Unconventional shale gas development: challenges for environmental policy and EA practice

Abstract

The growth of unconventional shale gas development has been accompanied by controversy over its environmental and social impacts. This paper reviews recent literature to clarify what is known and not known about the physical, chemical and toxicological properties of the process chemicals and wastewaters generated in hydraulic fracturing, the mechanisms and pathways by which they enter surface water and groundwater aquifers and the risks posed to human and ecosystem health. Assessing the impacts of unconventional shale gas development is clearly constrained by a lack of baseline information, complex hydrogeological histories for natural migration of hydrocarbons, lack of tracers to monitor and verify the source, timing and mechanism of contaminant migration into water resources. This is compounded by lack of transparency and accountability in policy decisions. The paper argues that managing the social and environmental risks of unconventional shale gas development calls for a new generation of impact assessment, one that marries the ideals of strategic environmental assessment, cumulative effects assessment, backcasting and deliberative and inclusive processes of community engagement towards collective risk management.

Keywords:

• Hydraulic fracturing
• risk-based environmental assessment
• unconventional shale gas development
• deliberative and inclusive process
• collective risk management

Introduction

The large-scale recovery of natural gas from tight shale formations (referred to in this paper as unconventional shale gas development or simply shale gas development) has been made possible and financially viable,¹ through the combination of two technologies: hydraulic fracturing and horizontal drilling. Hydraulic fracturing (also called hydrofracturing, hydrofracking, fracking and fraccing) creates additional permeability in a gas producing rock formation, facilitating the migration of fluids to the wellbore. It has been used in oil and gas wells since the 1940s. Horizontal drilling, also first used in the 1940s² enables a downward-plodding drill bit to bend as much as 90 degrees and continue drilling for several kilometres. The benefit of the combination is enormous. If a gas producing rock unit is 100ft thick, a vertical well drilled through it would have a production zone of a 100ft. If the well is then turned and drilled horizontally through the rock layer for 5000ft then this single well will now have a production zone 5000ft long.

As of 2013 recoverable shale gas resources worldwide, exclusive of Russia and the Middle East, were estimated at 7299 trillion cubic feet (Arent et al. 2015). Shale gas, typically more than 90% methane, can lead to energy security in many countries, and if coupled with fuel switching from coal and oil, is viewed as a potentially important transition fuel towards a decarbonised energy future (Arent et al. 2015). Natural gas is expected to make up 23% of world energy consumption by 2040 (Sieminski 2014). However, unconventional shale gas development has significant environmental burden with serious implications for human and ecological health in the medium and long-term (e.g. Goldstein et al. 2014). The dramatic growth of unconventional shale gas development in the last decade and a half has been accompanied by public protests (Long 2008) and civil suits (King et al. 2012), with hydraulic fracturing bans or moratoria imposed in various jurisdictions, e.g. Wales, Scotland, Maryland and New York in the U.S., Quebec, Newfoundland and Labrador, and New Brunswick in Canada, and attempted in several counties in Colorado, Texas, and Ohio³ and most recently, by an Indian Band in North Dakota (Van Gelder 2016).

Many see the control of social and environmental impacts from shale gas development through technology and policy as occurring in the absence of a strong body of empirical work on the science of the processes and consequences (e.g. [CCA] Council of Canadian Academics 2014; Small et al. 2014; Arent et al. 2015). This paper reviews recent literature (2006–2016), mostly peer-reviewed academic journals on unconventional shale gas development from different disciplinary perspectives to understand the nature and sources of risks in shale gas development. It seeks to explore the potential of risk-based environmental assessment to raise the quality of the debate in the negotiation for social licence for hydraulic fracturing projects. The focus is on chemical contamination of water resources in the drilling and production stages. The review is selective, seeking to provide an overview on the following questions:

(a)	What is known and not known about the process chemicals and wastewaters generated in shale gas development?
(b)	What is known and not known about the release mechanisms and transport pathways by which hydraulic fracturing chemicals and wastewaters enter surface water and groundwater aquifers?
(c)	What are the risks posed to human and ecosystem health?

By clarifying the sources and nature of the risks posed by shale gas development the paper to seeks to clarify the basis of public concern and envision the potential role that environmental impact assessment can play to promote responsible shale gas development.

The literature reviewed focuses almost exclusively on North America, the region where unconventional shale gas development is most advanced and the subject of most published empirical literature. Most is on the Marcellus shale formation (Pennsylvania, New York and West Virginia) because ‘very few samples of flowback and produced waters have been analysed and published especially for the region outside the Marcellus Shale’ (Jackson et al. 2014, p. 342).

The paper starts with a description of the basic steps involved in shale gas extraction, followed by an overview of the areas of public concern. The main discussion is structured along the three questions identified for the paper while taking into consideration the areas of public concern. The paper closes with recommendations for environmental policy and preliminary suggestions on EA practice.

Unconventional shale gas development

The typical process

Extracting natural gas from low permeability shale formation involves drilling a well vertically from the surface, typically 2–3 km down, until it reaches the shale. The drill bit is deviated to progressively greater angles (up to 90 degrees) and the drilling continued horizontally through the shale gas layer. The extent of this lateral leg varies by region, but averages about 1.5 km. During the drilling, a steel casing is cemented into the well to stabilise the wellbore wall and to prevent well fluids from contaminating the geologic formations being drilled, to protect shallow aquifers. The casing is perforated within the shale gas layer to create conduits for the fracturing fluid to flow into the target formation. Fracturing fluid is pumped at 500–700 times the standard atmospheric pressure (Lalanilla 2015) in a number of places or stages, spaced more or less evenly along the wellbore. This creates thin fractures (less than 1 mm) in the rock and a network of open fractures through which the gas can flow to the well and up to the surface (CCA, 2014; Zoback & Arent 2014) (Figure 1).

Figure 1. Drilling installation schema.

Once the fractures have been created, injection ceases, wellbore pressure is released and some fracturing fluids (up to 60%) return to the surface. Before actual commercial production starts, the natural gas that accompanies the flowback fluid is frequently collected, vented or flared from the first few wells while production rates and necessary modifications are determined. Once a single fracturing treatment is finished and the necessary modifications are completed, the fracturing process is repeated to optimise gas production. The next well on the pad is then rigged up and fractured.

Multiple wells are typically drilled on a single pad; each pad covers 3–5 acres. Flowback and produced waters constitute the hydraulic fracturing wastewaters (HFW) and are stored in impoundment ponds near the drilling sites before being sent off for disposal. They are deposited in deep injection wells where available as in the Barnett shale play⁴. Where injection wells are not available, the HFW are released to surface waters with or without prior treatment in industrial or municipal sewage treatment facilities (Rozell & Reaven 2011; Cooley & Donnelly 2012; Ferrar et al. 2013; Vidic et al. 2013; CCA, 2014; Shih et al. 2015). Some US jurisdictions (e.g. New York, West Virginia, Michigan) allow hydraulic fracturing wastewater to be spread on the road for dust suppression or de-icing (Jackson et al. 2014).

The technologies and practices in the sector are constantly evolving. There is a great deal of ‘learning-by-doing’ leading to better characterisation of the resource in the different geological environments. Horizontal laterals have grown longer (up to 3 km); formulation of fracturing fluids has been modified with experience; more stages are being done per well at higher injection rates and double the volumes; and more wells are being drilled from the same well pad (Jackson et al. 2014). Alternatives are also being developed to reduce production costs or in response to public concerns. In some areas, flowback water is reused for fracturing of subsequent wells; in others, waterless hydraulic fracturing fluids are used – saline groundwater, nitrogen gas, carbon dioxide and liquid petroleum gas. These reduce the need for new sources of water, truck transport and wastewater disposal. Under certain conditions some operators use ceramic beads, sintered bauxite and zirconium oxide to replace sand as proppants (See Table 1).

Table 1. Chemical classes of additives identified in hydraulic fracturing fluids.

Additive	% by weight	Purpose	Examples
Proppant	9	Props open fractures	Sand (0.4–0.8 mm in diameter), aluminium ceramic beads
Acid	0.400	Cleans up perforations, dissolve some rocks	Hydrochloric acid
Breaker	<0.001	Reduces viscosity, promotes flowback	Peroxydisulfates
Bactericide/biocide	0.020	Control bacterial growth	Gluteraldehyde, formaldehyde, methylisothiazolinone
Buffer agent		Adjusts/controls acidity	Sodium (or potassium) carbonate, acetic acid
Clay stabiliser		Prevents clay swelling/migration	Potassium chloride
Corrosion inhibitor	0.001	Prevents rusting of pipes	Ammonium bisulphate, methanol
Cross linker		These have higher viscosities and break down less	Potassium hydroxide, borate esters, vinylidene chloride/methyl acrylate copolymer
Friction reducer	0.090	Reduces friction between the fluid and the wellbore	Sodium acrylate, acrylamide copolymer, petroleum distillates (benzene, ethylbenzene, toluene, xylene, naphthalene)
Gelling agent	0.001	Increases fluid viscosity to carry more proppants into fractures	Guar gum, cellulose polymers, petroleum distillates (benzene, ethylbenzene, toluene, xylene, naphthalene)
Iron control	0.02	Prevents precipitation of iron oxides	Ammonium chloride, ethylene glycol
Solvent			Various polyaromatic hydrocarbons (PAHs), benzene, toluene
Surfactant	0.100	Reduce the fluid surface tension to aid fluid recoverys	Methanol, isopropanol, ethoxylated alcohol, ethylene dichloride

Sources:

Cooley & Donnelly 2012; CCA (2014); Stoiber et al. (2015).

Areas of public concern

Public concerns vary by region. The top three frequently cited are:

(a)	contamination of ground and surface water (including the safe disposal of large volumes of wastewater) (e.g. Warner et al. 2013; Arent et al. 2015);
(b)	risk of increased greenhouse gas (GHG) emissions (including fugitive methane emissions during and after production) (e.g. Howarth et al. 2011; Darrah et al. 2015); and
(c)	disruptive and psycho social effects on communities (Perry 2012) from land use changes, noise, truck traffic and health impacts (e.g. CCA, 2014; Powers et al. 2015).

Other concerns include impairment of local air quality, potential for triggering small- to moderate-sized earthquakes in seismically active areas (e.g. Mooney 2011), and the tremendous amounts of water used (e.g. Cooley & Donnelly 2012). Many of these are not unique to shale gas development but the density of hydraulic fracturing operations is much larger. Moreover, extensive high-density drilling is occurring in regions previously unexposed to oil and gas development, sometimes literally in people’s backyards (Mooney 2011).

What is known, what is not

This section considers what is known and not known about the major areas of public concerns: specifically the risks to water quality from fracturing fluid chemicals and wastewaters, the implications for human and ecosystem health, water consumption, induced seismicity, greenhouse gases and fugitive emissions.

Risks to water quality from fracturing fluid chemicals and wastewaters

The sources of potential contamination of water resources are the chemicals in the fracturing fluids, flowback and produced water:

•	Fracturing fluid is the mixture of water (ca 90% by volume), chemical additives (0.5–2%) and proppants (9%) used to stimulate the shale formation (Table 1). The formulation of fracturing fluid varies depending on fracturing type desired and the conditions of the formation being fractured (Rozell & Reaven 2011). Some of the chemicals in Table 1 are known to be toxic, carcinogenic or associated with reproductive harm. Many are considered hazardous water pollutants and are regulated in other industries (Shonkoff et al. 2014). They are rarely pure; they contain residuals from the manufacturing process, from solvents and other chemicals added to control product consistency or handling properties (CCST 2015).
•	Flowback is the fluid that returns to the surface via the wellbore within 2–4 weeks after the hydraulic fracturing treatment is completed and prior to gas production. It is a combination of the fracturing fluid, formation water and natural gas. It may also contain naturally occurring radioactive materials (NORM) (Ziemkiewicz & He 2015). The composition varies over time. The initial discharge appears dominated by the fracturing fluid chemistry, the later discharge, by the formation geochemistry (Stewart et al. 2015).
•	Formation water is water occurring naturally in the fractured geologic formation. It contains high levels of chloride (in some cases four times the concentrations found in seawater), arsenic, barium, manganese, total petroleum hydrocarbons, polyaromatic hydrocarbons and possibly NORM (Rozell & Reaven 2011).
•	Produced water is the fluid that flows to the surface with the gas through the wellbore during gas production. It comprises most of the wastewater in the life of the well since it continues to be generated even after the play is fully developed (Shih et al. 2015). Its chemical composition varies, determined by the chemical composition of the formation water and formation rock, the chemicals in the fracturing fluid, and the time since drilling (Ziemkiewicz & He 2015). The extreme pressure under which this interaction occurs shifts the chemical equilibrium; in what direction, to what extent and what products result remain poorly understood (Ziemkiewicz & He 2015). The total dissolved solids (TDS) content of produced water can range from below seawater to 10 times that of seawater. Because of its high chloride content, direct discharge of produced water to streams would hurt aquatic life and direct discharge to land would hurt vegetative growth (NSE 2008). Produced water may also contain dissolved inorganic compounds of arsenic, cadmium, cobalt, chromium, iron, nickel, vanadium and zinc, as well as NORM (Orem et al. 2014).
•	Hydraulic fracturing wastewaters (HFW) consist of flowback and produced water, and are enriched with materials from the shale formation, i.e. brines, hydrocarbons and NORM; the longer the fluids take to return to the surface the greater the concentration of formation materials.

A meta-analysis of the chemical content of hydraulic fracturing wastewaters (HFW) was done by Shih and co-investigators (2015). They drew data from wastewater generator reports filed in Pennsylvania from 2009 to 2011 and analysed 160 samples of flowback, produced water, and drilling wastes for 84 different chemicals. They conclude that typical shale gas wastewater may have higher concentrations of chloride, bromide and radionuclides than had previously been reported (p. 9563). The median concentration of eight major ions, total dissolved solids and NORM was higher in produced water than in flowback indicating the greater influence of formation conditions on produced water.

Ziemkiewicz and He (2015) focused on flowback chemistry in two studies. One looked at temporal variability in 4 Marcellus Shale well sites in northern West Virginia, with samples collected over six months during the well development and completion stages. The other looked at regional variability with samples from Ohio, Pennsylvania and West Virginia Marcellus and Utica. All were analysed for 15 organic, 17 inorganic and 13 radioactive constituents. The concentration of TDS, organics and NORMs in flowback were observed to vary over time, suggesting that the majority of these constituents originate from formation brine or from the interaction of fracturing fluid with formation minerals and organic compounds, consistent with the finding by Shih et al. (2015). The concentration of TDS, NORM, bromide, metals and organics, specifically, benzene, toluene and styrene, consistently exceeded maximum contaminant levels (MCLs)⁵ by up to three orders of magnitude. These agree with observations by Orem et al. (2014). The concentrations generally declined after onset of production but residual quantities persisted after more than eight months (Cluff et al. 2014), in contradiction to Ziemkiewicz and He (2015) and Hayes (2009) conclusion that fracturing fluids chemicals contribute little to flowback.

The amounts of chemicals can be very significant. For example, a 4 million-gallon fracturing job in the Marcellus shale used 937 gallons of hydrochloric acid and 29 gallons of methanol, yet both chemicals represent less than 0.01% of the total fluid by weight (Cooley and Donnelly 2012). In winter 2010 in northern Canada, a gas company boasted of having completed 274 fractures in 16 wells from a single-well pad over a 111-day period. The project used 5.7 million gallons of water, 50.3 million kilograms of sand and an estimated quarter of a million gallons of chemical additives (Kusnetz 2011).

As shown in Figure 2 these chemicals can move through the environment via several mechanisms and pathways.

Figure 2. Contaminant pathways in shale gas development (after Rozell & Reaven 2011).

One results from the ineffectiveness of wastewater treatment facilities. Warner et al. (2013) examined the water quality and compositions of discharged effluents, surface waters and stream sediments associated with a treatment facility site in western Pennsylvania. The treatment facility reduced the concentration of some chemicals but increased downstream concentrations of chloride and bromide above background levels. The radium levels in stream sediments (544–8759 Becquerel/kg) at the point of discharge were ~200 times greater than upstream and background sediments and above radioactive waste disposal threshold regulations (185–1850 Becquerel/kg), posing potential environmental risks of radium bioaccumulation in localised areas of shale gas wastewater disposal.

Another is failure of the drilling infrastructure. The very powerful technologies that enable access to gas reserves in low permeability rocks also induce tremendous stress on the built structure – the wellbore, steel casing and cement seals. Fluid leakage can result because of well deterioration from repeated fracturing treatments, or cement deterioration over time or improperly placed cement seals (CCA, 2014; Davies et al. 2014). These pathways may allow migration of gases and saline fluids over long time scales, with substantial cumulative impact on aquifer water quality (CCA, 2014). This was demonstrated by Darrah et al. (2015) who used noble gases tracers to study the source of fugitive gas migration in both Barnett and Marcellus shale formations. A study by Davies et al. (2014) using the Pennsylvania state database suggested that this contamination pathway is not trivial. They identified well barrier or integrity failure rate of 6.3% for the years 2005–2013.

These pathways are not unique to shale gas development but the potential is greater because of the greater well density and the repetitive fracturing treatment under such high pressures.⁶

Existing conventional oil and gas wells may also provide a subsurface pathway for shale gas contaminants to reach groundwater. In some jurisdictions hydraulic fracturing occurs in old reservoirs, where oil and gas had been produced for a long time. The risk stems from the likelihood that these ‘offset’ wells were constructed under less stringent regulations or are in degraded state and may not withstand the high pressures involved (EY 2015).

Chemical migration through the rock structure is a route that leads to groundwater contamination. It is frequently contended that the hydraulic fracturing process is not powerful enough to blow open new fissures through the thousands of feet of rock and connect horizontal wellbores to groundwater near the surface. The shale layers, the argument goes, can be a mile or more deep with thousands of feet of low permeability rock to the nearest shallow aquifer, which is precisely why gas reserves in tight shale have been so difficult to access until now. This has been used to explain the failure of many community lawsuits against hydraulic fracturing companies (King et al. 2012). However, others counter that while single fractures may not lead to wellbore-aquifer communication, the lateral for each well is fractured in multiple segments, multiple times. The probabilities increase; just how much is uncertain ([CCA] Council of Canadian Academics 2014). In British Columbia, Canada, regulators catalogued 19 separate incidents of ‘fracture communication’, one occurring between wells more than 600 m apart. Fracture communication is suspected to have caused high levels of benzene (50 times the legal maximum contaminant level), toluene and the solvent 2-butoxyethanol in groundwater in Wyoming (Jackson et al. 2014). Fracture lengths may go farther than anticipated because of weaknesses in the overlying rock layers (CCA, 2014). An examination of 44 000 wells in the US between 2010 and 2013 showed that while the average fracturing depth was 2.5 km, 16% were fractured 1.6 km from the surface, and 6%, less than a kilometre (Jackson et al. (2015).

A review of the empirical literature on the impacts of unconventional shale gas development on water resources in the US by Vengosh et al. (2014) identify several plausible risks to water quality: (a) groundwater contamination from stray gas, fracturing fluid and/or produced water from leaking wells; (b) surface water contamination from spills, leaks and poor wastewater treatment practices; and (c) accumulation of toxic elements in the soil and river sediments.

Rozell and Reaven (2011) used probability bounds analysis⁷ to model five pathways for water contamination – spills, well casing leaks, leaks through fractured rock, drilling site discharge and wastewater disposal (See Figure 2). Their process established that wastewater disposal carried the highest risk and epistemic uncertainty (lack of knowledge), reflecting the high volumes of wastewater generated (ca 200 m³ per well) and the lack of efficacy of industrial or municipal wastewater treatment facilities used to treat these wastewaters.

Potential risks to human and ecosystem health

Research on the health effects of 353 chemicals in fracturing fluid products with associated Chemical Abstract Service numbers⁸ found that 73% of the products were associated with between 6 and 14 different adverse health effects including skin, eye and sensory organ damage; respiratory distress including asthma; gastrointestinal and liver disease; brain and nervous system harms; cancers; and negative reproductive effects (Colborne et al. 2011).

However, there is paucity of scientific information on actual health effects of high volume shale gas development. There are several reasons. In the US data collection and analysis are hampered by regulatory exemptions at the federal level. Under the 2005 Energy Policy Act the natural gas industry is exempted from seven major federal laws including the Safe Drinking Water Act and the Clean Air Act. The industry is not required to disclose chemicals it considers to be proprietary. At the community level, non-disclosure compensation agreements between landowners and drilling companies for illnesses and contamination obstruct access to necessary data. It is difficult to monitor for unknown compounds and identify health risk factors in these circumstances. Uncertainties regarding the level, length and source of exposure make it difficult to attribute contamination incidents and health consequences to specific drilling activities. The toxicological complexity of the chemicals used adds further problems for causal inference. Epidemiologic studies (e.g. prospective cohort studies) frequently require significant time, rendered difficult by the rapid rate at which shale gas operations have expanded.

The studies that have been carried out give grounds for serious concern. Shonkoff et al. (2014) reviewed peer-reviewed empirical health impact, health risk literature – from trucking of the inputs to the disposal of hydraulic fracturing wastewater. A retrospective cohort study showed a positive association between density and proximity of pregnant mothers to shale gas development and the prevalence of congenital heart defects and possibly neural tube defects in their newborns (McKenzie et al. 2014, cited in Shonkoff et al. 2014). Another study measured oestrogen and androgen receptor activity in surface and groundwater samples in Colorado. Water samples from the more intensive areas of natural gas development exhibited statistically significant higher endocrine disruption activity than samples from sites with fewer or no operations (Kassotis et al. 2014, cited in Shonkoff et al. 2014).

As mentioned in several studies reviewed earlier HFW contain high levels of salinity, toxic metals and radioactivity. These wastewaters are frequently stored in impoundment ponds before being trucked for disposal. Concerns arise over accumulation of radiological material as well as the potential for spills or leaks due to overfilling or ruptures in the impoundment liners. Their release to rivers and streams even with prior treatment has likewise raised concern about the safety of drinking water.⁹ Industrial or municipal wastewater treatment plants do not have the capacity to remove high level of total dissolved solids (TDS) and naturally occurring radioactive materials (NORM) found in HFW (Vidic et al. 2013; Warner et al. 2013; [CCA] Council of Canadian Academics 2014). The surface water discharges from wastewater treatment plants – municipal or industrial – have been linked to increasing concentrations of salinity (bromides and TDS) in surface waters of Pennsylvania (Ferrar et al. 2013). High levels of iodide and bromide in hydraulic fracturing wastewaters have been shown to react with disinfection chemicals (chlorine, chloramine or ozone) to form compounds highly damaging to living cells and genetic material (Parker et al. 2014). When chlorinated, samples with as low as 0.01% HFW produce known mutagens and carcinogens (trihalomethanes and haloacetonitriles). In municipal wastewater-impacted river water, the presence of 0.1% hydraulic fracturing wastewater increased the formation of another carcinogen (N-nitrosodimethylamine) during chloramination (Parker et al. 2014).

The potential health impacts of air pollutants from shale gas development came to light in a study by John Hopkins University of first floor and basement indoor radon between 1987 and 2013 in Pennsylvania, a state with traditionally high radon readings. Radon is the second leading cause of lung cancer worldwide. Readings were significantly higher in counties with hydraulic fracturing activity compared to those without, a difference that did not exist prior to 2004, the beginning of hydraulic fracturing in the state. Buildings using water wells registered 21% higher readings than those using municipal water. Forty-two per cent of radon readings in Pennsylvania buildings now exceed US EPA action level of 148 Becquerel/m3 compared to 39% reported from 1999 to 2007. However, the researchers could not definitively establish the source of the radon increases, whether well water, air near gas wells, natural gas used in the home, or tighter building seals in recent years (Casey et al. 2015). There are also health concerns over other air pollutants such as NO_x, volatile organic compounds (VOC) and particulate emissions (PM_2.5). Both are precursors to ground level ozone which is associated with increased respiratory and cardiovascular morbidity and mortality (Jerett et al. 2009, cited by Shonkoff et al. 2014). The development of the Marcellus shale is predicted to contribute between 6 and 18% of NO_x and VOC emissions in the region in 2020¹⁰ (Roy et al. 2014).

The impacts on ecosystem health are examined the least in the literature. Shale gas development requires extensive infrastructure – roads, well pads, compressor stations, pipeline rights-of-way and staging areas (Linley 2011; Moran et al. 2015). Land impacts may include deforestation with resultant destruction and fragmentation of wildlife habitat, and adverse effects on existing land uses for agriculture and tourism. The fragmenting of habitat by the infrastructure can impair essential ecosystem goods and services. Preliminary evidence suggests two primary effects – patch shrinkage (indicator of habitat loss) and edge effects, i.e. the diverse physical and biotic changes associated with the boundaries of two habitats. Habitat fragmentation is among the most important threats to biodiversity and edge effects are drivers of change in fragmented landscapes as they can have serious impacts on species diversity and composition, community dynamics and ecosystem functioning (Laurance et al. 2007). (Moran et al. (2015) measured the changes in land use within the maturing Fayetteville shale gas development region in Arkansas between 2001/2002 and 2012. They found that on average, individual gas wells fully developed about 2.5 ha of land and modified an additional 0.5 ha of natural forest.

Water consumption

Estimates of water requirements in unconventional shale development range vary widely. The US EPA reports that extracting shale gas requires 40,000–1,000,000 gallons to drill each well and between 2.3 million and 3.8 million gallons of water to fracture it. Rogers (2011) estimates that a Barnett deep shale gas well requires approximately 250,000 gallons of water for drilling and an additional 3.8 million gallons on average for fracturing. These data reflect the significant variation among shale formations and differences in the depth to the target formation (Nicot et al. 2011). Estimating the water requirements is further complicated by the uncertainty about how many times a single well will be fracked over the course of its productive life and limited publicly available data. This said, a review of literature led Krupnick et al. (2014) to conclude that ‘the gross quantities of water being withdrawn are low per Btu of energy produced, trivial compared to withdrawals from other sectors (golf courses, for examples), and low relative to average daily flows of water in many source streams or rivers’ (p. 19) and that what matters is where and when the water is withdrawn. A similar conclusion was reached by Arent et al. (2015).

Induced seismicity

Minor earthquakes have been associated with shale gas development, but most studies have concluded that the quakes are due to reinjection of wastewater, not the fracturing process (e.g. Petersen et al. 2015). This was the conclusion in relation to the swarm of small earthquakes west of Fort Worth, Texas. The area, which had had no recorded earthquakes for 150 years, shook with 27 magnitude 2 or greater earthquakes in a three month period. Similar conclusion was reached on the sixfold increase in the number of earthquakes in Oklahoma from 2008 to 2014 over the period 1880–2008¹¹, including the 2011 magnitude 5.6 quake in central Oklahoma, the most powerful in its history and the biggest human-induced quake in the US (Pérez-Peña 2015). The trigger is argued to be the weakening of a pre-existing fault by fluid pressure (Ellsworth 2013). A different conclusion was reached in a study of three seismic events in north-eastern British Columbia and western Alberta Canada in 2014. Two were considered as likely induced by the fracturing treatment; the third was ‘enigmatic’ (Atkinson et al. 2015, p. 7).

Greenhouse gas and fugitive methane emissions

Estimates of greenhouse gas emissions from the natural gas sector are widely recognised as highly uncertain because of a critical lack of empirical data and lack of accessibility to industry data. This is viewed as particularly acute in the shale gas sector (Arent et al. 2015).

In a review of the literature, Arent et al. (2015) point to inconsistent approaches to data use and other assumptions as raising caution against cross-study comparison and hindering collective understanding of the magnitude of the problem. Some studies suggest that shale gas drilling leads to fugitive gas contamination in a subset of drinking water wells near drill sites (Darrah et al. 2015), with Howarth et al. (2011) estimating that 3.6–7.9% of the methane from shale gas production escapes to the atmosphere in venting and leaks over the lifetime of a well. Others suggest that the methane detected is unrelated to shale gas development (Molofsky et al. 2013).

Methane contamination of aquifers and drinking water has been highlighted as a problem in the Marcellus Shale gas play. Methane that enters wells is not considered a health hazard because it is not very soluble in water. However, methane can be oxidised by bacteria, resulting in oxygen depletion. Low oxygen concentrations can increase the solubility of arsenic or iron. Additionally, the proliferation of anaerobic bacteria under these conditions may create water- and air-quality issues. When methane degasses, it can create turbidity and, in extreme cases, explode (critical concentration at 10 mg/L). At least three explosions have occurred in Pennsylvania water wells since 2004 (Brantley 2015). However, some contend that more than 70% of wells in north-eastern PA may have natural gas (Molofsky et al. 2013). Methane naturally forms at great depth from high-temperature maturation of organic matter (thermogenic), but also at shallow depths through bacterial action (biogenic).

The work of Darrah et al. (2015) is helping to clarify the issue. They used noble gases (argon, krypton and neon) to trace the pathways for fugitive gas contaminations. Of the eight clusters of water wells examined where fugitive contamination was identified, all were related to well integrity, and not large-scale migration of gas following the hydraulic fracturing process. Fugitive gas migration, they conclude, occurs (a) along the void between pipings, tubings or casings (annulus); (b) through faulty or corroded casing; or (c) along legacy or abandoned wells.

Microbiology of HFW and implications for impact mitigation

An emerging area of research with potential implications for impact mitigation measures used by the industry, examines the relationship between the organic chemistry and microbiology of hydraulic fracturing wastewaters. Microbial activity appears to contribute to the degradation of organic compounds in produced waters, a major component of HFW. This has implications in terms of the effectiveness of impact mitigation measures and alternatives used in the industry.

HFW contain not only the chemicals and micro-organisms potentially introduced during the fracturing procedure, but also those acquired from the geologic formation, during the return to the surface, and subsequent storage. Despite the addition of biocides to fracturing fluids and the hyper salinity of produced water, anaerobic micro-organisms were successfully cultivated in produced water samples by Akob et al. (2015). The growth of anaerobic fermenters (i.e. methane-producing, and sulphide-producing bacteria) indicates that either the biocides were not effective at killing all micro-organisms in the fracturing fluid and the formation, or their effectiveness or concentrations declined over time, or there is microbial adaptation to environmental changes.

Unlike the spatial consistency of the inorganic chemistry of Marcellus shale flowback and produced waters demonstrated by researchers such as Shih et al. (2015), the organic chemistry and microbial viability were highly variable across 13 wells sampled by Akob et al. (2015), They hypothesise that these organisms may be fermenting organic compounds present in the fracturing fluids or in the formation, with the fermentation products serving as nutrients for other microbial populations. The work of Mohan et al. (2013) lends support. They compared the microbial ecology in prefracturing fluids (fracturing source water and fracturing fluid) and produced water at multiple time points from a natural gas well in south-western Pennsylvania. Aerobic species dominated in the prefracturing fluids, but declined in produced water with an increase in anaerobic (methane forming and sulphide forming) species. This implies that reuse of flowback and produced water in subsequent fracturing treatments could potentially accelerate sour-gas production, biofouling and scale formation, increasing production costs.

Strong et al. (2014) and Kekacs et al. (2015), on the other hand, examined the potential for biodegradation of organic compounds in HFW under aerobic conditions, a condition that obtains in accidental releases to surface waters and shallow soils. Microbial action removed between 57% to more than 90% of added dissolved organic carbon within 6.5 days. In other words, the organic compounds biodegraded. However, salinity concentrations of 40 g/L or more (characteristic of produced water, the main component of HFW) completely inhibited degradation even with microbial communities pre-acclimated to salt. This highlights once again the importance of pre-treatment of HFW to reduce total dissolved solids prior to disposal.

Table 2 gives an overview of what is known and what is not on the impacts of unconventional shale gas development.

Table 2. Summary of what is known and what is not.

Area of concern	Known	Not known
Process chemicals and wastewaters generated	• Seventy three per cent of HF fluid products with CAS numbers associated with 6–14 adverse health effects on skin, eye, sensory organ and reproduction • Marcellus shale region – Flowback and produced water contain levels of salinity (TDS), toxic chemicals and NORM exceeding regulatory thresholds • Wyoming – high levels of poly aromatic hydrocarbons (PAH) in groundwater • Pennsylvania Industrial and municipal WWTF do not effectively remove high levels of TDS and NORM • Minor earthquakes observed in Marcellus and Barnett shale regions attributed to reinjection of wastewater, not fracturing process; two events observed in Alberta Canada attributed to fracturing treatment	• Identity of over half of estimated 750 chemicals used in HF fluids • Outcomes of possibly synergistic interaction between HFF chemicals and formation chemicals under high pressure and temperature • Whether frequent and successive minor seismic events eventually trigger a big one
Release mechanisms and transport pathways to chemical contamination of water resources	• Major routes for toxic chemicals and NORM include spills, well casing leaks, leaks through fractured rock, drilling site discharge and wastewater disposal • Canada – ‘fracture communication’, between wells >600 m apart	• Baseline information on groundwater quality for different sites • Extent of well integrity failure • Extent of inter wellbore communication
Risks posed to human and ecosystem health	• Positive association between density and proximity of pregnant mothers to SGD and congenital heart defects and possibly neural tube defects in newborns • Colorado – water samples from areas with intensive SGD have significant higher endocrine disruption activity than samples from sites with fewer or no operations • High levels of iodide and bromide in HFW react with disinfection chemicals to form known mutagens and carcinogens • Pennsylvania – increased radon readings in buildings after SGD above MCLs • Arkansas – individual gas wells fully changed about 2.5 ha of land and modified an additional 0.5 ha of natural forest.	• Level, duration and source of exposure to water and air pollutants • Confounding variables in existing studies • Occupational health hazards posed by NORM in HFW • Source of radon emissions increase • Long-term impacts of land use changes on species diversity, composition, dynamics and ecosystem functioning
Risk of increased GHG and fugitive methane emissions	• Routes are leaks from well casing and through fractured rock	• Source of fugitive methane emissions, whether biogenic or thermogenic
Disruptive and psycho social effects on communities	• Mostly self-reported impacts on quality of life from traffic noise, increased population and threats to health from HFW	• Baseline health studies • Confounding variables on reported health impacts
Existing impact mitigation measures	• Microbial activity degrades organic compounds in produced waters • Pennsylvania – Aerobic micro-organisms dominate in pre-fracturing fluids, decline in produced water with an increase in anaerobes • Anaerobic micro-organisms successfully cultivated in produced water	• Impacts on flowback reuse

Recommendations for EA policy and practice

Managing the social, environmental and health risks from unconventional shale development requires an understanding of the chemistry, geochemistry and microbiology of the processes involved and waste streams generated specific to each geological formation. The discourse is constrained by several, frequently unacknowledged, factors. In most cases baseline information on key geological and environmental variables is not established, and in most jurisdictions, information on process inputs is proprietary.¹² For many drilling sites, data are not publicly available on total water withdrawals, number of wells drilled, water conservation approaches and wastewater management practices (Arent et al. 2015). Publicly available data are of variable quality and open to divergent interpretation ([CCA] Council of Canadian Academics 2014).¹³ In some studies the sample size and/or sampling protocol limit the generalisability of results. Water contamination studies and consequently health impact studies, are complicated by the multiplicity of potential sources and pathways, the complex hydrogeological histories for natural migration of hydrocarbons, lack of tracers to monitor and verify the source (natural versus anthropogenic), timing and mechanism of contaminant migration into shallow aquifers (e.g. [CCA] Council of Canadian Academics 2014). The debate is compounded by lack of transparency and accountability in policy decisions (e.g. Arent et al. 2015).

The Panel of Experts who wrote the CCA Report (2014, p. xvi) conclude:

Public acceptance of large-scale shale gas development will not be gained through industry claims of technological prowess or through government assurances that environmental effects are acceptable. It will be gained by transparent and credible monitoring of the environmental impacts.

Needed: a regional cumulative effects assessment guided by precaution and the community

The paucity of baseline information and post development studies renders questionable claims of evidence-based decisions. Given the uncertainties shown in Table 2, impact assessment of shale gas development must be guided by the precautionary principle. The assessment must extend beyond the project site and timeframe. A regional environmental assessment with a robust cumulative effects monitoring programme as argued by Therivel and Ross (2007) is called for since there will likely be multiple projects in the region, with multiple impacts on the community and the regional environment unfolding over a long period of time. The large number of wells and required infrastructure impose substantial cumulative impacts on communities and ecosystems. Cumulative effects assessment allows the linking of the different scales of impact assessment but maintains the focus on an agreed upon receptor (the community, the receiving environment) and how it is affected by the totality of interventions and activities.

The environmental assessment process arguably needs to be framed within a strategic (policy) perspective or it will ‘default to project driven approaches’ (Noble 2008, p, 86) and lack the ability to inform the management of region-wide cumulative impacts. However, rather than ‘projecting trends and identifying desirable futures from a range of competing possibilities’ as proposed by Noble (2008, p. 89), working ‘backwards from a particular desired future end-point or set of goals to the present’ to determine the feasibility of that desired future and the policy measures required to reach it (Robinson 2003; p.842) is more amenable to the participation of the ‘non experts’ from the community, hence more democratic. Holmberg and Robert (2000) argue that backcasting is particularly useful ‘when the problem to be studied is complex… dominant trends are part of the problem, the problem to a great extent is a matter of externalities, and the scope is wide enough and the time horizon long enough to leave considerable room for deliberate choice’ (p.294). These are precisely the conditions that obtain in large-scale shale gas development. To acknowledge community divisions, ‘backcasting from principles, rather than from scenarios’(Robinson 2003, p. 842) may be more appropriate.

Backcasting is explicitly normative and integrates policy choices in the analysis (Robinson 2003). Evidence to date suggests that the impacts of shale gas development are not only dependent on geology, but also operator-specific. Policy level actions are needed to enable project level impact management measures. For instance, if sustainability were one of the desired principles, certain policy directions (e.g. exempting HFW from regulations that protect public health) and industry choices (e.g. use of toxic chemicals in fracturing fluids and non-disclosure) would necessarily be precluded or systematically and actively discouraged.

Regulate siting and design

The precautionary principle ‘denotes a duty to prevent harm, when it is within our power to do so, even when all the evidence is not in’ (CELA, n.d.).¹⁴ Given scientific evidence, albeit limited, suggesting wellbore-aquifer communication ([CCA] Council of Canadian Academics 2014; Jackson et al. 2014) a precautionary approach would lead to shallow fracturing (<1 km from surface) near protected groundwater being banned¹⁵ or be subject to safeguards, as argued by Jackson et al. (2015). These include: (1) mandatory registry; (2) pre-project baseline of chemical concentrations; (3) different casing and cementing process and additional pressure tests and cement; (4) sufficient safety margin to accommodate subsurface uncertainties and (5) corrective action and mitigation plans in case of anomalous well behaviour. Fracturing in seismically active faults should be banned. If allowed, operators should be required to manage injection rates to minimise pore pressure increases at depth and to install local seismic monitoring arrays ([CCA] Council of Canadian Academics 2014; Zoback and Arent 2014). Regional level industrial wastewater treatment facilities should be considered as they impose a lower administrative burden and are easier to monitor for compliance (Rahm & Siha 2012). Operators should be required to contribute to develop the capacities of these facilities for removing high levels of TDS and NORM.

Operators and their suppliers should be required or incentivised to use:

•	closed-loop liquid dispensing system for bulk shipping containers to ensure worker safety and prevent spills,
•	separate water tanks to prevent mixing of hydraulic fracturing wastewater with freshwater,
•	protective liners for hydraulic fracturing wastewater pits and tanks to prevent water seepage,
•	non-toxic chemicals in fracturing fluids,
•	a tracer chemical in their hydraulic fracturing fluid, and a monitoring system for inter-wellbore communication.

Regulate transparency and accountability

The precautionary approach to risk management requires that in the absence of scientific consensus that an action is not harmful, the burden of proof that it is not harmful, falls on those taking an action that may or may not be a risk (Stewart 2011). With regards to hydraulic fracking fluid chemicals, the burden of proof of safety falls on those who use them. The chemical composition of hydraulic fracturing fluids should be posted on company websites along with the Chemical Abstracts Services (CAS) registry number. Proprietary claims should be justified and substantiated.

Full disclosure of fracturing fluid chemicals and the chemical composition of flowback/produced water is necessary, but insufficient for assessing the environmental and health risks associated with drilling and fracturing. Information is needed not only on concentration of each chemical but also its mobility, persistence in groundwater and surface water, and bio-accumulation properties on its own and as a mixture, and interactions under high temperature and pressure. Without such knowledge it is difficult to define how best to mitigate accidental releases of chemicals or flowback/produced water to the environment.

The permitting process should require from industry contribution to a research fund to finance independent studies, such as:

•	baseline on groundwater quality, and in rural areas, critical wildlife habitat,
•	sources and the mechanisms of actual and potential water contamination,
•	search for non-toxic chemical additives,
•	conditions for optimal bioremediation of HFW,
•	health and quality of life status of communities before and after unconventional shale gas development (e.g. [CCA] Council of Canadian Academics 2014; Arent et al. 2015).

Documentation and reporting systems for wastewater discharge and spills – when, what, where, how much – should be required, with penalties imposed for infractions. Company performance should be regularly audited and posted on government website, including infractions, local controversies and lessons learned.

Engage the community: iterative, analytic and deliberative process towards social licence

The history of community resistance to siting of hazardous waste and nuclear facilities suggests that that involuntary risks will be more easily accepted by the public if it is perceived that the distribution of risk and benefits is equitable, that the risks are [inherently] unavoidable, not imposed and the decision-making process is transparent (Kunreuther & Slovic 1996, p.118).

Environmental assessments for shale gas development should systematically and explicitly integrate risk considerations. The gaps and weaknesses in the data used should be identified, and probability distributions demystified. Using a tool similar to the environmental risk assessment matrix proposed by Hurley et al. (2013) might be useful to publicly acknowledge that assessing risk is not a purely scientific enterprise but ‘inherently subjective … a blending of science and judgement with important psychological, social, cultural, and political factors’ (Slovic 1999, p. 689). The rules over the framing of the risk information, definition of what constitutes risk, selection of endpoints, should be socially negotiated within the context of the community. Such a contextualist risk-based approach to EA could lead to greater public understanding of and trust in public and private sector decisions and move towards a collective risk management.

An iterative process of analysis and deliberation (Perry 2012, p.358) with a ‘community’ of communities, (i.e. a highly divided community) will no doubt be protracted and ‘inefficient’. Creative experimentation will be needed as environmental assessment practitioners stray away from time tested procedures.

The rapid rate of development of this economically important sector, the acute deficits in our understanding of the different risks, the diversity and complexity of the contexts in which the development takes place warrants a new generation of impact assessment, one that marries the ideals inspiring the pioneers in strategic environmental assessment, cumulative effects assessment, backcasting and deliberative inclusive processes of community engagement. It should rejuvenate the 40-year-old profession.

Notes

1. Unconventional oil and gas development in the United States is responsible for approximately 1.7 million jobs, $63 billion in federal, state and local taxes, and an overall contribution to the US economy of $238 billion (Zoback and Arent 2014).

2. From http://energy.wilkes.edu/pages/158.asp

3. From http://keeptapwatersafe.org/global-bans-on-hydraulicfracturing/

4. Play refers to a group of oil fields or prospects in the same region that are controlled by the same set of geological characteristics.

5. Maximum concentration of a chemical allowed in public drinking water systems; established by the U.S.E.P.A.

6. Gas production in stimulated wells decline sharply after two years (Jackson et al. 2015).

7. PBA has critiqued as a simple worst-case technique. Proponents argue that the computationally simple and mathematically rigorous bounds generated are useful where tail risks and best-case/worst-case scenarios are of interest. PBA can be used to determine if a desirable or undesirable outcome resulting from a decision is even possible, whether the current state of knowledge is appropriate for making a decision, or as a complement to other risk analysis methods (Rozell & Reaven 2011).

8. CAS registry number is a unique number assigned to each chemical; it is very useful when searching for information about a specific chemical structure, as well as polymers, mixtures, alloys and substances whose exact formula is unknown or variable. The CAS number provides index names, synonyms, structure diagrams, stereochemistry, molecular formulas, enriched with literature references to the substance and experimental and predicted property data.

9. EPA sets 500 mg/l as the MCL for drinking water. The WHO advises the same limit.

10. 2009 concentrations: NO_x = 60 tons/ day, VOCS = 70 tons/day; 2020: NO_x = 140 tons/day, VOCs = 100 tons/day.

11. Study Links Swarm of Quakes In Texas To Natural Gas Drilling, from http://www.manufacturing.net/news/2015/04/study-links-swarm-of-quakes-in-texas-to-natural-gas-drilling?et_cid=4529420&et_rid=561296081&location=top

12. Information on the chemical composition of additives used in hydraulic fracturing fluid is available through an online chemical disclosure registry, FracFocus (www.fracfocus.org), but public information on flowback and produced water constituents is, until very recently, non-existent.

13. For instance, state regulatory agencies confirmed 116 cases of well-water contamination in recent years from drilling activities in Pennsylvania, Ohio and West Virginia. But a recent scientific study in Arkansas’s Fayetteville Shale found no evidence of drinking water contamination for 127 homes in the region (Jackson et al. 2014). In groundwater in Wyoming the EPA found benzene at 50 times the safe level as well as toluene and 2-butoxyethanol, both common in fracturing fluids.

14. Definition by the Canadian Environmental Law Association, from http://www.cela.ca/collections/pollution/precautionary-principle

15. In British Columbia, Canada, a special permit is required for wells drilled above 600 m. In Germany, drilling is not allowed less than 3,000 m from the surface.