m wide traversing the overburden comprising a mixture of sandstone with subordinate components of shale, mudstone and limestone. The imposed head provided the driving force for flow.

3.1.5 Myers found that faults through the overburden could speed the upward travel time considerably in the steady-state model (before fracturing). When fracturing (or stimulation to increase the permeability) occurs the transport times of contaminated fluid from the fracked shale to the near surface can be reduced to a few tens of years “or less”. He argues for pre-stimulation fault mapping, a ‘setback’ distance between the stimulation zone and the nearest faults, and for a system of deep and shallow monitoring wells before development begins.

3.1.6 The comprehensive German study was published in German (Borchardt et al., 2012)[3] , and is therefore not widely known. The English summary of this report (Ewen et al., 2012)[4] does not provide details of the modelling. The report was later published as two peer reviewed papers, of which one (Kissinger et al., 2013)[5] deals with the modelling, but without all the detail of the original German-language report. The report deals in detail with passage of fluids up faults. Seven geological type-localities, or ‘settings’, were studied. One of them, Quakenbrück-Ortland in the Lower Saxony Basin, has a geological structure which is remarkably similar to the shale basins of the north of England. The modelling found that contaminated fluid could reach the groundwater resource zone in around 30 years.

3.1.7 Gassiat et al. (2013)[6] modelled the fluid transport up a 10 m wide fault zone; a width (the so-called ‘damage zone’) which they say is consistent with a regional fault having hundreds of metres of displacement. The fault is situated in a regional generic Canadian basin, with the shale being simulated having the properties of the Utica Shale. This shale has a low permeability; the value adopted is 10 nanodarcy for unfaulted shale. To contaminate the presumed shallow aquifer they find that the shale must be overpressured (a common occurrence in sedimentary basins, due to compression, burial, and/or subsequent uplift), and that it has to have been fracked. The timescale for the migration of contamination is of the order of 1000 years or less. The driving force for flow is overpressure in the shale, as is the case in the Bowland Shale of Lancashire.

3.1.8 They highlight some caveats in their modelling. An interesting and important result is that a tracer added to the shale fluid reaches the surface at 90% of its original concentration; in other words, ‘slugs’ of fluid travel upwards without getting significantly diluted. The ‘dilute and disperse’ model formerly used to justify ocean dumping of contaminants (for example, radioactive waste) evidently does not apply to migration of contaminants up a fault.

3.1.9 One modelling study (Cai and Ofterdinger, 2014)[7] concerns the Bowland Shale of Lancashire, of which the Holywell Shale and Pentre Chert are locally equivalent in Cheshire and Flintshire. The authors built a layer-cake computer geological model based on the geology at the Preese Hall-1 well, then added in hydraulic fractures in the Bowland Shale near the bottom of the model. No faults were built into the model initially. They did not have data on the physical properties of the Bowland Shale, so those from the Marcellus Shale were used as a proxy. The effect of faults was then crudely simulated in some models by extending six of the fractures upwards into the Sherwood Sandstone Group (SSG) aquifer, to simulate natural faults. They found that the SSG aquifer could become contaminated on the order of 100 years under certain conditions. Because they put in a sideways-directed head to simulate regional flow from the Bowland Fells west to the Irish Sea, most of the flow was diverted sideways within the Collyhurst Sandstone, which is a high permeability layer between the SSG (above) and the Bowland Shale at depth.

3.1.10 The Cai and Ofterdinger study is flawed as a fault study, principally because the representation of major faults by vertical fractures up to 1 mm in width is unrealistic as a model for major pre-existing faults. Their critique of Myers (2012) reveals this misunderstanding. However, the study may be applicable for the unlikely case of induced fractures propagating a long distance upwards.

3.1.11 Reagan et al. (2015)[8] studied two generic failure scenarios: “(1) communication between the reservoir and aquifer via a connecting fracture or fault and (2) communication via a deteriorated, pre-existing nearby well.” The results are similar in both cases, except that migration via the faulty well is faster. Overburden thicknesses (between the shale and the aquifer) were set at either 200 m or 800 m. Note that the latter depth corresponds approximately to the depth of the target shale at Ellesmere Port.

3.1.12 Their conclusions are principally that gas transport is aided by high permeability of the connecting pathway (fault or well) and by its overall volume – an unremarkable result. They also conclude that production of gas from the shale formation will mitigate upward migration – again, not a surprising result. The breakthrough times, read from diagrams, are of the order of 0.02 to 200 days (the former figure is under one hour).

3.1.13 Birdsell et al. (2015)[9] reviewed all the earlier work summarised above, and described their own generic study, applicable to a conducting conduit which could be either a fault or a wellbore. Their results show that only a few percent, or even a fraction of a percent, of tracer reaches the aquifer. However, their permeable pathway geometry to simulate a fault is unrealistically small.

3.1.14 Wilson et al. (2018)[10] have recently published a modelling study of possible aquifer contamination. They use a 2D model based on a cross-section through the Preston New Road site currently being appraised by Cuadrilla. This might appear to be relevant to Ellesmere Port, but the geometry of the model excludes the passage of produced water (naturally occurring water collected as a by-product of the stimulation, usually highly saline) passing up faults from the Bowland Shale frack zone. But the faults that are present in the model, being away from the direct path upwards from the shale to the near-surface, only have a secondary influence in acting as regional fluid conduits or barriers. Therefore I have omitted the research from Table 1 above.

3.2 Hydrocarbon seepages

3.2.1 Two American oil industry professional societies, the American Association of Petroleum Geologists and the Society of Exploration Geophysicists, have recently published a memoir on hydrocarbon seepage (Aminzadeh, F. et al. 2013 – extract provided at Appendix 1). Such seepages have been known since antiquity, and now provide an indication of an active petroleum production and migration system for the oil explorer. But their relevance here is that successful imaging of such underground chimneys, where upward migration of fluid is localised, or of more diffused gas clouds above a hydrocarbon accumulation, can provide an indication of whether or not particular faults are acting as seals, and/or whether it is the overburden itself is acting as an efficient seal. Three-dimensional seismic reflection data are invariably required as a basis for the specialised imaging to determine this situation.

3.2.2 Selley (1992)[11] has compiled an inventory of oil and gas seepages in the UK. Two of these seeps are in Lancashire. Oil leaked from the Carboniferous through Triassic sediments at Formby (grid ref. SD 310068), and led to the discovery of the eponymous oil field in 1939. A gas seep from Coal Measures occurred at Wigan (grid ref. SO 5805).

3.3 Empirical evidence of faults as conduits

3.3.1 The relevance of such evidence to the Appeal is that the geology around the well is pervaded by many faults, even though these cannot be mapped in detail in the subsurface.

3.3.2 It is widely known and accepted that faults may act as barriers to flow, or else act as conduits for flow. The same fault may even act as a barrier at one epoch and as a conduit later on, or vice-versa. Important classes of mineral deposits have been laid down along fault lines. In the conventional hydrocarbon industry faults are carefully considered, both as barriers and as conduits. As pointed out above, sophisticated 3D seismic methods can now even image the upward flow of gas through permeable geology and/or upwards along fault zones.

3.3.3 Very little evidence of faults acting as conduits for contamination from unconventional shale gas or oil production is available. This is partly because faulting in the principal US shale basins is very rare (unlike the UK shale basins), but also because many of the (admittedly rare) cases in which shale gas production has resulted in damage to homeowner wells have been settled out of court, with non-disclosure clauses signed as part of the settlement. Many more cases of contamination, however, are related to faulty wells rather than geological faults. However it must be stressed that the density of faulting in the UK shale basins is typically 500 times greater than in the USA, which is what provides the specific concern with unconventional hydrocarbons in the UK.

3.3.4 I have carried out a systematic investigation of contamination incidents in the USA, in which faults and not faulty wellbores may have provided the pathway, and have discovered around a dozen such case histories. The timescale in all the examples I have found is that contamination occurs in a matter of weeks or months. This is at the low end of the various computer modelling estimates discussed in section 3.1.

3.4 Wellbore failure as a conduit for contamination

3.4.1 A commentary by Jackson (2014)[12] describes the problem:
“In a technical sense, “well integrity” refers to the zonal isolation of liquids and gases from the target formation or from intermediate layers through which the well passes. In a practical sense, it means that a well doesn’t leak.”

3.4.2 Fairly early on in the history of shale gas exploitation in the USA, Osborn et al. (2011)[13] were able to correlate statistically the concentration of methane in drinking water wells in NE Pennsylvania to nearness to gas wells. The correlation was corroborated by the increased concentration of thermogenic methane and ethane, the latter being a gas which does not occur biogenetically (Jackson et al. 2013)[14].

3.4.3 Well integrity, however, is not a problem restricted to unconventional wells. This is because the techniques used to drill and case a well are much the same, whether the well is conventional or unconventional. It is just that the problem has come to the fore with the advent of unconventional gas drilling.

3.4.4 The Northern Territory of Australia conducted a scientific inquiry into unconventional hydrocarbon extraction. The results were published in 2017. One of its reports comprises a detailed worldwide desktop review of shale well integrity (Huddleston-Holmes et al. 2017)[15]; that is, it was a review of existing research, and not the reporting of original new research. In summary, the authors consider that rates of well failure are most likely to be due to:
“migration of methane gas up the outside of the well, caused by a loss of integrity of the bond between cement and casing or cement and formation. The rates of gas leakage on a per well basis are likely to be small …”

3.4.5 Although otherwise comprehensive, the Northern Territory shale gas well integrity review did not discuss the risk of wells failing due to seismic activity. Such risk has to be taken into account in NW England, where the Earth’s crust appears to be stressed compressively in a N-S direction, and is near to failure. This has been demonstrated at the Preese Hall-1 well in the Fylde, drilled in 2011 by Cuadrilla. Fracking in this well triggered nearly fifty events, of which two events, of magnitudes 1.6 and 2.3, caused the wellbore to be damaged (ovalised) by slip on the fault, such that the well had to be abandoned. Subsequently, abnormal pressure measurements at the wellhead demonstrated cement bond failure of the well. The problem has allegedly been rectified, and the well has been plugged and abandoned. No further monitoring is taking place to validate the success of this plugging and to confirm the lack of upward contamination over an extended time.

3.4.6 The region around Ellesmere Port is not known for unusual or abnormal natural seismic activity. According to the BGS online database there was a magnitude 2.2 event in the Mersey in 1993, a magnitude 1.6, just 200 m from the well in 1992, a magnitude 2.4 at Ashton in 2001, a magnitude 2.2, 4 km from the well in 2013, and a magnitude 1.1, 2km from the well in 2017 (database from 1956 to date).

3.4.7 In the Environmental Risk Assessment (CD1.9) the Appellant states that the Mersey Estuary is a Special Protection Area and a Site of Special Scientific Interest (page 7), and gives an accurate grid reference (SJ 39793 78266 ) to the edge of the features. However, the assessment misleadingly gives the distance from this point to the well as 2.4 km, when in fact the correct distance is about 300 m. This is illustrated in Figure 3.1.

fig 3.1
Figure 3.1. Mersey Special Protection Area (blue dotted area) 300 m NE of Ellesmere Port-1. Grid is at a 1 km spacing. (data from //magic.defra.gov.uk/MagicMap.aspx)

3.4.8 The map is relevant to the potential for contamination, because it shows that potential methane contamination from the well and its immediate surrounds could affect the protected area. Research results from NE Pennsylvania, a region heavily exploited for unconventional gas, show that contamination by gases (methane and ethane) of deep (thermogenic) origin, plus volatile organic compounds, can occur within 1-2 km distance from a wellhead, by passage outwards through the shallow subsurface after having migrated up the faulty well.

3.5 Hydrogen sulphide

3.5.1 The Bowland Shale Group (aka the Holywell Shale) is a source rock for several oil and gas fields in the eastern Irish Sea and Lancashire (Haig et al. 1997)[16]. Figure 3.2 shows a map of these fields.

fig 3.2
Figure 3.2. Oil and gas fields of the eastern Irish Sea and Lancashire.

3.5.2 The gas in the following fields is sour, i.e. it contains a high hydrogen sulphide (H2S) content: Darwen, Hamilton, Calder, Lennox and Asland. These fields within the map area of Figure 3.2 are shown coloured brown. Douglas and Lennox, coloured blue, have sour oil. The onshore Formby oilfield lies within the cluster of wells in Lancashire shown in Figure 3.2. Haig et al. state that the oil in the Douglas, Lennox and Formby oilfields has been generated from the same Bowland Shale Group source rock. It is therefore possible, or even probable, that any gas produced from the same source rock at EP-1 will also have a high H2S content.

3.5.3 Brook et al. (2012)[17] discussed the risk of high-sour exploration wells. They state, in a paper read at a meeting of the Society of Petroleum Engineers, the leading society for that profession:
“Approximately one-third of the world’s gas fields contain ‘sour gas’, contaminated by hydrogen sulphide, (H2S) …. H2S is one of the most deadly hazards in the oil and gas industry, making such fields more difficult to develop, especially when they are located in the vicinity of populated areas. Conventional approaches to wellsite layout generally use conservative consequence-based criteria, which mean a large number of high-sour wells cannot be drilled due to the potential impact on population centres from an uncontrolled release of sour gas.”

3.5.4 They concluded that the risk was low; however, they modelled a desert environment with the risk to one habitation at 5 km distance being estimated. This conclusion is evidently not applicable to Ellesmere Port, where, if we were to follow their risk criteria, the well would not be developed to production.

3.5.5 The Appellant has not considered the risk of production of H2S as a by-product of its proposed flow testing. The health risk of H2S is both that it is deadly poisonous to humans and animals. It smells of rotten eggs at low concentrations, but at higher, lethal concentrations it becomes odourless. It is also a fire hazard. In addition, leaks of H2S to the biosphere may occur via the pathways discussed above.

3.6 Discussion of potential conduits for contamination

3.6.1 The modelling studies to date on flow up faults all confirm that fluid from the stimulated shale may use faults as a migration route; however the estimated transit times to reach the near-surface vary by two orders of magnitude, from less than ten years to a thousand years in the case of liquid. Gas transit is very rapid, of the order of one hour to hundreds of days. The empirical case histories, on the other hand, suggest that the pollution event can occur very rapidly, on the order of weeks or months, and not decades as suggested by the modelling studies.

3.6.2 For a fault to be a potential pathway it does not have to have a large displacement; it merely has to connect the hydrocarbon-bearing strata to the near-surface groundwater resources at risk; but a large fault will have a wider damage zone than a small one, so will be a greater conductor. The Bowland numerical study, although geologically unrealistic, demonstrates that a 1 mm wide fracture can transmit fluid upwards in the order of 100 years (Cai and Ofterdinger, 2014).

3.6.3 Stimulation of the shale, whether by fracking or by acidisation, may trigger earthquakes. This is because both techniques reduce the natural strength of the rock which, in a tectonically stressed geology, as is the case in the UK, then permits this natural tectonic (crustal) stress to released, causing a seismic event. In addition, long-term extraction of oil or gas may lead to seismic activity. The concern is not that such activity will ever be likely to cause damage to structures at the surface, but they might disrupt wellbores and watercourses leading to contamination. Such disruption is not always likely to lead to serious damage to well casing, as occurred at Preese Hall-1, but on the other hand is more likely to lead to long-term degradation of the cement bonding around the casing, leading to leaks at the surface.

4 Structure and stratigraphy at the wellsite

4.1 Introduction

4.1.1 This section takes into account the application documents provided by the Appellant, (CD1.5, CD1.9, CD2.4) and the further information requested by FFEP&U, and belatedly provided by the Appellant on 23 November 2018, viz. the EP-1 well logs and final report, and the IG-14 set of seismic reflection data. The latter are currently not in the public domain, but will become public on 1 January 2019. It also incorporates the change in the Appellant’s belated view of the target geology, from being conventional to unconventional in nature, discussed in section 2.1 above.

4.2 The IGas cross-section

4.2.1 The Appellant states in the Site Condition Report (CD 1.9) 5.2.1:
“Geologically, the Pentre Chert sits stratigraphically above the Teilia Shale and below the Holywell Shale Formation. Following the analysis of data acquired during the 2014 drilling operation it is likely that the Pentre Chert formation thins to the north. A thinning to the north is also identifiable within the 3D seismic data. This geometry allows for an accumulation of conventionally trapped hydrocarbons within the Pentre Chert in a stratigraphic pinch out.
The reservoir comprises of [sic] interbedded chert, silicious [sic] shales and occasional sands deposited in an offshore, deep water environment towards the bottom of a siliclastic ramp.”

4.2.2 The statement is accompanied by a cross-section, reproduced below as Fig 4.1.

4.2.3 This cross-section is of very limited use, because:
• It has neither horizontal nor vertical scales,
• The upper part is evidently truncated, such that most of the Sherwood Sandstone Group is missing,
• There is no location map to accompany it, and
• Faulting is not shown in the area of interest.

fig 4.1
Fig 4.1. North-south cross-section allegedly through EP-1 provided by the Appellant.
The unexplained letters within black circles may denote: C – caprock, R – reservoir, and S – source.

4.2.4 The cross-section does not correspond to any approximately N-S line through EP-1, whether the vertical scale is measured in depth or in two-way reflection time (TWT). I demonstrate this with a couple of thickness measurements. In the diagram the Coal Measures are shown as 1.34 times as thick as the sequence from the top of the Teilia Shale to basement. But at EP-1 the upper of these two intervals is 274 m thick and the lower is just 72 m thick – a ratio of 3.8. The ratio is even more extreme if we assume that the vertical scale is in TWT, in which case it becomes 5.5. Therefore the geology at the well labelled EP-1 evidently does not correspond to the data from the well. It would appear that the cross-section has been taken either from IG-14-01 or from a vertical display from within the North Dee-15 3D seismic survey (still confidential), situated about 8 km to the east of EP-1. In fact, the well labelled EP-1 in the diagram may correspond more closely to Ince Marshes-1, although this latter well did not penetrate to basement.

4.2.5 This conclusion is supported by the velocity survey, discussed next.

4.3 Velocity survey (supplied by IGas 23/11/18)

4.3.1 Baker Hughes carried out a vertical seismic profile (VSP) at EP-1. This is a survey designed to enable a tie-in of reflectors on seismic data to the geology in the well, and to calibrate the sonic log to make an accurate integrated travel time (ITT) log. The ITT, multiplied by two, then yields the two-way travel times (TWTs) of seismic reflection profiles.

4.3.2 Baker Hughes states:
“A 2D surface seismic line (IG-14-01) was provided by the client. The line is on average more than 6000m away from the well head. A corridor tie is therefore impossible but similarities between the events seen on the corridor stack and on the surface seismic were found, … Different time shifts were applied to focus on the Warwickshire, Coal and Basement. The corridor stack was tied with the IG-14-01 line at 346448E and 374931N.”

4.3.3 It is significant that the Appellant provided a portion of seismic line situated between 7 and 8 km east of the well for the attempt at a corridor tie (figure 4.2). It would have been appropriate for the Appellant to have provided a segment of one of the seismic lines IG-14-02, -03 or -06 for this task, since these lines lie within 1 km of the well.

fig 4.2
Figure 4.2. Shotpoint map of the IG-14 2D survey lines. The portion of IG-14-01 used for the corridor tie to EP-1 (red star) is highlighted in blue.

4.3.4 The significance of this highly unusual tie, from the well to a seismic line 7-8 km distant, instead of the normal procedure which is to tie in to a seismic line which runs through, or is adjacent to the well, is that it demonstrates that no reliable tie of the well ‘tops’ (the geological information from the well) to the seismic data around the well has, or can be, made. This in turn demonstrates that the Appellant does not have an adequate understanding of the geology around the well. This lack of understanding will not be ameliorated by the proposed works.

4.4 Quality of seismic data in the vicinity of the well (supplied by IGas 23/11/18)

4.4.1 The nearest seismic lines to the well is IG-14-03. The well tie to the final migrated version of the line is shown in Figure 4.3.

fig 4.3
Figure 4.3. Migrated version of part of seismic line IG-14-03. The inset shows the well tops in more detail.

4.4.2 The well, which is essentially vertical, lies 270 m NE of the seismic line. The seismic data quality is extremely poor; none of the well tops shown in the inset to Figure 4.3 can be correlated with any reflectors. Even the concave-upward strong reflector at 700 ms to the east of the wellbore may be a migration ‘smile’, i.e. an artefact, possibly from an out-of-plane reflection.

4.4.3 The other seismic lines in the vicinity of the well, including the 1980s vintage data not shown in Figure 4.2, suffer from the same problem of poor quality.

4.5 The nature of the Appellant’s proposed acidisation

4.5.1 The Applicant claims, concerning the chert to be flow-tested (CD 1.9 5.2.1):
“The zone of interest is from 1,795mMD to 1,849mMD, with the primary interval being between 1,846mMD and 1,849mMD.”

4.5.2 First, the depths quoted here seem to be in error. Figure 4.4 shows an extract from the EP 1 composite log covering the Pentre Chert interval. The interval extends from 1767 m MD (measured depth) to 1864 m MD (figures circled in red), a total interval of 97 m. The detailed zone of interest is the limestone from about 1845.5-1850.5 m MD, about 5 m in thickness.

4.5.3 Second, what level, or type, of acidisation is proposed? I refer to the informative paper on acidisation by McWhirter (EP05). She has summarised how the different levels of acid treatment are defined, in order of increasing severity:
• Acid washing / acid maintenance
• Matrix acidising
• Fracture acidising / acid fracking.

4.5.4 The Applicant claimed originally that 45 m3 of 15% strength hydrochloric acid would be required for the testing:
“To re-establish natural flow in the formation, a dilute acid (most commonly hydrochloric acid) at 15% concentration with water is applied to the near wellbore formation through the perforations. This is a standard oilfield practice and similar to practices used in the rehabilitation of public water supplies and commercial water wells.”

4.5.5 The Appellant has revised this figure upwards in Revision 1 of their document. Its paragraph 7.1.3 states:
“The proposed dilution of hydrochloric acid is 15%, which is circulated across the perforations. The process of washing the perforations will be repeated if necessary. Following the washing of the perforations, HCI is then selectively squeezed into the formation. It is anticipated that collectively a total of 95m3 of HCI will be pumped into the formation. IGas may undertake a maximum of three (3) acid wash and squeeze operations. The volume quantity for each acid treatment will vary and an exact figure cannot be provided at the time of permit application, however for absolute clarity the total volume of acid will not exceed 95m3 for the entire operation.'”

fig 4.4
Figure 4.4. Composite log for EP-1 covering the Pentre Chert interval. Some log tracks have been removed, leaving only the gamma ray (green) and caliper (dashed black) on the left of the lithology column, and the mud gas track on the right. The vertical depth scale is gridded at a 2 m interval.

4.5.6 The acid strength of 15% is double what is required and customary for an acid wash, which would normally be 7%. It therefore appears that the Appellant is actually proposing matrix acidising, which cannot be considered as a form of ‘rehabilitation’. Furthermore the EA (EP20) considers matrix acidising to be a form of well stimulation. This is incompatible with the Appellant’s initial claim that the chert is a conventional reservoir.

4.5.7 The volume requirements for the acid treatment also show that the Appellant intends to carry out matrix acidisation. The California Department of Conservation (CDC) has published a discussion paper (California Department of Conservation 2014)[18] on the threshold volumes for acid stimulation of wells. It states:
“For the purposes of present regulation, the Acid Volume Threshold [AVT} will be a determining factor in distinguishing between well stimulation treatments designed to enhance the permeability of the formation, versus uses of acid for wellbore clean-up, wellbore maintenance, and removal of formation damage.”

4.5.8 The CDC has determined that a radius of 36″ from the wellbore axis is a suitable basis for calculating the AVT. Below that threshold the amount of acid used is considered adequate for an acid wash; above the threshold it is considered as a form of stimulation.

4.5.9 The AVT is calculated as the volume of pore space in the rock out to the radius of 36″, per foot of wellbore length, subtracting the volume of the wellbore itself where the rock has been drilled out. In the absence of data from the Pentre Chert, an appropriate proxy for porosity/permeability data is the Woodford Shale of the USA, as discussed in section 2.2 above. Fishman et al. (2010)[19] give the porosity of their data samples collected from outcrop as 0.59-3.46%, and a permeability of 0.001 to 0.033 μD. The latter figures are clearly in the range of tight formations, as seen in Figure 2.1 above.

4.5.10 To calculate the AVT for the Pentre Chert, using the CDC methodology, let us round off the 36″ radius criterion to 1 m, assume a 6″ wellbore diameter, and generously assume a 5% porosity for the chert. The AVT per metre of well is then 0.14 m3 of acid, which should be of 7% strength and not 15%, if the purpose really is for an acid wash.

4.5.11 The total acid volume requirement for the acid wash of the limestone with the gas shows, 5 m thick, is therefore 0.14 x 5 = 0.7 m3. Even if the entire Pentre Chert interval is to be acid washed, the AVT is 0.14 x 97 = 13 m3. This figure is one-seventh of the Appellant’s alleged requirement of 95 m3.

4.5.12 It is therefore reasonable to conclude that the Appellant is actually seeking to acid stimulate the chert, and not merely to clean up the wellbore with an acid wash.

4.6 Regional faulting

4.6.1 In the region of interest, NW Cheshire, the Bowland Basin, of Carboniferous age, is overlain by Upper Carboniferous Millstone Grit and by Triassic Sherwood Sandstone Group sediments. Post-Triassic normal extensional faulting dominates the structure, but faulting with various senses of displacement – strike-slip, normal and reverse – is present in the shale layers and basement.

4.6.2 Figure 4.4 shows a regional cross-section compiled by the BGS as part of its Bowland Shale study (Andrews 2013)[20]

fig 4.4b
Figure 4.4. Extract from BGS regional cross-section, which passes through Ince Marshes-1.

4.6.3 The cross-section illustrates the large-scale normal faults cutting the section. The portion nearest to EP-1 is marked.

4.7 Faulting in the neighbourhood of the wellsite

4.7.1 The nearly 800 m thickness of Sherwood Sandstone Group (SSG) is an important Primary Aquifer (Many Cheshire villages take their drinking water from it). It has been, and continues to be, used for industrial and agricultural purposes, and for that reason cannot be subject to risk of contamination.

4.7.2 Figure 4.5 shows a compilation of the four BGS solid geology maps around the site.

4.7.3 The various shades of yellow or buff denote the Triassic-age SSG at outcrop (or, rather, subcrop below unconsolidated sediments), on which I have superimposed faults in red from the Edina Digimap digital database. The colours change across map boundaries because the mapping and subsequent map publication has taken place at different epochs.

fig 4.5
Figure 4.5. 1:50,000 solid geology map sheet compilation. The location of EP-1 is denoted by the red star. Red lines are faults from the digital database.

4.7.4 Because of the poor exposure, the recognition of faults by surface mapping is very difficult. The faults shown in Figure 4.5 are probably only a small fraction of those that exist.

4.7.5 I have interpreted faults on the IG-14 set of seismic lines, loaded into the IHS Kingdom v8.8 interpretation program. I have also inspected images of a dozen or so other older 2D released lines in the area, available on the UKOGL website. The quality of the IG-14 lines within 1-2 km of the vicinity of EP-1 is extremely poor, but somewhat better a few kilometres further east. The same problem applies to the other seismic data, for reasons discussed below.

4.7.6 Figure 4.6 shows an unmigrated version of IG-14-03. In view of the almost complete absence of continuous coherent reflectors it is better to use this form of processing to try to identify faults, which I have marked as black lines. The migration smile feature seen in the migrated version (Fig. 4.3) is now seen as a narrow faulted anticlinal feature just east of the wellbore.

fig 4.6
Figure 4.6. Unmigrated version of IG-14-03 with fault interpretation marked as black lines. EP-1 is positioned by a red star. The inset shows the location of the line in red; grid at 1 km interval.

4.7.7 It is apparent that mapping outwards (away from) the well tops is quite impossible

4.7.8 In conclusion, the SSG is pervaded by many near-vertical faults. These, together with the wellbore itself, may act as conduits for contamination and upward migration of stray gas and liquids.

4.8 Shallow water supply boreholes

4.8.1 There are at least a dozen deep boreholes archived by the BGS within a radius of 4 km of EP-1 (Fig. 4.7). The map gives the depths in metres in the box to the right or the left of the maroon borehole symbol. Many of the boreholes, denoted by a black circle, remain confidential. Shallow boreholes (under 10 m depth in blue; 1-30 m depth in green) have not been inspected. The boreholes marked, from 79 m depth or greater, are very unlikely to have been bored for construction purposes.

4.8.2 This analysis does not preclude the likelihood that some of the boreholes less than 30 m deep may also be for water supply. The deepest borehole, Standing Wood, labelled 685 m depth on the map, is about 400 m south of EP-1. It was sunk originally for coal in 1902, then either deepened or re-bored in 1932 for Bowaters Paper Mill. The flow rate was measured at over 300 m3 per hour. It is not known whether this and other boreholes are still in use.

fig 4.7
Figure 4.7. BGS borehole viewer map of the area around EP-1. The circle is 4 km in radius centred on the well. Deep boreholes (red) are defined as more than 30 m deep; depths are labelled in metres on the right or the left of the symbol. Note that black symbols are confidential boreholes; these may include more deep water boreholes.

4.9 Discussion of the geology at the wellsite

4.9.1 The SSG at the wellsite is moderately to poorly reflective, comprising fairly weak short reflections. This could be due to one or more of several explanations:
• Poor seismic acquisition and/or processing techniques.
• Intrinsically unreflective beds, comprising conglomerates.
• Gas chimneys, which tend to break up reflector coherency.
• Pervasive faulting.

4.9.2 The comparison of the stacked (unmigrated) and the final migrated versions of the west end of line IG-14-03 is instructive (Figure 4.8). The tops from EP-1 are marked. Seismic layering is seen in the stacked version within the Triassic; the SSG extends from outcrop down to the top of the Carboniferous, marked by the Warwickshire Group top (abbreviated as WarGroup). If the processing had been better then these reflectors would have been equally good or better in the migrated version.

4.9.3 Many small faults can be interpreted on the stacked version of the line, as has been shown in Figure 4.6. I conclude that the poor reflection continuity is a product of faulting, together with mediocre processing. The alluvial SSG, in addition, does not lend itself to strong continuous reflectors. Gas chimneys can be discounted as a possible cause of the poor reflector quality.
fig 4.8
Figure 4.8. Comparison of stacked and migrated versions of part of seismic line IG-14-03.

4.9.4 The Carboniferous section, however, is poorly imaged in both the stacked and the migrated versions. It follows that the amount of detail shown in the Appellant’s cross-section purporting to be through EP-1 (reproduced in Figure 4.1 above) is unjustified. There is no evidence for any of the continuous layering portrayed.

4.9.5 There are no impermeable capping rocks, such as clays or marls, lying between the shale to be exploited and the SSG. The intervening Coal Measures and Millstone Grit Group are both largely arenaceous (sandy, permeable) formations.

4.9.6 There is no realistic prospect of structural mapping any of the geological layers as seen at the well using the existing data.

4.9.7 The Appellant states (EP23 letter dated 22 November 2018) that “the Pentre Chert encountered at Ellesmere Port is considered to be naturally fractured”. No evidence has been provided to support this assertion. The Weatherford fracture report concerns only two cores from above the Pentre Chert, which Weatherford states to be from within the Bowland Formation and the Sabden Shale, respectively. The two cores are noted on the composite log at 1532-7-1550.6 m and 1645.3-1663.3 m (MD); according to this log both are from the Holywell Shale, which is an obsolete name for the Bowland Shale Formation. I have found no evidence in the well logs and reports supplied to date to support the contention of natural fracturing in the chert.

4.9.8 The geological interpretation problems may have had their origin in the initial purpose of the well. Drilling was completed on 12 December 2014. The processing of the IG-14 seismic dataset was undertaken in November-December 2014. So the results of this new dataset were not available for planning and drilling the well, which must therefore have been designed on the basis of interpretation of the older seismic data from the 1980s.

4.9.9 The well was originally planned to explore for coalbed methane within the Coal Measures. It may have been considered that the seismic data available were adequate for this purpose, which was to drill to a minimum depth of 900 m. The Appellant has argued (CD2.7 letter to Cheshire West and Chester Council dated 10 August 2017) that since the permit was granted for this minimum depth, it automatically had permission to drill to more than twice that depth (total depth was 1950 m MD). But this argument is invalid; the Nexen planning statement of 14 October 2009, which formed the basis of the permit subsequently granted, states (para. 9.3.6):
“During the Appraisal Drilling phase two appraisal boreholes will be drilled to an estimated minimum depth of circa 900METRES (3,000′) Total Vertical Depth (TVD). The borehole will typically decrease in diameter from 500mm (20 inches) at the top section to 152 mm (6 inches) in diameter at the maximum depth in the coal seam.” [underlining added]

4.9.10 This states clearly that the coal seam target is to be the maximum depth, at whatever depth that may be, and deeper than 900 m. The top of the Pennine Coal Measures came in at 867 m TVD, and two coal seams were drilled at around 1,010 and 1,040 m MD. Once the drillbit left the Coal Measures and penetrated the Millstone Grit Group at 1,141 m TVD it was therefore drilling without permission. There is no possibility that coal could have been encountered at 1,141 m or greater depths.

4.9.11 The minimum depth of 900 m quoted above probably comes from a condition in the PEDL award. In my experience of PEDL award conditions (section 1.1.6 above) it is customary for the award of a PEDL, if the drilling of a well has been offered by the applicant, to put a minimum depth on the well. This is to avoid the possibility of a licensee later claiming that it had fulfilled the conditions of the award by subsequently drilling a well to shallow depth. The specified minimum depth corresponds to the target formation described in the licence application. The licence, PEDL184, started on 1 July 2008. It was awarded for coal bed methane extraction. It is likely that the geological rationale for the siting of the well was optimised for coal bed methane exploration, and it is the later switch of emphasis to prospecting for the deeper shales that has led to problems of interpretation. In conclusion, a well to target the shales, and in particular, the Pentre Chert, should probably have been sited elsewhere within the PEDL.

5 Conclusions and recommendations

5.1 Summary of findings

5.1.1 The EP-1 well was permitted to drill only to the maximum depth of a coal seam. The Appellant had no permit to drill deeper than the Pennine Coal Measures.

5.1.2 The Appellant has misleadingly transferred its geological interpretations from 7-8 km further east and presented them as if they were a factual interpretation of its findings at EP-1. The detailed interpretation presented simply cannot be justified, given the poor quality of the existing 2D seismic data around EP-1.

5.1.3 It is possible that the interpretation presented is based on the North Dee-15 3D seismic survey to the east. However, since this survey is still confidential we have no way of scrutinising the veracity or otherwise of the interpretation. Even if it is correct within the 3D survey area itself, there is no justification for transferring such an interpretation some 7-10 km west to the EP-1 locality.

5.1.4 The fact that the Applicant was apparently obliged to provide to its velocity survey contractor a useable sample of 2D seismic data from a distance of 7-8 km away from the well, to try to tie in the wellhead velocity survey to the grid of 2D seismic reflection data, is independent evidence that that seismic data around the well are, in effect, unusable for characterising the Carboniferous shales.

5.1.5 The structural layering of the chert target at EP-1 is therefore quite unknown. The concept of a stratigraphic pinch-out, to provide a trapping mechanism, is therefore also mere speculation. The Applicant has failed to provide the suite of maps that would normally be made and presented to justify its interpretation. I conclude that such maps do not exist.

5.1.6 The Appellant claimed initially that the chert target is a conventional reservoir, but has lately conceded that it is unconventional in nature. Therefore the claimed conventional stratigraphic trapping mechanism to retain the hydrocarbons is no longer valid. It is a tacit admission that the hydrocarbons will require stimulation to extract.

5.1.7 The volume and concentration requirement for hydrochloric acid is clearly within the range of that required for well stimulation – an unconventional treatment – and not for wellbore cleaning.

5.1.8 There is no cap rock to prevent upward migration of contaminating fluids (liquids and gas) from the shale zone to the Principal Aquifer of the Sherwood Sandstone Group (SSG). The aquifer is extensively used for industrial and agricultural ends, and cannot be put at risk of contamination.

5.1.9 The Mersey Estuary Special Protection Area and Site of Special Scientific Interest, both adjacent to the site, may therefore be at risk of harm. The Appellant has misled the Inspector over the distance to these areas, which are just 300 m distant, and not the claimed distance of 2.4 km.

5.1.10 The SSG is pervaded by many near-vertical faults. These, together with the wellbore itself, may act as conduits for contamination and upward migration of stray fluids.

5.1.11 The stray gas component of the fluids may include not just hydrocarbons, but highly poisonous hydrogen sulphide, because the source rock being investigated is known to contain high levels of this gas.

5.1.12 Given the lack of free flow of gas within the chert at EP-1, it is a priori unlikely that the chert is a conventional high-permeability reservoir. If gas had flowed freely the well could have been classed as a conventional discovery well.

5.1.13 The reasonable presumption is that the Pentre Chert is a tight rock (very low permeability) which, by definition, will require unconventional methods to be employed to stimulate the flow, such as acidisation or fracking. The Appellant is actually planning to apply matrix acidisation, and not merely the claimed acid wash.

5.2 Conclusion and recommendations

5.2.1 In conclusion, and bearing in mind the misleading and incomplete nature of many aspects of the Appellant’s case for flow testing the chert, there is little confidence that the proposed development will not lead to contamination of the aquifer or escape of gas. Therefore the appeal should be dismissed.

5.2.2 However, if the Inspector is minded to allow the appeal, it should be granted on condition that the acidisation be restricted to what the Appellant claims to wish to achieve by this method; that is, the use of a volume and strength of acid appropriate to wellbore cleaning, applied at a pressure just above formation pressure.

APPENDIX 1 – Hydrocarbon Seepage: From Source to Surface

Aminzadeh, F. et al. (eds.) 2013. Hydrocarbon Seepage: From Source to Surface. Society of Exploration Geophysicists and American Association of Petroleum Geologists, [selected pages reproduced below]

app pic 1

app pic 2

app pic 3

app pic 4

app pic 5

References

[1]Watney, W.L., Guy, W.J. and Byrnes, A.P. 2001. Characterization of the Mississippian chat in southcentral Kansas. Bulletin of the American Association of Petroleum Geologists., 85(1),  85–113.

[2]Myers, T. 2012. Potential contaminant pathways from hydraulically fractured shales to aquifers. Ground Water 50, no. 6: 872–882. DOI: 10.1111/j.1745-6584.2012.00933.x.

[3]Borchardt, D., Ewen, C., and Hammerbacher, R. 2012. Abschätzung der Auswirkungen von Fracking-Maßnahmen auf das oberflächennahe Grundwasser- Generische Charakterisierung und Modellierung (Assessment of the impact of fracking operations on the near-surface groundwater – Generic characterisation and modelling), http://www.dialog-erdgasundfrac.de/

[4]Ewen, C., Borchardt, D., Richter, S. and Hammerbacher, R. 2012.  Hydrofracking Risk Assessment Executive Summary [English summary version of Borchardt et al.,  http://www.dialog-erdgasundfrac.de ]

[5]Kissinger, A., Helmig. R., Ebigbo, A.,Class, H., Lange, T, Sauter, M., Heitfeld, M., Klünker, J., and Jahnke, W. 2013. Hydraulic fracturing in unconventional gas reservoirs: risks in the geological system, part 2, Environ. Earth Sci., 70(8), 3855– 3873, doi:10.1007/s12665-013-2578-6, 2013

[6]Gassiat, C., Gleeson, T., Lefebvre, R., and McKenzie, J. 2013. Hydraulic fracturing in faulted sedimentary basins: Numerical simulation of potential long term contamination of shallow aquifers, Water Resour. Res., 49(12), 8310-8327, doi:10.1002/2013WR014287.

[7]Cai, Z. and Ofterdinger, U. 2014. Numerical assessment of potential impacts of hydraulically fractured Bowland Shale on overlying aquifers. Water Resources Research, 50, 6236–6259, doi:10.1002/2013WR014943, 2014.

[8]Reagan, M. T., Moridis, G. J., Keen, N. D., and Johnson, J. N. 2015. Numerical simulation of the environmental impact of hydraulic fracturing of tight/ shale gas reservoirs on near-surface groundwater: Background, base cases, shallow reservoirs, short-term gas, and water transport, Water Resour. Res., 51, 2543–2573, doi:10.1002/2014WR016086

[9]Birdsell, D. T., Rajaram, H., Dempsey, D. and Viswanathan, H. S. 2015.  Hydraulic fracturing fluid migration in the subsurface: A review and expanded modeling results, Water Resources Research, 51(9), 7159–7188.

[10]Wilson, M. P., Worrall, F., Davies, R. J., & Hart, A. 2017. Shallow aquifer vulnerability from subsurface fluid injection at a proposed shale gas hydraulic fracturing site. Water Resources Research, 53, 9922–9940. https://doi.org/10.1002/2017WR021234

[11] Selley, R.C. 1992. Petroleum seepages and impregnations in Great Britain. Mar. Petrol. Geol. 9, 226-244.

[12] Jackson, R. B. 2014. The integrity of oil and gas wells. Proc. Nat. Acad Sci. https://doi.org/10.1073/pnas.1410786111

[13]Osborn, S. G., Vengosh, A., Warner, N. R. and Jackson, R. B. 2011. Methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing, Proc. Natl. Acad. Sci., 108(20), 8172–8176, doi:10.1073/pnas.1100682108.

[14]Jackson, R.B. et al. 2013. Increased stray gas abundance in a subset of drinking water wells near Marcellus shale gas extraction. Proc. Natl. Acad. Sci. http://www.pnas.org/cgi/doi/10.1073/pnas.1221635110

[15]Huddlestone-Holmes, C.R., Wu, B., Kear, J, and Pandurangan, R. 2017. Report into the shale gas well life cycle and well integrity. EP179028. CSIRO, Australia.

[16]Haig, D. B., Pickering, S.C. and Probert, R. 1997. The Lennox oil and gas Field. From Meadows, N. S., Trueblood, S. E, Hardman, M. & Cowan, G. (eds), 1997, Petroleum Geology of the Irish Sea and Adjacent Areas, Geological Society Special Publication No. 124, pp. 417-436.

[17]Brook, G. and Bates, M. 2012. Development of a Risk-Based Approach for High-Sour Exploration Wells. Society of Petroleum Engineers, SPE Middle East Health, Safety, Security, and Environment Conference and Exhibition, Abu Dhabi, UAE, 2–4 April 2012. SPE 154533.

[18]California Department of Conservation 2014. Discussion of calculated acid volume threshold. SB4 well stimulation treatment regulations.  https://bit.ly/2zz95bP

[19]Fishman, N. et al. 2010.  From radiolarian ooze to reservoir rocks: microporosity in chert beds in the Upper Devonian Lower Mississippian Woodford Shale, Oklahoma. Am. Assoc. Petrol. Geol. Search and Discovery Article #10268 https://bit.ly/2Q5ZK5C

[20]Andrews, I. J. 2013. The Carboniferous Bowland Shale gas study: geology and resource estimation, British Geological Survey for Department of Energy and Climate Change, London, UK, https://www.gov.uk/government/publications/bowland-shale-gas-study.