Scholarly article on topic 'Life cycle environmental impacts of UK shale gas'

Life cycle environmental impacts of UK shale gas Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Applied Energy
Keywords
{"Shale gas" / "Hydraulic fracturing" / Fracking / "Life cycle assessment" / "Climate change" / "Environmental impacts"}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Laurence Stamford, Adisa Azapagic

Abstract Exploitation of shale gas in the UK is at a very early stage, but with the latest estimates suggesting potential resources of 3.8×1013 cubic metres – enough to supply the UK for next 470years – it is viewed by many as an exciting economic prospect. However, its environmental impacts are currently unknown. This is the focus of this paper which estimates for the first time the life cycle impacts of UK shale gas, assuming its use for electricity generation. Shale gas is compared to fossil-fuel alternatives (conventional gas and coal) and low-carbon options (nuclear, offshore wind and solar photovoltaics). The results suggest that the impacts range widely, depending on the assumptions. For example, the global warming potential (GWP100) of electricity from shale gas ranges from 412 to 1102g CO2-eq./kWh with a central estimate of 462g. The central estimates suggest that shale gas is comparable or superior to conventional gas and low-carbon technologies for depletion of abiotic resources, eutrophication, and freshwater, marine and human toxicities. Conversely, it has a higher potential for creation of photochemical oxidants (smog) and terrestrial toxicity than any other option considered. For acidification, shale gas is a better option than coal power but an order of magnitude worse than the other options. The impact on ozone layer depletion is within the range found for conventional gas, but nuclear and wind power are better options still. The results of this research highlight the need for tight regulation and further analysis once typical UK values of key parameters for shale gas are established, including its composition, recovery per well, fugitive emissions and disposal of drilling waste.

Academic research paper on topic "Life cycle environmental impacts of UK shale gas"

Contents lists available at ScienceDirect

Applied Energy

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

Life cycle environmental impacts of UK shale gas

Laurence Stamford, Adisa Azapagic *

School of Chemical Engineering and Analytical Science, Room CIS, The Mill, Sackville Street, The University of Manchester, Manchester M13 9PL, UK

HIGHLIGHTS

• First full life cycle assessment of shale gas used for electricity generation.

• Comparison with coal, conventional and liquefied gas, nuclear, wind and solar PV.

• Shale gas worse than coal for three impacts and better than renewables for four.

• It has higher photochemical smog and terrestrial toxicity than the other options.

• Shale gas a sound environmental option only if accompanied by stringent regulation.

ABSTRACT

Exploitation of shale gas in the UK is at a very early stage, but with the latest estimates suggesting potential resources of 3.8 x 1013 cubic metres - enough to supply the UK for next 470 years - it is viewed by many as an exciting economic prospect. However, its environmental impacts are currently unknown. This is the focus of this paper which estimates for the first time the life cycle impacts of UK shale gas, assuming its use for electricity generation. Shale gas is compared to fossil-fuel alternatives (conventional gas and coal) and low-carbon options (nuclear, offshore wind and solar photovoltaics). The results suggest that the impacts range widely, depending on the assumptions. For example, the global warming potential (GWP100) of electricity from shale gas ranges from 412 to 1102 g CO2-eq./kWh with a central estimate of 462 g. The central estimates suggest that shale gas is comparable or superior to conventional gas and low-carbon technologies for depletion of abiotic resources, eutrophication, and freshwater, marine and human toxicities. Conversely, it has a higher potential for creation of photochemical oxidants (smog) and terrestrial toxicity than any other option considered. For acidification, shale gas is a better option than coal power but an order of magnitude worse than the other options. The impact on ozone layer depletion is within the range found for conventional gas, but nuclear and wind power are better options still. The results of this research highlight the need for tight regulation and further analysis once typical UK values of key parameters for shale gas are established, including its composition, recovery per well, fugitive emissions and disposal of drilling waste.

© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CCBY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/3XI/).

CrossMark

ARTICLE INFO

Article history:

Received 27 November 2013 Received in revised form 14 August 2014 Accepted 17 August 2014 Available online 6 September 2014

Keywords: Shale gas

Hydraulic fracturing Fracking

Life cycle assessment Climate change Environmental impacts

1. Introduction

The discovery and exploration of shale gas in the UK is a very recent phenomenon about which much remains unknown - while exploration is occurring, commercial extraction has not yet begun. In 2010, the official reserve estimate stood at 150 billion cubic metres (Gm3) [1] - or 5.3 trillion cubic feet (Tft3). Only a year later, in 2011, Cuadrilla Resources claimed to have discovered 5600 Gm3 (200 Tft3) in Lancashire alone [2]. In 2013, IGas declared that up to 4800 Gm3 (170 Tft3) were available in their licensed area in the North West [3]. By July 2013, it had been estimated that

* Corresponding author. E-mail address: adisa.azapagic@manchester.ac.uk (A. Azapagic).

37,600 Gm3 (1329 Tft3) of gas-in-place existed in central Britain, although not all of this will be recoverable [4]. By comparison, the total gas input to the UK transmission system in 2011 was 79.9 Gm3 [5], meaning that the most recent estimate equates to about 470 years' worth of shale gas supply.

Regardless of the obvious uncertainties in the recoverable reserve size, shale gas could clearly be a 'game-changing' resource that could transform the UK energy market and contribute significantly to the national security of supply. However, while the economic potential is obvious - for example, in the US, shale gas is estimated to have a value of nearly $34bn in 2011 [6] - its environmental and social implications are currently unknown, making it a controversial issue. Some of the reasons for this include earthquakes caused due to the extraction of shale gas by hydraulic

http://dx.doi.org/10.1016/j.apenergy.2014.08.063 0306-2619/® 2014 The Authors. Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

fracturing or 'fracking' and leakage of fracking chemicals and gas (methane) into the water table, the latter of which has been reported in several US states (see, for example, Pennsylvania DEP [7]). In the UK, low-intensity earthquakes (measuring 2.3 and 1.5 on the Richter scale) were observed in April 2011 due to fracking in North West England which led the Government to suspend shale gas extraction nationally from May 2011 to December 2012. New regulations were subsequently imposed and fracking in the UK has since continued [8].

In addition to the above issues, there is an ongoing debate on whether, and how, shale gas might fit into a world attempting to reduce fossil fuel usage and prevent dangerous climate change (see, for example, Friends of the Earth [9] and Wood et al. [10]). Some scenario analysis has already been conducted by the 1EA in an attempt to address these issues, with the conclusion that unconventional gas exploitation should have little impact on climate change [11]. However, this is far from certain and is highly dependent on what shale gas displaces. 1f it is consumed mainly in countries with emissions restrictions (carbon and otherwise) -for instance if UK shale gas is consumed in the UK - it will likely displace imported gas and coal leading to a net reduction in direct emissions. This has also been the case in the US, where the price signal created by shale exploitation has discouraged the burning of coal for electricity, ultimately allowing US energy sector emissions to fall from a 2007 peak of 6023 Mt CO2 to a two-decade low of 5290 Mt CO2 in 2012 [12]. However, availability of cheap shale gas is also likely to depress investment in low-carbon technologies such as nuclear and renewables. Moreover, even if shale gas displaces coal, a reduction in coal demand in countries with shale gas can lead to reduction in global coal prices, potentially increasing coal consumption in other countries. For example, it has been argued that increased coal exports from the USA, driven by shale gas availability, are the cause of higher coal consumption in Europe [13]. In the UK, coal's deflated price has made it a more profitable fuel for energy companies than gas [14] so that its consumption grew by 33% in 2011/2012, increasing its contribution to the electricity mix from 29.5% to 39.4%, the highest since 1996 [5]. However, other factors may have contributed, such as the sharp rise in European gas prices and the approach of the deadline for the implementation of the EU Large Combustion Plant Directive. Ultimately, the impact of shale gas will depend mainly on the cost relationships of different fuels and technologies in a complex, dynamic, semi-global market. Thus, its true effect is extremely difficult to predict and a debate should continue to explore the ethics and future market perturbations of exploiting this new resource; however, this is beyond the scope of this paper which focuses on environmental impacts of shale gas over its life cycle.

Limited literature is available on its life cycle impacts owing to the relative immaturity of shale gas fracking. Most such studies originate in the US where shale gas extraction is much greater than in the UK: shale gas represented 30% of total US natural gas production in 2011, up from 8% in 2007 [6]. Most studies estimate the greenhouse gas (GHG) emissions and related global warming potential (GWP) of shale gas extraction [15-24]. Of these studies, only that by MacKay and Stone [23] is UK-oriented and still relies almost entirely on data from the other US-based papers cited above. Outside of the US and UK, some LCA work has also been conducted for shale gas in China [25], again considering GHG emissions but only for the production stage. As far as the authors are aware, there are no studies considering life cycle environmental impacts of shale gas other than GWP, with the exception of life cycle water consumption [24,26] and some qualitative discussion of potential non-GWP impacts to air, water and land [27].

Among the available estimates, the most controversial (and often cited) is that of Howarth et al. [17] which asserts that the

GWP of shale gas is greater than that of coal. Although the authors calculated the GWP of the extracted gas at 84-224 g CO2-eq./MJ, they focused on the top end of the estimated range to arrive at their conclusion. Primarily, this high estimate is due to the use of a high GWP for methane over a 20-year time horizon (GWP20) of 105 kg CO2-eq./kg CH4, based on work by Shindell et al. [28] rather than the more-widely accepted value of 75 kg CO2-eq./kg as estimated by the 1PCC [29]. By comparison, the GWP over 100 years (GWP100), also estimated by the 1PCC and used more commonly, is 25 kg CO2-eq./kg CH4. Furthermore, the authors express the results per MJ of energy contained within the fuel as opposed to kWh of electricity generated, thereby ignoring the higher average efficiency of gas compared to coal-fired power plants. Several other issues also contributed to the high estimates, such as overestima-tion of fugitive methane emissions and the assumption of no flaring to reduce the GWP by converting methane into carbon dioxide, despite flaring being common and/or legally required in the US. Readers are directed to several works by other authors for more detailed criticism of the study by Howarth et al. in regard to the above points [16,18-20,22,30].

Estimates in the literature have tended to express the GWP of shale gas either per unit of energy contained in the fuel (CO2-eq./ MJ) and/or per unit of electricity generated by a power plant (CO2-eq./kWh). These estimates are shown in Fig. 1a and b, respectively, with the bar representing the central estimate and the error bars indicating the range of estimated values. 1n the case of expressing GWP per MJ of fuel, the major differences between estimates are due to assumptions regarding fugitive methane emissions during shale gas extraction and due to different estimated ultimate recovery (EUR) values, representing the total output of a shale gas well over its lifetime. MacKay and Stone [23] and Logan et al. [19] in particular cite this as the major uncertainty in shale gas GWP estimates. Of the studies shown in Fig. 1a, the average GWP100 per unit of energy contained in the fuel is estimated at 72 g CO2-eq./MJ with a range of 56-114 g. Note that these values are skewed by the Howarth et al. result of 84-114 g CO2-eq./MJ: the other five authors have an average central estimate of 66 g CO2-eq./MJ and the results are all within 15% of each other.

A major differentiator between the estimates in Fig. 1b (expressed per unit of electricity generated) is the assumed efficiency of the power plant. Moreover, it is not always possible to determine the exact efficiency assumed as it was not specified whether the higher heating value (HHV) or lower heating value (LHV) of natural gas was used. Logan et al. [19] specify an HHV efficiency of 51% (about 56.6% on an LHV basis). Stephenson et al. [21] quote a lower efficiency of 43% on an HHV basis, or 47.6% for LHV. Other assumed efficiencies in the literature range from 39% to 55% but it is not clear whether these are based on HHV or LHV.

Of the studies shown in Fig. 1b, the average GWP100 per unit of electricity from shale gas is 504 g CO2-eq./kWh with a range of 416730 g. As mentioned previously, Howarth et al. [17] do not provide an estimate for electricity over a 100-year timeframe, using instead only the GWP20 factors. Their maximum value of 2878 g CO2-eq./ kWh is based on a very low-efficiency gas power plant (28%), much lower than the efficiency of the predominant technology, i.e. combined cycle gas turbines (CCGTs). For example, the average efficiency of CCGTs, which account for >90% of gas-fired installed capacity in the UK, is 48.5% for HHV or 53.8% for LHV [5], with the efficiency of new plants approaching 54% HHV (60% LHV).

A focus of previous literature has been comparison of shale gas to conventional natural gas. With the exception of Howarth et al.

[17] all estimates indicate that electricity from shale gas has a similar GWP to that of conventional gas. For example, Hultman et al.

[18] conclude that a shale gas-fired CCGT has 11% higher GWP than conventional gas; Jiang et al. [15] estimate this difference at 7.4%

250 200 150

□ GWP100 DGWP20

(a) GWP per MJ of energy contained in the fuel

^ 3000

S 2500

SJ 2000

O 1500

3 1000 Q.

□ GWP100 DGWP20

(b) GWP per kWh of electricity generated

Fig. 1. Global warming potential estimates for shale gas expressed (a) per unit of energy contained in the fuel and (b) per unit of electricity generated in a gas power plant. (Where available, estimates are given for both 100- and 20-year timeframes. Howarth et al. [17] and MacKay and Stone [23] do not give a central estimate, therefore the bar height shown is an average of their lowest and highest results.)

and Stephenson et al. [21] at 2.4% in favour of conventional gas. Similarly, MacKay and Stone [23] find that the GWP of shale gas lies within the range of possible results for conventional gas while Burnham et al. [16] suggest that shale gas in fact has a lower GWP than conventional gas (by about 6%).

Shale gas has also been compared to coal with most studies finding that shale gas has between 36% and 63% lower GWP than coal [15,16,18,23]. The exception to this is again the previously mentioned study by Howarth et al. [17] who assert that coal has a lower GWP.

However, as mentioned previously, few existing studies have considered any other impacts apart from the GWP. Therefore, this paper sets out to estimate wider environmental consequences of shale gas, in addition to the GWP. As far as the authors are aware, this is the first study of its kind globally. 1t is assumed that shale gas is used for electricity generation and its impacts are compared to electricity from conventional gas, coal, nuclear power and renewables. The focus is on UK shale gas whose exploitation is, at the time of writing, at a very early stage of development so that the results can help companies and policy makers understand broader environmental impacts of shale gas and make more informed decisions.

2. Methodology

Life cycle assessment (LCA) has been used as a tool to estimate the environmental impacts, following the LCA methodology in 1SO 14040/44 [31,32]. This is detailed below.

2.1. Goal and scope of the study

The goal of the study is to estimate the life cycle environmental impacts of electricity from UK shale gas and compare them to other electricity options (as detailed below). 1t is assumed that electricity from shale gas is generated by a CCGT, which represents more than 90% of installed gas-fired capacity in the UK [5].

As it is not clear what options shale gas would actually be replacing at this stage, it is compared to the following current (2012) electricity options in the UK [5]:

• conventional gas which supplies 27.7% of electricity; this comprises gas from the North Sea, liquefied natural gas (LNG) imported from Algeria and LNG imported from Qatar, accounting together for ~90% of UK gas supply;

• coal power plants, which generate 38.4% of electricity;

• nuclear power plants (pressurised water reactor, PWR), which provide 18.1%; and

• solar photovoltaics (PV) and offshore wind, supplying 0.34% and 2.1% of electricity, respectively; these are renewables with the fastest uptake in the UK with 284- and 47-fold increases in installed capacity, respectively, from 2003 to 2012 [5].

As shown in Fig. 2, the scope of the study for electricity from shale gas and all other options is from 'cradle to grave'. These are described in more detail in the following sections. The functional unit is defined as 1 kWh of electricity generated at the power plant (i.e. transmission is excluded from the system boundary).

2.2. System description, data and assumptions

2.2.1. An overview of the shale gas life cycle

The shale gas life cycle begins with exploration and site preparation which may include exploratory drilling during which any gas released is normally flared or simply vented to air.

Drilling may be carried out with diesel-powered or electric equipment. Exploration in the UK has focused on the Sabden and Bowland shales in North West England, requiring drilling to a depth of 1600-2800 m [33]. After a site is selected and a vertical well drilled, horizontal drilling gives access to a greater volume of the shale body containing the trapped gas.

This is followed by hydraulic fracturing: the steel casing of the well is perforated and pressurised fluid is injected into the shale. Again, this is typically achieved with diesel internal-combustion equipment. The fracturing fluid is a mixture of water, additives and proppant, the latter being a fine-grained substance that helps to keep fissures open thereby maximising gas flow; sand is the most common proppant used in fracking. The Environment Agency must approve the composition of any fracturing fluid used in the UK and any substances contained must be 'non-hazardous within the specified situation' [34].

Following hydraulic fracturing, the well formation is complete. The well completion process often involves allowing gas to escape without capture in order to clear debris and return some of the fracturing fluid (this is known as 'flowback').

After well completion, production begins. This may be interrupted by periods of 'liquid unloading' and 'workovers': the former refers to temporary disconnection of piping and venting of gas to clear excess fluid in the well; the latter is the act of re-fracking (i.e. repeating the hydraulic fracturing process) to replenish gas production from the well if it begins to slow down.

Following gas production, the life cycle is the same as that of conventional gas: the gas is treated for distribution before being fed into the national network. Alternatively, it may be liquefied to produce LNG which is then imported to (or exported from) the

Fig. 2. The life cycles of gas, coal, nuclear, wind and solar PV electricity generation. (PWR: pressurised water reactor; PV: photovoltaics. In the gas life cycle, stages unique to shale gas are indicated by black boxes, stages unique to LNG are indicated by grey boxes and white boxes apply to all three options: shale, LNG and conventional gas.)

UK via a specialised LNG carrier at LNG terminals (such as Grain LNG, near London). Once in the UK, it is regasified and fed into the national network.

In the above stages, it is clear that there is considerable potential for methane to escape directly into the atmosphere, including during fracking and from the compressors used to increase pipeline pressure during transmission. Furthermore, methane dissolved in the flowback water will slowly be released if stored in open tanks, adding to fugitive emissions. The flowback water also typically has higher levels of salinity, heavy metals and naturally occurring

radioactive materials (NORMs) than the input water because it has dissolved components of the shale formation. These contaminants may reach the environment if not handled properly: for example, disposal of untreated flowback water to surface water bodies is not permitted in the UK or the US; in the US, much of it is reused in fracking fluid or disposed of via underground injection or wastewater treatment facilities [27]. In future, alternative fracturing fluids may be used that do not dissolve as many contaminants, such as liquefied petroleum gas gels [35], but these are not yet commercially available.

Table 1

Summary of best, central and worst values for key parameters considered in this study.

Parameter

Best case

Central case

Worst case

Data source(s)

Shale well specification

Energy source for drilling and other equipment Drilling fluid

consumption per well

• Triple-cased design

• Length 5773 m (vertical 2773 m, horizontal 3000 m)

• Steel 327.61

• Concrete 207.7 m3

UK grid electricity Diesel generator

Drilling waste

disposal Fugitive emissions from drilling and completion Emissions from production phase (flaring and venting)

Estimated ultimate recovery (EUR) per well Gas composition

Fracking fluid consumption per wellc

• Water 246 m3

• Barite 19,886 kg

• Oil 4419 kg

• Unspecified inorganic chemicals (mainly salts) 3108 kg

• Bentonite 1473 kg

• Unspecified organic chemicals (emulsifiers, polymers, bio-cide) 666.6 kg

• Lignite 14.7 kg 100% Landfill

No gas vented

• CO2 11.4 g/Nm3 gas produced3

• CH4 0.264 g/Nm3

• SO2 0.1 g/Nm3

• CO 0.08 g/Nm3

• NMVOCb 0.018 g/Nm3

• NOx 0.0128 g/Nm3

• Hg 1.67 x 10~7 g/Nm3

• Rn 0.334 Bq/Nm3 84.95 Mm3 (3 bcf)

CH4 0.61 kg/m3 C2H6 0.04 kg/m3 C4H10 0.04 kg/m3 Other alkanes 0.02 kg/m3 CO2 0.13 kg/m3 He 0.001 kg/m3 Hg 2 x 10~7 kg/m3 Rn 400 Bq/m3

Water 7500 m3 (98.0%) Sand 3911 (1.917%) Polyacrylamide 2264 kg (0.0296%)

Surfactant 1313 kg (0.0214%) Hydrochloric acid 1459 kg (0.0159%)

Scale inhibitord 829 kg (0.0108%)

Biocide 424 kg (0.00554%) Sodium chloride 0.238 kg (2.58 x 10~6%)

• Water 1000 m3

• Barite 80,838 kg

• Oil 17,964 kg

• Unspecified inorganic chemicals (mainly salts) 12,635 kg

• Bentonite 5988 kg

• Unspecified organic chemicals (emulsifiers, polymers, biocide) 2710 kg

• Lignite 59.9 kg

60% Landfarming, 40% landfill

4.1 m3 gas vented per metre drilled (23,669 m3 total per well)

28.32 Mm3 (1 bcf)

CH4 0.555 kg/m3 C2H6 0.075 kg/m3 C3H8 0.05 kg/m3 C4H10 0.02 kg/m3 Other alkanes 0.03 kg/m3 CO2 0.115 kg/m3 H2S 0.045 kg/m3 N2 0.03 kg/m3 He 0.001 kg/m3 Hg 2 x 10~7 kg/m3 Rn 400 Bq/m3

Water 12,000 m3 (96.2% vol.) Sand 1215 t (3.65%) Polyacrylamide 7032 kg (0.0564%)

Surfactant 4078 kg (0.0408%) Hydrochloric acid 4532 kg (0.0302%)

Scale inhibitord 2576 kg (0.0206%)

Biocide 1318 kg (0.0106%) Sodium chloride 0.738 kg (4.92 x 10~6%)

Diesel generator

• Water 2271 m3

• Barite 183,584 kg

• Oil 40,796 kg

• Unspecified inorganic chemicals (mainly salts) 28,693 kg

• Bentonite 13,599 kg

• Unspecified organic chemicals (emulsifiers, polymers, biocide) 6153 kg

• Lignite 136 kg 100% Landfarming

54 m3 gas vented per metre drilled (312,000 m3 total per well)

Cuadrilla Resources Ltd. [37]

Ecoinvent [36]

Water consumption based on statements by Cuadrilla in the central case [40]. Range based on data from Chesapeake Energy [39]. Other components Ecoinvent [36]

Ecoinvent [36]

Ecoinvent in the central case [36]. Range based on U.S. EPA. [41]

Ecoinvent [36]

2.832 Mm3 (0.1 bcf)

CH4 0.5 kg/m3 C2H6 0.11 kg/m3 C3H8 0.10 kg/m3 Other alkanes 0.04 kg/m3 CO2 0.1 kg/m3 H2S 0.09 kg/m3 N2 0.06 kg/m3 He 0.001 kg/m3 Hg 2 x 10-7 kg/m3 Rn 400 Bq/m3

Water 29,000 m3 (90.0% vol.) Sand 82361 (9.584%) Polyacrylamide 47,657 kg (0.148%)

Surfactant 27,642 kg (0.107%) Hydrochloric acid 30,718 kg (0.0794%)

Scale inhibitord 17,457 kg (0.0542%)

Biocide 8932 kg (0.0277%) Sodium chloride 5.00 kg (1.29 x 10~5%)

Based on data from U.S. Geological Survey [42]

Ecoinvent [36]

Water consumption based on statements by Cuadrilla in the central case [40]. Range based on Wood et al. [10] and American Petroleum Institute [43]. Composition of fluid based on data from Cuadrilla Resources Ltd. [44] and the Royal Society [35]

a Nm3: normal m3 at standard conditions at t = 0 "C and p = 1 bar. b NMVOC: non-methane volatile organic compounds. c All percentages denote volume percentage. d Polycarboxylates are assumed to be used as scale inhibitors.

The above stages are described in detail in the sections that follow, with the assumptions used in the study summarised in Table 1. The analysis is based on UK-specific data where possible, drawing on the data in the public domain, including from Cuadrilla, the largest shale gas company in the UK. Where no UK data were available, other sources have been used to fill in data gaps, primarily the Ecoinvent database v2.2 [36]. Note that the 'central case' in Table 1 and the rest of the paper refers to the authors' estimate of the most likely average conditions for shale gas in the UK.

2.2.2. Well design and drilling

Well design is based on that shown in Fig. 3 using specifications provided by Cuadrilla Resources for their Preese Hall 1 well [37]. It is a triple-cased design with depth of 2773 m. Extending out from the vertical well, a typical production well has additional horizontal piping in order to access more of the shale body. These horizontal pipes may be at the same depth or 'vertically stacked'. As no commercial-scale horizontal pipes have yet been deployed in the UK, we assume that a total length of 3000 m of single-cased horizontal

pipe is added to the vertical well (this compares to Cuadrilla's first proposed horizontal pipe of 1400 m which it describes as 'modest in length' [38]). Therefore, the total length of the borehole (vertical + horizontal) is 5773 m. Using data on bore and steel casing width [37], total steel requirements for the well are estimated at 327.6 tonne and total volume of concrete at 207.7 m3 (Table 1).

Drilling fluid composition is given in Table 1 and, in addition to water, includes components such as barite (to increase drilling fluid density), bentonite (lubricant and borehole sealant), oil and various other chemicals. Drilling water consumption per well is estimated at 1,000 m3 based on a range of 246-2271 m3 given by Chesapeake Energy [39]. Cuadrilla have also informally estimated their drilling water requirements at 1000 m3 per well [40]. The range given by Chesapeake Energy is explored through sensitivity analysis later in the paper.

2.2.3. Drilling waste

Drilling a long borehole - 5773 m in this case - produces considerable quantities of waste, disposal of which may vary depending on the site and its operator. Landfilling and landfarming are the most common routes, with the latter involving spreading the waste onto agricultural land. This is normally permitted by the Environment Agency in the case of drill cuttings, whether separated or mixed with bentonite (from drilling fluid) [45]. However, the drilling waste also contains toxic components such as barite, thus the disposal route is an important parameter. In the central case, it is assumed that 60% of waste (237 kg/m drilled) is spread on agricultural land and 40% (158 kg/m) is landfilled (see Table 1). This ratio is varied as part of the sensitivity analysis.

2.2.4. Fracking fluid

Much controversy in the US has centred around the composition of fracking fluid and the refusal of companies to disclose the composition on the basis of commercial confidentiality: companies are exempt from various federal regulations that would force disclosure to the EPA [35]. In the UK, regulations are stricter: operators should disclose their fluid composition on their website and

t Conductor: a shallow casing

to aid construction of the well

Aquifer

Surface casing

Cement

Intermediate casing

Production casing

Hydraulic fracturing in the shale formation

Fig. 3. Typical modern design of shale-gas well (based on Cuadrilla design [37]).

the Environment Agency is able to demand full disclosure under the Water Resources Act 1991 if necessary [35].

1n this work we assume a fluid composition based on data from the Royal Society (RS) & the Royal Academy of Engineering (RAEng) [35] and actual usage by Cuadrilla in 2011 [44] (see Table 2). The influence of fluid composition on the results is explored within the sensitivity analysis.

As shown in Table 2, water and sand (proppant) typically comprise >99% of the fracking fluid. Based on the data from Cuadrilla, <0.05% of the fluid is made up of chemical additives, which is considerably less than has been discussed in some literature. For instance, Wood et al. [10] state that chemical additives in fracking fluid used in the UK comprise 2% of the fluid: a factor of 40 higher compared to the data reported by Cuadrilla. Thus the Wood et al. estimate of 180-580 m3 of chemicals injected per well may need revising down to 4.5-14.5 m3, much of which may be returned in flowback water. However, fracking by Cuadrilla is still at an early stage and it is not clear how their fluid composition will change as they ramp up to commercial-scale production. The RS and RAEng [35] estimate that additive chemicals comprise 0.17% of the fluid as shown in Table 2: this is more than Cuadrilla has used thus far but still 12 times lower than Wood et al.'s figure.

The total volume of fracking water reported in literature ranges from 7500 to 29,000 m3 per well [10,43]. Cuadrilla put this estimate at 12,000 m3 [40]. This study assumes this value as its central estimate. 1t should be noted that fracking fluid can be recycled to reduce resource requirements, thus lower consumption of water may be possible as shown by the above consumption range reported in literature. On the other hand, workovers (re-fracturing operations) may be performed, meaning more fluid is required, again as the upper ranges found in other studies show. These ranges are also considered as part of the sensitivity analysis.

Typically, 20-40% of the fracking fluid returns to the surface during well completion as flowback water, with more gradually returning throughout the life of the well [46]. However, how much fluid ultimately returns will vary from location to location and is not well established for UK shale formations [35]. In this study it is assumed that 80% returns to the surface over the lifetime of the well. As is current practice in the UK, it is assumed that this flowback water is stored in sealed tanks (thus avoiding direct emissions of methane and other VOCs to the atmosphere) and sent for treatment in a wastewater facility.

Some estimates have been made of the composition of shale well flowback water [47]. However, as there are no data on treatment of waste water with such composition, in this study data for treatment of black chrome coating effluent have been used [36] because this has the closest chemical composition to the flowback water. This is inevitably an approximation as concentrations of metal contaminants in flowback water differ by up to an order of magnitude. However, in the absence of specific data, it is preferable to use proxy data rather than not consider water treatment at all.

2.2.5. Estimated ultimate recovery

As mentioned above, the estimated ultimate recovery (EUR) from a shale well is a key determinant of its life cycle impacts. Logan et al. [19] use a central figure of 1.42 bcf (40.21 Mm3), while other studies have tended to assume higher figures; for instance, Jiang et al. [15] assume 2.7 bcf (76 Mm3) and Burnham et al. [16] 3.5 bcf (99 Mm3). More recent data from the U.S. Geological Survey [42] suggest that average EUR values for shale gas wells might be slightly lower than the above, albeit in a large range of 0.01-20 bcf (0.28-566 Mm3): while individual well data are not provided, the median EURs for 26 groups of shale gas wells range from 0.03 to 2 bcf (0.85-57 Mm3).

1n this study, a value of 1 bcf (28.32 Mm3) is assumed in the central case, within a range of 0.1-3 bcf (2.8-85 Mm3). Given the

Table 2

Calculation of fracking fluid components in the central case.

Cuadrilla Resources Ltd. [44] RS and RAEng [35] Assumed in this study

l/m3 Water % vol. l/m3 Water % vol. l/m3 Water % vol.

Water 1000 97.93 1000 94.6 1000 96.19

Sand 20.65 2.023 55.29 5.23 37.97a 3.652

Friction reducer (polyacrylamide) 0.438 0.043 0.734 0.069 0.586a 0.056

Surfactant - - 0.424 0.0401 0.424 0.0408

Hydrochloric acid - - 0.315 0.0298 0.315 0.0302

Scale inhibitorb - - 0.215 0.0203 0.215 0.0206

Biocide - - 0.110 0.0104 0.110 0.0106

Sodium chloride 5.120 x 10~5 0.000005 - - 5.120 x 10~5 4.92 x 10~6

a Average.

b Polycarboxylates are assumed to be used as scale inhibitors.

lack of UK production data, it is assumed that only gas is produced; there is a possibility that oil could be co-produced with gas, but where and to what extent this might be feasible has not been established in the UK and will depend on economics as well as technical ability [4].

2.2.6. Gas composition

The composition of shale gas itself varies from region to region, as does that of conventional gas. For instance, Barnett shale gas in the US comprises about 75% to >95% methane; the remainder is primarily a mix of non-methane hydrocarbons (ethane, propane, butane), carbon dioxide, nitrogen and hydrogen sulphide [48]. Gas with a higher proportion of methane is generally referred to as 'dry', whereas gas with a considerable proportion of other hydrocarbons is 'wet'; gas containing hydrogen sulphide is referred to as 'sour', while that without is 'sweet'. Since the average composition of raw UK shale gas (i.e. before any treatment) has not yet been established, and to enable like-for-like comparisons with conventional gas, it is assumed that both the conventional and shale gas have the same composition - as shown in Table 1 - assuming a 50:50 split between sweet and sour gas in the central case. In the best case scenario, it is assumed that the produced gas is sweet and, in the worst case, all the gas is sour.

In order to be fed into the National Transmission Network, natural gas must meet the requirements of the Gas Safety (Management) Regulations [49], which specify acceptable energy content and contaminant ranges. Thus the 'sour' gas component is 'sweetened' by removing hydrogen sulphide following its production. The 'sweetening' process is considered in this study.

2.2.7. Fugitive emissions

During drilling and completion of the well, i.e. prior to production, gas may be vented to the atmosphere. As there are no data on the typical amount of gas vented during this process in the UK, the estimates used in this study are based on the conventional onshore well dataset in Ecoinvent: this gives 4.1 m3 of vented gas per metre drilled [36]. With a total well length of 5773 m (see Section 2.2.2), this equates to 23,669 m3 of gas per well.

Although there is not yet a prescriptive policy on well completions in the UK, statements thus far have reaffirmed the stance that venting of gas to the atmosphere must be reduced to the 'minimum technically possible' and flaring to the 'economic minimum' [50]. This, together with recent policy decisions in the US to make so-called 'reduced emissions completions' (RECs, or 'green completions') a legal requirement, suggests that RECs will be the industry norm in the UK. These involve additional equipment being used to separate gases and solids expelled during well completion so that the gas can be fed into the production pipeline rather than being vented or flared. The US EPA has estimated [41] that an average unmitigated shale well completion (i.e. one in which all gas is

vented) emits 312,000 m3, while a REC reduces emissions by ~90%, giving 31,000 m3: a figure similar to that assumed in this study. Given the discussed uncertainty over UK completions policy, unmitigated completions are also considered as part of the sensitivity analysis (see Table 1).

For lack of UK data, emissions to air due to flaring and venting during production are assumed the same as those for conventional onshore gas production (Table 1). This should be a reasonable approximation insofar as the only major difference between shale gas and conventional onshore gas extraction is the fracking process.

2.2.8. Electricity generation

After extraction, processing and distribution, it is assumed that the gas feeds a CCGT of modern design: 400 MW capacity (265 MW gas and 140 MW steam turbine) with net efficiency of 52.5% based on LHV. This was the average efficiency of CCGTs operating in the UK in 2012 [5]. The LCA data for this plant are sourced from Ecoin-vent [36].

2.2.9. Other electricity sources

As mentioned earlier, other electricity sources considered in this study for comparison with shale gas are conventional gas, coal (subcritical pulverised), nuclear (PWR), wind (offshore) and solar PV. Their respective life cycle diagrams and the system boundaries are outlined in Fig. 2. The LCA data are sourced from prior work by the authors [51] where further details can be found.

For conventional gas, the current UK gas supply is considered. The majority (45% of gross supply in 2012) comes from the British North Sea, with a further considerable contribution (29%) from the Norwegian North Sea [5]. In addition, LNG contributes an increasing share - 15% of gross supply in 2012 - primarily from Qatar, Algeria and Trinidad & Tobago [5]. Since 2008, Qatar has been the main supplier, accounting for over 95% of the UK's LNG imports in 2012 [5]. Therefore, the following options are considered:

• pipeline gas extracted solely from the North Sea;

• LNG from Qatar; and

• LNG from Algeria.

For the other technologies (coal, nuclear, wind and solar PV), a range of values for key parameters has been considered, as follows, based on the prior work by the authors [51]:

• Coal: power station efficiency has been varied from 36% to 42%, SO2 capture by flue gas desulphurisation from 77% to 90% and NOx removal by selective catalytic reduction from 0% to 79%. End-of-life recycling of power plant components is also considered, assuming current recycling rates.

• Nuclear: the proportion of mixed oxide (MOX) fuel used has been varied from 0% to 8%. The proportion of enrichment

achieved via centrifuge has been varied from 70% to 100% with the remainder being achieved via diffusion. As for coal, recycling of power plant components at the end of life is also considered, assuming current recycling rates.

• Offshore wind: turbine capacity has been varied from 2 to 5 MW and capacity factor from 30% to 50%. End-of-life component recycling is also considered.

• Solar PV: the range of technologies considered includes mono-and multi-crystalline Si, amorphous Si, ribbon Si, CdTe and CIGS (cadmium-indium-gallium-selenide). Each technology is mounted on a slanted roof, while mono- and multi-crystalline Si are also modelled as façade and flat-roof installations. End-of-life component recycling is also included in the study.

3. Results and discussion

GaBi v6.0 LCA software [52] has been used to model the life cycles of shale gas and other electricity systems. The environmental impacts are estimated following the 2010 update of the CML 2001 methodology [53,54]. All 11 impacts included in the CML method are estimated to obtain a full picture of environmental consequences of shale gas, as opposed to the previous studies which considered mainly the GWP. The results are presented in Fig. 4 and discussed in the sections below. The estimated ranges indicated as error bars in the figure are a result of varying the parameters described in Section 2 and detailed in Table 1.

3.1. Global warming potential (GWP)

As shown in Fig. 4, the central estimate for the GWP100 of shale gas is equal to 462 g CO2-eq./kWh, with the values ranging from 412 to 1102 g CO2-eq./kWh. The main contributor to this impact (84%) is combustion of shale gas to generate electricity. This is in line with the range of 416-730 g CO2-eq./kWh reported by other authors for other regions [15,16,18,19,21]. The high maximum estimate of 1102 g CO2-eq./kWh is primarily due to the assumption of a very low EUR value of 0.1 bcf (2.83 Mm3) in the worst case compared to the worst case value of 0.5 bcf (14.2 Mm3) assumed in the studies cited above.

As also indicated in Fig. 4, the GWP100 of electricity from shale gas is comparable to that from conventional gas. While the extraction stage of the shale gas life cycle is more carbon intensive than the extraction of conventional gas (by 4-23 times in the central case), its overall life cycle GWP100 is comparable to - and in fact slightly lower than - imported LNG due to the emissions saved by avoiding liquefaction, oceanic transportation and regasification.

Assuming the worst case simultaneously for all parameters in Table 1 demonstrates that electricity from shale gas could potentially have a GWP 2.75 times higher than North Sea gas (1102 c.f. 401 g CO2-eq./kWh; see Fig. 4), which is broadly comparable to that of a coal-fired power plant (1068 g CO2-eq./kWh). However, this is an extreme case that is probably not realistic: amongst other things, it would require a very low EUR from the shale well (0.1 bcf compared to the typical >1 bcf) and no attempt to mitigate gas venting to the atmosphere during completion. Furthermore, with such a low EUR, it seems unlikely that the site would be economically attractive. It is also very unlikely that unmitigated venting would be practised and/or allowed. Nevertheless, these results indicate what could happen without appropriate legislation and regulation.

In the opposite case, in which all parameters are favourable, electricity from shale gas may only have a 2.8% higher GWP100 than North Sea gas and about 15-19% lower than LNG. Natural gas transport accounts for 17-23% of the total GWP for LNG, whereas in the shale gas life cycle transport is minimal (~2%); therefore, life cycle emissions are lower.

Compared to coal power, the GWP100 of shale gas is 51-58% lower in the central case, depending on the efficiency and fuel mix of the coal plant. However, even in the best case, shale gas still has much higher GWP100 than solar PV, offshore wind and nuclear power (by 78%, 97% and 98%, respectively).

Shifting to a 20-year time horizon for GWP does not have a considerable effect on the relative difference between shale gas and the other options compared to GWP100 (Fig. 4). This is in contrast with some previous studies, particularly Howarth et al. [17] in which GWP20 is much higher for shale gas (832-2878 g CO2-eq./kWh). This is mainly due to the relatively low level of gas venting assumed in this study in the central case (see Table 1). Consequently, direct methane emissions to air are relatively low and the impact of increasing the GWP of methane from 25 kg CO2-eq./kg CH4 for the 100-year time horizon to 72 g or more (105 g in the case of Howarth et al.) over 20 years is not so apparent.

As stated previously, the extent to which reduced-emissions completions will be adopted (or regulated) in the UK is not yet clear; if adoption is low, methane emissions will be higher and the actual GWP of shale gas will be closer to the worst case estimates of 1102 (GWP100) and 1565 (GWP20) g CO2-eq./kWh shown in Fig. 4.

3.2. Abiotic depletion potential - elements (ADP-E)

In the central case, at 0.446 ig Sb-eq./kWh, shale gas is comparable to coal power (Fig. 4). However, its ADP-E is 81% higher than for North Sea gas; in the worst possible case the shale impact rises to 0.69 mg Sb-eq./kWh, 29 times higher than North Sea gas. The latter is due to the high volume of drilling fuel assumed per unit of gas extracted in the worst case and particularly due to barite, the main contributor (44%) to this impact. However, as previously stated, this is unlikely because such a high impact is related to an extremely low EUR together with drilling fluid consumption of more than twice the volume estimated by Cuadrilla. Even so, in the worst case it is still 18% lower than the ADP-E of offshore wind power and 94% lower than that of PV (Fig. 4) because of the use of various metals in their respective life cycles. Solar PV in particular has a very high ADP-E, primarily due to depletion of silver and tellurium during the manufacture of the metallisation pastes required for silicon cell production (although copper and silver components in capacitors also contribute to this impact).

3.3. Abiotic depletion potential - fossil fuels (ADP-F)

As shown in Fig. 4, the depletion of fossil fuels in the shale-gas life cycle is 6.65 MJ/kWh in the central case. This is comparable to conventional gas (6.21-7.87 MJ/kWh) and is 43-49% lower than for coal power (10.77-12.04 MJ/kWh). Depletion of shale gas accounts for 88% of this impact with the remainder comprising gas and diesel usage during its extraction and processing. However, at 12.47 MJ/kWh in the worst case, the ADP-F of shale gas could exceed that of coal (12 MJ). This is only possible if the EUR is very low, as in the worst case (0.1 bcf), which, combined with loss of gas via venting, increases the fossil fuel depletion in the drilling and processing stages.

In all cases, solar PV, offshore wind and nuclear power have much lower ADP-F: nuclear, in particular, depletes 99% less fossil fuel than shale gas in the central case.

3.4. Acidification potential (AP)

The results shown in Fig. 4 suggest that shale gas electricity is likely to have a higher AP than conventional gas: about 4.1-7.5 times higher in the central case. 1n the worst case, at 7.17 g SO2-eq./kWh, the impact could be around 4 times worse than a

□ Extraction & Processing 0 Operation

□ Fuel Transport

□ Power plant construction

£ j 400 200 ¡S 0

□ Extraction & Processing

□ Operation ^ 1600 CL 1400 § 1200 " 1000

800 600 400 200 0

ÊO Ë"

Fuel Transport Power plant construction

□ Extraction & Processing

□ Operation

1000 900 800 700 600 500 400 300 200 100 0

□ Fuel Transport

□ Power plant construction

10,932*

9 »r if

□ Power plant construction

□ Extraction & Processing aFuel Transport

□ Operation

14 t ® 12 -

Ü 10 -8 8 -

□ Extraction & Processing üFuel Transport

aOperation

□ Power plant construction

□ Extraction & Processing aFuel Transport

Operation

Power plant construction

£■ o> S E

2000 1800 -1600 -1400 -1200 -1000 -800 600 400 200 0

v>ô vi

□ Extraction & Processing DFuel Transport

- SS Î

Operation 300 -, 250 -200 150 -100 -50 0

□ Power plant construction

Extraction & Processing Fuel Transport

Operation 300 250 200 150 100 50

Power plant construction

Extraction & Processing Fuel Transport Operation Power plant construction

1,181 +

400 ■S 350 J — 300 g.| 250 200 S 150 100 S ° 50

□ Extraction & Processing »Fuel Transport

□ Operation 60 i 50 " 40 -30 20 -10 -0

□ Power plant construction

9 X-if

Extraction & Processing Fuel Transport

0 Operation

Power plant construction

250 200

Ü ra 50

* v> V

□ Extraction & Processing aFuel Transport

aOperation

9 A §

« g 1

□ Power plant construction

Fig. 4. Life cycle environmental impacts of electricity from shale gas, conventional gas, coal, nuclear, offshore wind and PV. ("Abiotic depletion potential - elements: the range for solar PV is 3143-15,862 ig Sb-eq./kWh; ''Marine eco-toxicity potential: the range for coal is 957-1338 kg DCB-eq./kWh; ^Photochemical oxidation creation potential: the worst case for shale gas is 2502 mg C2H4-eq./kWh; ^Terrestrial eco-toxicity potential: the worst case for shale gas is 54.4 g DCB-eq./kWh).

typical UK coal power plant. However, in the best case shale gas has an AP of 0.17 g SO2-eq./kWh which is at least 30% lower than the conventional gas options and 60% lower than solar PV. Nuclear and wind power, however, are superior in terms of AP in all cases.

More than 80% of this impact is due to the combination of on-site diesel combustion for drilling and other equipment as well as for raw gas sweetening (see Table 1). Thus, if on-site equipment is powered from the grid and the extracted gas is low in H2S, as in the best-case scenario, this impact is greatly improved. Therefore, until UK production data are available to provide a typical gas

composition and typical on-site power source, the acidification impact of shale gas remains uncertain.

3.5. Eutrophication potential (EP)

With an EP of 138 mg PO|-eq./kWh in the central case (see Fig. 4), electricity from shale gas is broadly comparable to conventional gas, lying in between offshore wind (60 mg) and solar PV (280 mg). About half of the eutrophication impact is caused by diesel combustion in the on-site electric generator used for drilling

and other equipment. Hence, in the best case (where equipment is powered by the grid and the EUR of the well is high), the impact of shale gas is 25% lower than that of North Sea gas.

The high impact in the worst case (1120 mg) is mainly due to diesel usage for drilling being high relative to the very low gas output of the well. 1n that case, the EP is 38% lower than that of a coal plant. Similar to the AP, this impact is in the central case also lower for shale gas than for PV (by 51%) but higher than for nuclear (89%) and wind (57%).

3.6. Freshwater aquatic eco-toxicity potential (FAETP)

At 8.36 g DCB-eq./kWh in the central case, the FAETP of electricity from shale gas is comparable to conventional gas (2.19-5.19 g). This makes it an order of magnitude better than nuclear, offshore wind and solar PV (21, 21 and 64 g, respectively) and two orders of magnitude better than coal power (290 g).

However, it is notable that extraction of shale gas has a FAETP 11 times greater than that of Qatari LNG, with savings in transportation making up the difference. Additionally, disposal of drilling waste via landfarming is the major contributor (56%) due almost entirely to barium deposition to soil. Therefore, the drilling waste disposal route and the EUR of the well (i.e. the volume of gas extracted per mass of drilling waste produced) are the key parameters for FAETP. 1n the worst case, where all drilling waste is landf-armed (compared to 60% in the central case) and the EUR is lower, the result increases to 105 g DCB-eq./kWh. However, even in the worst case, the FAETP of shale gas is 2.75 times lower than that of coal power; in that case it is also 20% greater than the worst case scenario for solar PV (88 g DCB-eq./kWh).

3.7. Human toxicity potential (HTP)

1n the central case, electricity from shale gas has a HTP 2.9-4.4 times higher than conventional gas (see Fig. 4). As for FAETP, the main cause of this impact is disposal of drilling waste to landfarm-ing. Since drilling waste is normally classified as non-hazardous this disposal route is quite common, but these results demonstrate that disposal solely to landfill has benefits with respect to this impact: in the best case, shale gas HTP is slightly lower than LNG from either Qatar or Algeria.

However, in both the central and best cases shale gas has low HTP relative to its alternatives: nuclear, solar and coal power all have much higher impact (5, 6 and 10 times, respectively). It is only in the worst case (100% landfarming, low EUR) that shale gas exceeds the HTP of coal power (by 20%).

3.8. Marine aquatic eco-toxicity potential (MAETP)

Like the other toxicity impacts, landfarming of drilling waste is the main contributor to the MAETP of electricity from shale gas (65% of the total). For this reason, shale gas has a MAETP around two to five times higher than conventional gas in the central case (Fig. 4).

However, compared to the other electricity options, both conventional and shale gas have low MAETP: nuclear, offshore wind and solar PV are 1.6-7.8 times worse in the central case, while coal has an impact 45 times higher. With all drilling waste being landf-armed together with a low EUR, the result for shale gas might reach as high as 352 kg DCB-eq./kWh compared to an overall range for coal power of 957-1338 kg.

3.9. Ozone layer depletion potential (ODP)

1n the central case, shale gas is estimated to have an ODP directly comparable to conventional gas (15% lower than North

Sea gas; see Fig. 4). However, it should be noted that gas in general causes relatively high ozone layer depletion when compared to its alternatives. This is due mainly to leakage of fire retardant gases used in pipelines and processing facilities. Nuclear power and offshore wind, for example, have ODPs around 25 times lower than North Sea gas.

The main cause of ODP in the shale gas life cycle is leakage of halon 1211 during pipeline transport (67% of the total impact in the central case); halons are used as fire retardants and coolants in various processes related to gas pipeline use and maintenance. However, in the worst case this is accompanied by very high consumption of diesel in on-site generators relative to gas output; in this case the combined leakage of Halons in the diesel and gas transport life cycles lead to an ODP about 85 times higher than that of wind power.

Also of note is the fact that the potential range of ODP results for shale gas (9.4-51.2 ig CFC-11-eq./kWh) overlaps with the potential range for solar PV (3.6-25.2 ig). In the PV life cycle this impact is mainly due to the manufacture of tetrafluoroethylene, the polymer of which (Teflon) is often used in solar cell encapsulation.

3.10. Photochemical ozone creation potential (POCP)

Leakage of VOCs during the removal of H2S (sweetening) is the main potential cause of POCP (photochemical smog) in the life cycle of shale gas (56% in the central case). As shown in Fig. 4, the assumption in the central case that half of the gas requires sweetening results in an impact about 9 times higher than that of North Sea gas. Therefore knowledge of the average raw gas composition for UK shale is essential to quantifying this impact: as is the case for acidification potential, if the gas is low in hydrogen sulphide, the overall impact can be comparable to North Sea gas and lower than imported LNG. Any conclusion is therefore dependent on further exploration and analysis of UK shale deposits.

In the worst case, shale gas might be 98 times worse than North Sea gas and 18 times worse than the coal power life cycle. However, in this case the major contributor (70%) is venting of gas during well drilling and completion due to the worst case assumption that all gas is vented (see Table 1). Therefore, gas venting regulations (such as the requirement for reduced emissions completions) are critical to avoiding such high impacts.

In comparison to other technologies, shale gas has high POCP. In the central case it is worse than solar PV, offshore wind and nuclear power by factors of 3,26 and 45, respectively. Even in the best case, wind and nuclear power are still preferable (by factors of 3.3 and 5.6, respectively).

3.11. Terrestrial eco-toxicity potential (TETP)

As indicated in Fig. 4, the estimated TETP of electricity from shale gas in the central case is 13-26 times higher than that of conventional gas. This impact is also higher than for coal, nuclear, wind or solar PV by between 2 and 4.4 times. However, as is the case for human toxicity potential, most of this impact (over 90%) is due to disposal of drilling waste via landfarming and the subsequent deposition of heavy metals and barium in soil. Therefore, in the best case (where all drilling waste is landfilled), this impact is around a third lower than for imported LNG and an order of magnitude lower than for solar PV, offshore wind and coal. In the worst case, however, the TETP of shale gas electricity might be as much as 30 times higher than that of coal power (the next worst option). Thus, this impact is critically dependent on the disposal of drilling wastes and any regulation put in place to monitor the composition of those wastes. Greater experience of UK shale extraction is therefore necessary to properly evaluate TETP.

3.12. Sensitivity analysis

In addition to simultaneously varying the key parameters shown in Table 1, sensitivity analysis was undertaken to explore the effect of varying individual parameters on the impacts of the central case. The assumptions and the results of the sensitivity analysis are presented in Fig. 5.

As demonstrated in Fig. 5e, the EUR of the well has the greatest impact on the results with ADP-E, EP, FAETP, HTP, MAETP and TETP all increasing by seven times or more if the EUR is very low (0.1 bcf compared to 1 bcf in the central estimate). This is understandable given that much of the emissions throughout the life cycle of shale gas relate to fixed factors - energy and material consumed when drilling the well, emissions during completion etc. - therefore a well that has incurred those emissions but fails to produce a large quantity of gas has high impacts per cubic metre of shale gas output.

Fig. 5c illustrates a point made frequently throughout Section 3: the toxicity impacts (FAETP, HTP, MAETP, TETP) are highly sensitive to disposal of drilling waste. TETP in particular varies from -92% to +62% relative to the central case depending on whether waste is landfilled or landfarmed. Clearly this sensitivity will be influenced by the composition of the drilling waste which could not be considered in more detail in this study due to lack of data.

Gas composition creates variance of no more than ±4%, except for AP and POCP which vary by ±56% and ±68%, respectively (see Fig. 5f). In both cases this sensitivity is caused by the H2S in sour gas and the subsequent sweetening process required to remove the sulphur. The acidification impact is caused by direct release of H2S to the atmosphere when sour gas is vented and the subsequent release of SO2 during sweetening. The POCP sensitivity results mainly from direct release of non-methane volatile organics (NMVOCs) to the atmosphere during sweetening. Fig. 5d illustrates the fact that the POCP result is also sensitive to the volume of gas vented during drilling and completion: venting releases methane, butane, ethane and other VOCs to air, all of which have high photochemical smog creation potential.

Fig. 5a demonstrates that the choice of diesel- or grid-powered on-site equipment involves several trade-offs. While the effect on the GWP in the central case is negligible, there are benefits of 13-34% to switching to grid electricity in terms of AP, EP, ODP and POCP. Conversely, grid electricity has 65% and 37% higher MAETP and FAETP, respectively.

Changing the volume of drilling fluid independently of other parameters affects ADP-E by -36% to +60% (see Fig. 5b). This is primarily due to depletion of barite in drilling fluid. The effect of this on the other impacts is negligible (±2%). As shown in Fig. 5g and h, composition and volume of fracking fluid also have negligible effects on most impacts. Again, ADP-E is the only impact affected by more than ±2%, primarily as a result of the hydrochloric acid and biocides in the fluid.

4. Conclusions and recommendations

With the exception of a few countries, notably the US, shale gas exploration and extraction is at a very early stage of development. In the UK, while exploration is occurring, commercial extraction has not yet begun, yet its potential has stirred controversy over its environmental impacts, its safety and the difficulty of justifying its use to a nation conscious of climate change. The work reported in this paper has demonstrated that potential variation of different parameters can result in a wide range of life cycle environmental impacts for shale gas. Some of these are favourable relative to conventional gas and other alternatives, others very unfavourable.

The findings for the GWP of electricity from UK shale gas are in line with most of those reported in literature for other regions: shale gas is broadly comparable to conventional gas sources and the combustion stage is the major contributor to the impact; the total impact may be considerably worse depending on the amount of gas vented and the estimated ultimate recovery per well (among other factors). A power plant with 52.5% efficiency (LHV basis) fuelled by conventional gas has a GWP of 401-508 g CO2-eq./

(a) grid electricity 100%

diesel generator

60% 40% 20% 0% -20% -40% -60% -80% -100%

LULi-0_0_Q_000_0_0_0_0_

100% 80% 60% 40% 20% 0% -20% -40% -60% -80% -100%

246 m3 »2,271 mi

LUL1-0_0_0_000_0_0_0_0_ n' ¿<CLUI— 2 £2— I— QOI—

(c) 100% landfill 100%

60% 40% 20% 0% -20% -40% -60%

■ 100% landfarm

LUL1-0_0_0_000_0_0_0_0_

(d) 0 m3 gas vented per well

■ 312,000 m3 vented per well 100%

60% 40% 20% 0% -20% -40% -60% -80% -100%

LULi-0_0_Q_000_0_0_0_0_

700% 600% 500% 400% 300% 200% 100% 0% -100%

84.95 Mm3 (3 bcf) ■ 2.832 Mm3 (0.1 bcf)

-1 ---+T

IHD.CLÛ.

OCLCLCLCLQ. ^l-l-QOI-^ХШООШ

80% 60% 40% 20% 0% -20% -40% -60%

100% 'sweet' ■ 100% 'sour'

LULLCLCLCL < <

OCLCLCLCLQ. ' l-QOI-

УООШ

< CL1-

7,797.1 m3 (7,500 mi water) 130,148.9 m3 (29,000 m3 water)

6% 5% 4% 3% 2% 1% 0% -1% -2%

LIJU_Q_Q_Q_

QQ << 2

OQ.CLQ.CLCL WhhQOh

> < CL1-

7% ■ 6% 5% 4% 3% 2% 1% 0% -1% -2% -3%

98% water B90% water

LULLCLCLCL OQ_Q.Q_Q.Q_

Fig. 5. Sensitivity analysis for eight key parameters considered in this study, relative to central case. (a) Energy source for drilling and associated equipment (central case: diesel generator); (b) drilling water consumption per well (central case: 1000 m3); (c) drilling waste disposal (central case: 60% landfarming, 40% landfill); (d) fugitive emissions from drilling and completion (central case: 23,669 m3); (e) estimated ultimate recovery (EUR) per well (central case: 28.32 Mm3 (1 bcf)); (f) gas composition (central case: 50% sweet, 50% sour); (g) fracking fluid volume (central case: 12,475.4 m3 (12,000 m3 water)); (h) fracking fluid composition (central case: 96.2% water).

kWh while the corresponding range for shale gas is 412-1102 g CO2-eq./kWh with a central estimate of 462 g.

In the central case, the other main outcomes for shale gas are as follows:

• abiotic depletion of elements is around 50-80% higher than for conventional gas but 19-244 times lower than offshore wind or solar PV;

• acidification potential is 4.1-7.5 times higher than conventional gas and similarly worse than nuclear, wind or PV;

• eutrophication potential appears quite comparable to LNG, with results 8.8 and 2.3 times worse than nuclear and offshore wind, respectively, but 2 and 13 times better than coal and solar PV, respectively;

• freshwater eco-toxicity potential is 60% to 3.8 times worse than conventional gas but an order of magnitude better than any of the non-gas technologies;

• human toxicity potential is 2.9-4.4 times worse than conventional gas but an order of magnitude better than nuclear, solar PV or coal power;

• marine toxicity is 2-5 times worse than conventional gas but is comparable to nuclear and offshore wind and better than solar PV or coal power by 7.8 and 45 times, respectively;

• ozone layer depletion potential is lower than for North Sea gas (by 15%) and Qatari LNG (17%) but other technologies are superior in this regard, particularly nuclear power and offshore wind which have impacts two order of magnitude lower;

• photochemical oxidant creation potential is about nine times higher than for North Sea gas and is also 60% worse than coal power (the worst of the other technologies considered); and

• terrestrial eco-toxicity potential is 13-26 times worse than conventional gas and is also worse than any of the other technologies (by a factor of 2-4).

It should be noted, however, that toxicity-related impacts are associated with some uncertainty owing to a lack of data - this is not specific to LCA alone but to toxicological research in general. Therefore, the absolute values for human and eco-toxicity should be interpreted with caution. Nevertheless, the relative comparisons among the options are valid as the results are based on the same methodology and involve an equivalent level of uncertainty.

In addition to the central estimates summarised above, sensitivity analysis was conducted on eight key parameters: on-site power source, drilling fluid consumption, drilling waste disposal route, fugitive emissions from drilling and completion, estimated ultimate recovery per well, gas composition, fracking fluid volume and fracking fluid composition. Simultaneously varying these parameters can result in impacts from shale gas that are worse than all other options considered in terms of global warming, acidification, human toxicity, ozone layer depletion, photochemical smog and terrestrial toxicity.

However, it should be borne in mind that, since primary data are currently sparse owing to the early stage of development of shale gas extraction, some parameters are uncertain. The refinement of the results will be possible if/when commercial-scale production begins in the UK. This would provide further data on typical UK-specific drilling and completion emissions, gas composition, on-site power supply, emissions during the production phase and flowback water treatment.

It should also be noted that much of the controversy related to the impacts from fracking fluid is centred on the potential contamination of groundwater owing to accidents and/or malpractice. This could involve contamination with fracking fluid components or naturally occurring substances that have been mobilised by the extraction process, such as heavy metals. At present, this cannot be addressed well by LCA so that additional research is needed to complement the LCA.

This should include long-term monitoring and reporting of ground-water before and after shale gas exploration as well as better characterisation of the possible compositions of flowback water and their consequences for wastewater treatment. In particular, the concentrations of heavy metals, NORMs, oils and other components in flow-back water should be topics of future research.

In the meantime, based on the conclusions of this study, the recommended outcome depends on the perceived importance of each impact and the regulatory structure under which shale gas operates. From the government policy perspective - focusing mainly on energy security, cost and climate change - it appears likely that shale gas represents a good option for the UK energy sector (assuming it can be extracted at reasonable cost). However, a wider view must also consider other impacts and the effect of widespread shale gas use: given the large variation in potential impacts, shale gas should only be perceived as a sound environmental option if accompanied by tight regulation. This should include, for instance, compulsory reduced-emission completions, stringent controls on drilling waste disposal and careful estimation of ultimate recovery before commencing drilling in order to avoid high emissions associated with a low-output well. Appropriate regulation might also increase the extraction cost of shale gas to a point where it is competitive but discourages extraction of low-EUR sites and does not reduce investment in more sustainable technologies. Weak regulation may well result in shale gas having higher impacts than coal power, in turn failing to meet climate change and sustainability imperatives and undermining the deployment of renewable and nuclear technologies.

Acknowledgements

This work has been funded by the University of Manchester pump-priming fund and the UK Engineering and Physical Sciences Research Council, EPSRC (Grant No. EP/K011820/1). This funding is gratefully acknowledged.

References

[1] DECC. The unconventional hydrocarbon resources of Britain's onshore basins -shale gas. London: Department of Energy & Climate Change, British Geological Survey; 2012.

[2] Cuadrilla Resources Ltd. Energy security, <http://www.cuadrillaresources.com/ benefits/energy-security/>; 2011 [cited March 2013].

[3] IGas Energy plc. Shale gas in place in IGas' North West licences of up to ca. 170 Tcf, <http://ir1.euroinvestor.com/asp/ir/IGas/NewsRead.aspx?storyid=12357 614&ishtml=1>; 3 June 2013 [cited 2013].

[4] Andrews IJ. The Carboniferous Bowland Shale gas study: geology and resource estimation. London: British Geological Survey for Department of Energy & Climate Change; 2013.

[5] DECC. Energy statistics, <https://www.gov.uk/government/organisations/ department-of-energy-climate-change/about/statistics>; [cited October 2013].

[6] U.S. Energy Information Administration. Natural gas: data, <http://www.eia. gov/naturalgas/data.cfm>; [cited September 2013].

[7] Pennsylvania DEP. OG compliance - report viewer, <http://www.depreporting services.state.pa.us/ReportServer/Pages/ReportViewer.aspx?/Oil_Gas/OG_ Compliance>; [cited September 2013].

[8] DECC. Oil and gas: onshore exploration and production, <https://www.gov.uk/ oil-and-gas-onshore-exploration-and-production>; 2013 January 2013 [cited August 2013].

[9] Friends of the Earth. Unconventional, unnecessary and unwanted: why fracking for shale gas is a gamble the UK does not need to take. London: Friends of the Earth; 2013.

[10] Wood R, Gilbert P, Sharmina M, Anderson K, Footitt A, Glynn S, et al. Shale gas: a provisional assessment of climate change and environmental impacts. Manchester: A report commissioned by The Cooperative and undertaken by The Tyndall Centre, University of Manchester; 2011.

[11] IEA. Golden rules for a golden age of gas: world energy outlook special report on unconventional gas. Paris: OECD Publications; 2012.

[12] U.S. Energy Information Administration. Total energy: monthly energy review, <http://www.eia.gov/totalenergy/data/monthly/>; [cited May 2013].

[13] Broderick J, Anderson K. Has US shale gas reduced CO2 emissions? Tyndall Manchester: A research briefing commissioned by The Co-operative; 2012.

[14] RAEng. GB electricity capacity margin. London: Royal Academy of Engineering; 2013.

[15] Jiang M, Griffin WM, Hendrickson C, Jaramillo P, Van Briesen J, Venkatesh A. Life cycle greenhouse gas emissions of Marcellus shale gas. Environ Res Lett 2011;6:034014.

[16] Burnham A, Han J, Clark CE, Wang M, Dunn JB, Palou-Rivera I. Life-cycle greenhouse gas emissions of shale gas, natural gas, coal, and petroleum. Environ Sci Technol 2011;46:619-27.

[17] Howarth RW, Santoro R, Ingraffea A. Methane and the greenhouse-gas footprint of natural gas from shale formations. Climatic Change 2011;106:679-90.

[18] Hultman N, Rebois D, Scholten M, Ramig C. The greenhouse impact of unconventional gas for electricity generation. Environ Res Lett 2011;6:044008.

[19] Logan J, Heath G, MacknickJ, Paranhos E, Boyd W, Carlson K. Natural gas and the transformation of the U.S. energy sector: electricity. Golden, Colorado: Joint Institute for Strategic Energy Analysis, National Renewable Energy Laboratory; 2012.

[20] O'Sullivan F, Paltsev S. Shale gas production: potential versus actual greenhouse gas emissions. Environ Res Lett 2012;7:044030.

[21] Stephenson T, Valle JE, Riera-Palou X. Modeling the relative GHG emissions of conventional and shale gas production. Environ Sci Technol 2011;45:10757-64.

[22] Weber CL, Clavin C. Life cycle carbon footprint of shale gas: review of evidence and implications. Environ Sci Technol 2012;46:5688-95.

[23] MacKay DJC, Stone TJ. Potential greenhouse gas emissions associated with shale gas extraction and use. London: Department of Energy & Climate Change; 2013.

[24] Dale AT, Khanna V, Vidic RD, Bilec MM. Process based life-cycle assessment of natural gas from the marcellus shale. Environ Sci Technol 2013;47:5459-66.

[25] Chang Y, Huang R, Ries RJ, Masanet E. Shale-to-well energy use and air pollutant emissions of shale gas production in China. Appl Energy 2014;125:147-57.

[26] Clark CE, Horner RM, Harto CB. Life cycle water consumption for shale gas and conventional natural gas. Environ Sci Technol 2013.

[27] Jenner S, Lamadrid AJ. Shale gas vs. coal: policy implications from environmental impact comparisons of shale gas, conventional gas, and coal on air, water, and land in the United States. Energy Policy 2013;53:442-53.

[28] Shindell DT, Faluvegi G, Koch DM, Schmidt GA, Unger N, Bauer SE. Improved attribution of climate forcing to emissions. Science 2009;326:716-8.

[29] Solomon S, Qin D, Manning M, Alley RB, Berntsen T, Bindoff NL, et al. Technical summary. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, et al., editors. Climate change 2007: the physical science basis contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change (IPCC). Cambridge: Cambridge University Press; 2007.

[30] Cathles LM, Brown L, Taam M, Hunter A. A commentary on "the greenhouse-gas footprint of natural gas in shale formations'' by R.W. Howarth, R. Santoro, and Anthony Ingraffea. Climatic Change 2012;113:525-35.

[31] ISO. ISO 14040:2006: Environmental management - life cycle assessment -principles and framework. Geneva, Switzerland: International Organisation for Standardisation; 2006.

[32] ISO. ISO 14044:2006: Environmental management - life cycle assessment -requirements and guidelines. Geneva, Switzerland: International Organisation for Standardisation; 2006.

[33] Fretwell B, Haines T, Glazier J, Marsland T. Review of assessment procedures for shale gas well casing installation. Bristol: Environment Agency; 2012.

[34] Environment Agency. Unconventional gas - shale gas and coal bed methane, <http://www.environment-agency.gov.uk/business/topics/126689.aspx>; 2013 [cited 19.03.13].

[35] RS, RAEng. Shale gas extraction in the UK: a review of hydraulic fracturing. London: The Royal Society & The Royal Academy of Engineering; 2012.

[36] Ecoinvent Centre. Ecoinvent database v2.2, <http://www.ecoinvent.org/ database/>; [cited 2010].

[37] Cuadrilla Resources Ltd. Wellbore integrity - Cuadrilla land based wells. Lichfield: Cuadrilla Resources Ltd.; 2012.

[38] Cuadrilla Resources Ltd. Anna's road site, Westby, <http:// www.cuadrillaresources.com/our-sites/westby/>; 2013 [cited August 2013].

[39] Chesapeake Energy. Water use in deep shale gas exploration. Oklahoma City: Chesapeake Energy Corporation; 2012.

[40] The Energy and Climate Change Committee. Shale gas: fifth report of session 2010-12. London: The Stationery Office; 2011.

[41] U.S. EPA. Oil and natural gas sector: standards of performance for crude oil and natural gas production, transmission, and distribution. background supplemental technical support document for the final new source performance standards. Washington DC: United States Environmental Protection Agency; 2012.

[42] U.S. Geological Survey. Variability of distributions of well-scale estimated ultimate recovery for continuous (unconventional) oil and gas resources in the United States. Reston, Virginia: U.S. Department of the Interior & U.S. Geological Survey Oil and Gas Assessment Team; 2012.

[43] American Petroleum Institute. Water management associated with hydraulic fracturing. Washington: American Petroleum Institute; 2010.

[44] Cuadrilla Resources Ltd. Composition of components in Bowland shale hydraulic fracturing fluid for Preese Hall-1 well. Lichfield: Cuadrilla Resources Ltd.; 2012.

[45] Environment Agency. Regulating the recovery of waste drilling muds. The Environment Agency; 2012.

[46] Cuadrilla Resources Ltd. Fracturing fluid, <http://www. cuadrillaresources.com/what-we-do/hydraulic-fracturing/fracturing-fluid/>; 2013 [cited 22.04.13].

[47] Environment Agency. Shale gas: north west - monitoring of flowback water. The Environment Agency; 2011.

[48] Stuart ME. Potential groundwater impact from exploitation of shale gas in the UK. Nottingham: British Geological Survey; 2012.

[49] Gas Safety (Management) Regulations 1996. London, United Kingdom: Her Majesty's Stationery Office; 1996.

[50] Davey E. Written ministerial statement by Edward Davey: exploration for shale gas 2012.

[51] Stamford L, Azapagic A. Life cycle sustainability assessment of electricity options for the UK. Int J Energy Res 2012;36:1263-90.

[52] PE International. GaBi 6.0. Stuttgart, Echterdingen: PE International; 2012.

[53] Guinée JB, Gorrée M, Heijungs R, Huppes G, Kleijn R, Koning Ad, et al. III: Scientific background. Handbook on life cycle assessment: operational guide to the ISO standards. Dordrecht: Kluwer Academic Publishers; 2002.

[54] van Oers L. CML-IA characterisation factors, <http://cml.leiden.edu/software/ data-cmlia.html>; [cited 22.09.12].