Academic year 2015/2016

THE WATER-ENERGY-FOOD NEXUS OF PRODUCTION FROM SHALE OIL, SHALE GAS AND OIL SANDS

By Lorenzo Rosa

Supervisor: Maria Cristina Rulli Co-supervisor: Paolo D’Odorico Co-supervisor: Kyle Frankel Davis

September 2016

Abstract Unconventional fossil fuel resources have recently emerged as new important energy sources and they are expected to play a fundamental role in meeting energy demand in the near future. We consider global shale oil, shale gas and oil sands assessing consequences on the Water-Energy-Food security Nexus.

Shale resources are globally abundant and widespread. Extraction of shale oil and natural gas is performed through hydraulic fracturing, a water intensive process that is not free from environmental and social impacts. It is unclear to what extent and where the development of shale resources could compete with water and food security. Here we consider the global distribution of known shale deposits suitable for oil and gas production and evaluate the impacts on water resources for food production and other human and environmental uses in the same region. We find that 39% of world’s high quality shale deposits is located in areas affected by water stress and 7% is situated in regions where irrigation is expected to meet the growing food demand. In these regions shale oil and gas production would likely threaten water and food security. These results highlight the need for adequate policies to avert social, economic, and ecological consequences of shale resource extraction.

Oil sands deposits account for a third of globally proven oil reserves, extend over large natural areas, and have extraction methods requiring large volumes of freshwater. Little work has been done to quantify some of the environmental impacts of oil sands operations. Here we examine forest loss and water use for the world’s major oil sands deposits. We calculate rates of water use and forest loss both in Canadian deposits where oil sand extraction is already taking place and in other major deposits worldwide accounting for ≈93% of global oil sand reserves. We estimated that their full exploitation could result in 3 -1 2 -1 1.31 km yr of freshwater, 8700 km of forest loss, and 383 Mtonne CO2eq yr in greenhouse emissions. The expected escalation in oil sands extraction thus portends extensive environmental impacts. While unconventional fossil fuels extraction have multiple environmental impacts, there have however been substantial economic benefits that bring to an ever-evolving technological innovation, which is lowering environmental impacts. Furthermore, energy security of some countries has been strengthen. Thus there are clear and ongoing tradeoffs between economic development, energy, and the environment. Lorenzo Rosa Department of Civil and Environmental Engineering, Politecnico di Milano, Milan I-20133 Italy

Prof. Maria Cristina Rulli, Thesis Supervisor Department of Civil and Environmental Engineering, Politecnico di Milano, Milan I-20133 Italy

Prof. Paolo D’Odorico, Thesis Co-Supervisor Department of Environmental Sciences, University of Virginia, Charlottesville, VA 22904

Dr. Kyle Frankel Davis, Thesis Co-Supervisor Department of Environmental Sciences, University of Virginia, Charlottesville, VA 22904

Acknowledgements First of all, I would like to thank my thesis committee, Prof. Maria Cristina Rulli (chair), Prof. Paolo D’Odorico, Dr. Kyle Frankel Davis. They have provided me a tremendous challenge and support, and each has been an outstanding mentor.

I would like to thank my parents, Manuela and Marco, for all the support and for all the opportunities that have given during these years of study. I would like to thanks my brother, Andrea, for being a great mate in all these years.

I would like to thank all my Italian friends and all the amazing people that I met in the past few years in Stockholm and Charlottesville. A great thanks to Matteo and Marco.

I would like to thank Madagascar Oil for providing information about Madagascar oil sand concession areas and Chris W. Baynard (Baynard Geospatial Consulting: http://www.baynard-geospatial.com) for sharing the shapefile of Orinoco Heavy Oil Belt in Venezuela.

CONTENTS INTRODUCTION ...... 5 WATER-ENERGY-FOOD NEXUS ...... 9 DEFINITIONS ...... 11 PART 1: SHALE GAS AND SHALE OIL ...... 13 HISTORY ...... 17 EXTRACTION ...... 17 EXPLORATION/PLANNING ...... 18 DIRECTIONAL DRILLING ...... 18 HYDRAULIC FRACTURING ...... 20 PRODUCTION ...... 21 METHODS ...... 24 SHALE DEPOSITS IN WATER STRESSED REGIONS ...... 28 SHALE DEPOSITS OVER IRRIGATED AREAS ...... 31 WATER FOR FRACKING OR WATER FOR IRRIGATION? ...... 40 BALANCING ENERGY AND FOOD SECURITY ...... 41 REFERENCES ...... 43 PART 2: OIL SANDS ...... 52 CANADA...... 55 RECOVERING THE OIL ...... 57 UPGRADING ...... 69 OIL SANDS IN THE WORLD ...... 71 METHODS ...... 77 ENVIRONMENTAL IMPACTS OF OIL SANDS PRODUCTION ...... 84 TRADEOFFS BETWEEN ECONOMY, ENERGY, AND THE ENVIRONMENT ...... 95 REFERENCES ...... 99 CONCLUSIONS ...... 105 DISCUSSION ...... 109

INTRODUCTION In recent years oil and gas corporations have shown a rising interest in unconventional fossil fuels, likely in response to the increasing global energy demand (U.S. EIA, 2013; Exxon Mobil, 2016), its predominant reliance of fossil fuels (U.S. EIA, 2013; Exxon Mobil, 2016; British Petroleum, 2016), scarcity of conventional fossil fuel resources (Gordon, 2012), and technological innovations that have substantially reduced extraction and processing costs (Speight, 2013). Unconventional fossil fuels are hydrocarbons found in deposits that cannot be tapped with standard production methods (i.e., based the extraction of hydrocarbons that naturally flow into production wells) (Rogner, 2007) but require more complex and advanced technology (Chew, 2014). These fossil fuels exist both as oil (oil sands, tight/shale oil, deep-sea oil, heavy and extra-heavy crude oil) and natural gas (tight/shale gas, coal bed gas and gas hydrates) (Rogner, 2007). Reliance on unconventional fossil fuels is dramatically transforming the exploration and production industries (Gordon, 2012). While the global oil demand is projected to rise by about 20% from 2014 to 2040 (Exxon Mobil, 2016), it is expected that within the same time period the contribution of unconventional oil will increase from 25% to about 40% (Exxon Mobil, 2016). By 2040, 10% of world oil production will come from oil sand deposits (U.S. EIA, 2013; Exxon Mobil, 2016), which are bigger by an order of magnitude than conventional crude oil deposits (Mossop, 1980). Moreover, shale oil production (also known as tight oil) is expected to more than double contributing from 4% in 2015 to 10% in 2040 of world oil demand (U.S. EIA, 2016 a-b; British Petroleum, 2016; Exxon Mobil, 2016). Natural gas consumption is expected to rise of 40% from 2014 to 2040 (British Petroleum, 2016; Exxon Mobil, 2016). Unconventional gas will cover a big share of this increase, shale gas production will surge from 10% in 2014 to 30% in 2040 of world natural gas supply (U.S. EIA, 2016 a-b; British Petroleum, 2016; Exxon Mobil, 2016).

Figure 1. Hydrocarbon resources pyramid. This work is separated in two different parts. Part 1 focuses on world shale resources and their consequences on the water-energy-food nexus. Part 2 mainly focuses on world oil sands deposits and their environmental implications on water use and forest loss related to oil sands extraction and processing, with a little focus on the water-energy-food nexus. Shale are low permeability sedimentary rocks containing high quantities of hydrocarbons (Holditch, 2007). The oil and natural gas contained in these shale deposits can then be tapped using a water-intensive process known as hydraulic fracturing (Jiang, 2013; U.S. DOE, 2013; Vidic, 2013), a method not free from environmental and social impacts (Vidic, 2013). Indeed, various studies have shown how hydraulic fracturing is associated with the use of substantial amounts of water (Nicot, 2012; Clark, 2013; Godwin, 2014; Scanlon, 2014; Gallegos, 2015; Chen, 2016) as well as declines in regional water quality (Osborn, 2011; Warner, 2012; Vidic, 2013; Stokstad, 2014). Other related consequences are methane migration from faulty seals around well casings (Jiang, 2011; Howarth, 2011; Vidic, 2013; Brandt, 2014), health hazards (i.e., groundwater contamination), impacts on regional air quality (Kargbo, 2010; McKenzie, 2012; Bunch, 2014), seismic triggering (Rutqvist, 2013), forest cover loss, habitat fragmentation (Droham, 2012; Kiviat, 2013; Brittingham, 2014), and biodiversity loss (Kiviat, 2013; Souther, 2014; Brittingham, 2014). Of the various socio-environmental consequences of shale deposit development, its impacts on water availability or accessibility are arguably the most profound yet poorly understood (Kargbo, 2010; Moniz, 2011; Nicot, 2012; Reig, 2014; Mauter, 2014). In some regions, the relatively high rates of water

use for horizontal drilling and hydraulic fracturing could lead to a competition between water appropriation for shale rock stimulation and other human and environmental needs (e.g., food production, environmental flows) (Nicot, 2012; Mauter, 2014; Freyman, 2014). Therefore the withdrawal and depletion of water resources for shale deposit development is a significant challenge situated at the water- energy-food nexus (Howells, 2013; Rulli, 2016). Meeting humanity’s increasing water demand for food and energy production, while protecting ecosystem needs (Hoekstra, 2014) is expected to be a major task of this century. Despite growing interest in shale resources, there is only a limited understanding of the pressure that their extraction could place on local water resources. It remains unclear to what extent the water requirements of shale gas and shale oil production would compete with other uses by ecosystems and society and contribute to unsustainable water use (World Economic Forum, 2015). Such a trade-off is especially worrisome for regions prone to water stress, where additional water will already be needed to enhance food production and prevent crop water stress induced by climate change (Hoekstra, 2014; Haddeland, 2014). Further, in areas with an active water market (Debaere et al, 2014), hydraulic fracturing could contribute to an increase in water prices, as already observed in the case of South Texas (Nicot, 2012; Mauter, 2014). The limited understanding of the water demands of shale deposit development thus prevents the implementation of a sound management plan for the use of these energy sources (Mauter, 2014). There is therefore a pressing need for a quantitative assessment and mapping of where shale resource development could induce or exacerbate water stress as well as intensify the competition for water between food and energy production (Ayensu, 1999; Fiksel, 2006; Bazilian, 2011). Here we examine the global distribution of known shale deposits suitable for oil and gas production (Kuustra, 2013) and identify the regions in which water use for hydraulic fracturing could compete with agriculture. We analyze the average annual surface water stress at 0.5° degrees resolution (~ 50 km at the Equator) for the world’s high-quality shale deposits (also known as “shale plays”); unlike previous studies (Reig et al., 2014), we account for the water needed for shale oil and gas development, agriculture, industry, as well as environmental flows required to maintain key ecosystem functions. We contextualized shale development impacts on water resources considering other human and environmental uses in the same region, while also considering local groundwater stress (Gleeson et al., 2012) and expected increase in water demand for irrigation (Muller et al., 2012).

Little is known about the environmental implications of oil sands extraction and processing. For the case of Alberta (Humphries, 2008), previous investigations have focused on GHG emissions (Charpentier, 2009) and the human health effects resulting from oil sands production (Kelly, 2010). There are also other important environmental impacts associated with these extraction and treatment processes. First, the total water footprint of oil sand extraction remains poorly understood to date, except for some values of freshwater withdrawal reported in Alberta (Alberta Environment and Parks, 2015). It is also unclear whether these deposits and extraction operations occur in areas of relative water scarcity and whether they may enhance water stress. Moreover, the effects of oil sand extraction on vegetation cover have yet to be quantified despite conspicuous losses of forest during the excavation of shallow deposits and habitat fragmentation from in situ infrastructure and exploration (Schneider, 2006). With all of these apparent environmental impacts in mind, we examined five countries whose deposits account for 93% of the global reserves of recoverable crude oil from oil sands and cover 162,750 km2 – equal to the size of Tunisia. Specifically, the aim of this study was to estimate the amount of water required for the extraction process, the ongoing and expected loss in forest cover, associated GHG emissions due to extraction and processing, and the number of people that could be potentially affected by the development of oil sand deposits.

WATER-ENERGY-FOOD NEXUS Renewed debate over food security has emerged after 2007-2008 and 2011 food crises (Howells, 2013). In response commercial pressures on land and water are increasing worldwide (Rulli, 2012). Furthermore, access to water is a concern, with an increasing number of people living in water stressed areas (Mekonnen and Hoekstra, 2016). Moreover, global energy demand is expected to increase of 25% by 2040 (U.S. EIA, 2016 a-b; British Petroleum, 2016; Exxon Mobil, 2016). At the same time water, energy and food are three pillars fundamental to preserve security, prosperity and equity (Bazillian, 2011). Water, energy and food are also important to reach the Eight Millennium Goals (United Nations, 2005) and the Seventeen Sustainable Development Goals, such as: eradicate malnutrition bringing food to 1 billion people actually malnourished, ensure safe and constant water supply to 1.2 billion people facing water shortages, bringing access to electric energy to 1.3 billion people. Hence, an efficient management of these resources needs to be taken in the coming decades (Bazillian, 2011). The Seventeen Sustainable Development Goals commit subscribing countries to new action targets aimed at achieving sustainable water use, energy use and agricultural practices, as well as promoting more inclusive economic development (United Nations, 2014).

The Water-Energy-Food Nexus describes the complex and inter-related nature of our global resources systems (Biggs, 2015). It has emerged as a useful concept to describe and address the complex and interrelated nature of our global resource systems, on which we depend to achieve different social, economic and environmental goals. In practical terms, it presents a conceptual approach to better understand and systematically analyze the interactions between the natural environment and human activities, and to work towards a more coordinated management and use of natural resources across sectors and scales (FAO, 2014). This can help us to identify and manage trade-offs and to build synergies through our responses, allowing for more integrated and cost-effective planning, decision-making, implementation, monitoring and evaluation. A Nexus approach helps us to better understand the complex and dynamic interrelationships between water, energy and food, so that we can use and manage our limited resources sustainably (Biggs, 2015). It forces us to think of the impacts a decision in one sector can have not only on that sector, but on others. Anticipating potential trade-offs and synergies, we can then design, appraise and prioritize response options that are viable across different sectors (FAO, 2014).

Energy, water and food are valuable resources that are interdependent and determine human well- being. Water is used in power generation; in extraction, transport and processing of fossil fuels; in irrigation to grow crops used to produce biofuels. Energy is necessary to water provision, to power

systems that collect, transport, distribute and treat water. Energy is needed to produce fertilizers and to prepare land, harvest crops and process agriculture products.

Water-energy-food nexus has and will have greatest consequences (IEA, 2012). Energy and water will have a rising demand due to global population and economy growth, climate change effects, higher standard of living and higher food demand, particularly a dietary shift to more water intensive food (i.e., from a plant towards a meat based diet) (World Water Assessment Programme, 2012). All these drivers will create a more water-constrained future, which will amplify the mutual vulnerability of energy, food production and water (IEA, 2012).

For the energy sector, constraints on water can challenge the reliability of existing operations as well as the physical, economic and environmental viability of future projects. Water constraints can occur naturally (i.e., droughts and heat waves), or be human-induced (i.e., growing competition among users or regulations that limit access to water). Equally important to water-related risks confronted by the energy sector, the use of water for energy production can impact freshwater resources, affecting both availability (the amount downstream) and quality (their physical and chemical properties).

Water is becoming so important that is used as a criterion for assessing the viability of energy projects. The availability and the access to water could become an issue for shale gas, shale oil and oil sands expansion and development. Understanding energy-water stressed areas is of fundamental importance for future planning action of development of unconventional fossil fuels resources. This, on the one hand, will create less environmental issues, less risked capital expenditures and less failures. On the other hand, it will require the use of better technologies and a better integration between water and energy policies.

Agenda 21 highlighted the need for an integrated assessment of water, energy and food management to support the decision making process and help to reach the Eight Millennium Goals (Howells, 2013). Here we consider an integrated regional assessment of water, energy and food for global shale resources and oil sands. While, extraction of fossil fuels from unconventional fossil fuels has environmental impacts, the issues generated from an insecurity of access to water, energy and food suggest that the economic and security related issues may be a stronger motivators to change in future planning actions (Bazilian, 2011).

DEFINITIONS from Kuustra, 2013

Remaining oil and natural gas in-place. Original oil and gas in-place minus cumulative production. Technically recoverable resources. The volumes of oil and natural gas that could be produced with current technology, regardless of oil and natural gas prices and production costs. Economically recoverable resources. Resources that can be profitably produced under current market conditions and under current technological development. Proved reserves. The most certain oil and gas resource category, but with the smallest volume, is proved oil and gas reserves. Proved reserves are volumes of oil and natural gas that geologic and engineering data demonstrate with reasonable certainty to be recoverable in future years from known reservoirs under existing economic and operating conditions. Proved reserves generally increase when new production wells are drilled and decrease when existing wells are produced.

Figure 2. Stylized representation of oil and natural gas resource categorizations (Kuustra, 2013).

Page intentionally left blank

PART 1: SHALE GAS AND SHALE OIL Shale is a fine-grained sedimentary rock formed from the compaction of silt and clay (Ridley, 2011). Shale deposits typically contain organic material (Speight, 2013), which is usually transformed in natural gas and oil and stored in the relatively high porosities. Shale has very low permeability; in other words, the degree to which a fluid or gas will flow through a rock is very low, oil and gas can only be removed and produced economically if the permeability is increased fracturing the rock (U.S. DOE, 2013). Shale rocks are not only the source of the oil and natural gas, but also the reservoir and seals for conventional oil and gas accumulations (Denney, 2009). Interstitial spaces in shale formations are very tiny. However, these spaces can occupy a significant volume of the rock. This allow shale to hold a great amount of oil and gas that cannot flow due to low permeability. Oil and gas industries discovered new methods to create artificial porosity and permeability within the rock (Speight, 2013).

SHALE GAS

Shale gas is gas that remains tightly trapped in shale and consists mostly of dry gas with 60%-90% v/v of methane, but some formation produce wet gas (i.e., ethane, propane, butane and pentane) (Speight, 2013). Gas generated is stored in situ and can be found in three forms:

 Free gas in the pore spaces and fractures;  Adsorbed gas, where gas is electrically stuck to the organic matter and clay;  Dissolved in the organic matter.

TIGHT GAS

It is used to indicate natural gas produced from low-permeability sandstone and carbonate reservoir (U.S. EIA, 2016), it can be extracted fracturing the rock. The difference between tight gas and shale gas is the higher amount of sandstone in the tight gas reservoir (Speight, 2013).

COAL BED METHANE

Coal bed methane can be found in every coal seam (Al-Jubori, 2009). Historically was considered a nuisance during coal mining extraction. Nowadays, development in technologies let to exploit this resource of natural gas using similar methods used for shale gas (Ridley, 2011). Deep coal seams undevelopable for mining operations can be exploited for the extraction of coal bed methane (Al-Jubori, 2009). Generated gas is adsorbed by the organic material that forms coal. Large volumes of stored gas

are possible because the internal surface area of the micro porosity where the gas is adsorbed is very large (Al-Jubori, 2009).

TIGHT/SHALE OIL

Tight oil is oil produced from low permeability sandstones, carbonates and shale formations (U.S. EIA, 2016 a). Tight oil production is a term used by oil and gas industry rather than shale oil production (U.S. EIA, 2015). Tight oil production is a more encompassing term with respect to different geologic formations producing oil at any particular well.

Figure 3. Illustration of conventional and unconventional oil and natural gas reservoirs. Source: U.S. EIA. As other unconventional resources, also shale accumulations tent to be distributed over a larger area than conventional accumulations and usually require advanced technology to be economically productive. Technological advances in horizontal drilling, hydraulic fracturing and 3D mapping has yielded substantial cost reductions making unlocking huge quantity of shale resources (shale gas, shale oil) in USA and Canada. This commercial success has opened up the possibility to produce oil and natural gas from shale resources in other parts of the world. However, commercial shale development of the type demonstrated in the United States requires the ability to rapidly drill and complete a large number of

wells in a single productive geologic formation. The logistics and infrastructure necessary to support this level of activity, including the drilling and completion processes, the manufacturing of drilling equipment, and the distribution of the final product to market are not yet evident in countries other than the United States, Canada, China, and to some extent, Argentina. Other above the ground factors such as ownership of mineral rights, taxation regimes, and social acceptance also play a role in decisions regarding the development of shale and tight resources.

Although the shale resource estimates will likely change over time as additional information becomes available, it is evident that shale resources constitute a substantial share of overall global technically recoverable oil and natural gas resources (Kuustra, 2013). Two-thirds of the assessed, technically recoverable shale gas resource is concentrated in six countries: USA, China, Argentina, Algeria, Canada and Mexico (Figure 4). Similarly, two-thirds of the assessed, technically recoverable shale oil resource is concentrated in six countries: Russia, U.S., China, Argentina, Libya and Venezuela (Figure 5) (Kuustra, 2013). The portion of technically recoverable resource which is translated into reserves in each of these plays will depend on economic decision made by companies (U.S. National Energy Technology Laboratory, 2013).

Technically recoverable resources of shale gas (Trillion cubic meters) 50 45 40 35 30 25 20 15 10 5 0

Figure 4. World technically recoverable resources of shale gas expressed in trillion cubic meters of natural gas. Source: Kuustra, 2013.

Technically recoverable resources of shale oil (Billion barrels) 80 70 60 50 40 30 20 10 0

Figure 5. World technically recoverable resources of shale oil expressed in billion barrels of oil. 1 barrel =0.159 m3. Source: Kuustra, 2013.

HISTORY Shale gas and shale oil extraction come through a combination of existing technologies:

 The knowledge that shale rock contains gas and oil;  Hydraulic fracturing of rock to open the pores and allow the extraction of hydrocarbons. Hydraulic fracturing dates 1949;  Horizontal drilling, which is in use in the oil industry from 1970s;  Seismic exploration and growing computer power led to the development of 3D reconstruction of rock strata in 2000s.

In 1990s George Mitchell brought these four elements together in Texas and discovered that large quantity of natural gas could be extracted from deep shale. This make shale permeable enough for oil or gas to escape. Indeed shale often forms the cap that holds in place the profitable oil and gas reservoirs that have migrated into permeable sandstones beneath. In 2000s shale gas extraction developed quickly in the USA starting from Texas and then to North Dakota, Pennsylvania and Ohio thanks to a rapid increase in natural gas prices between 1998 and 2008 (from 2$/Mcf to 10$/Mcf) (U.S. DOE, 2013).

These technological breakthroughs have made commercially viable to recover gas trapped in tight formations, such as shale or coal and at the same time to recover oil from shale oil reservoirs.

EXTRACTION Extraction is performed stimulating the reservoir by creating a fracturing network to give enough surface area to allow sufficient production from the additional enhanced reservoir permeability (Speight, 2013). All shale and tight reservoirs require fracture stimulation to connect the natural fracture network to the wellbore (Gale, 2007). The use of horizontal drilling in conjunction with hydraulic fracturing has greatly expanded the ability of producers to profitably recover natural gas and oil from low permeability geologic plays.

Basic steps of shale gas and shale oil production are (Ridley, 2011):

1. Seismic exploration. Underground rock formations are mapped using sound waves and 3D reconstruction to identify the depth and thickness of appropriate shale. 2. Well pad construction: a platform for the drilling rig is levelled and hard-cored. Roads and infrastructures needed are built. 3. Horizontal drilling: a drilling derrick drills up to 12 holes down to the shale rock.

4. Hydraulic fracturing: the casing of the horizontal pipe is perforated with small explosive charges and water mixed with sand is pumped through the hole at high pressure to fracture the rock. 5. Slick water fracturing: is a method of hydro-fracturing which involves adding chemicals to increase the fluid flow. 6. Water disposal: tanks collect water that flows back out of the well the water is generally reused for future fracking or disposed as waste. 7. Production: a ‘Christmas tree’ valve assembly and a tank remain on site to collect gas, which then flows through pipeline. While a Pump Jack is installed to lift oil to the surface and storage tank facilities.

EXPLORATION/PLANNING The realization of shale oil and shale gas extraction is more expensive and labor intensive than conventional reservoirs. Since the construction of a well pad is very expensive and intensive activities of exploration with highly specialized expertise are needed to characterize the reservoir and plan future eventual extraction.

Characterization of shale gas and shale oil resources using geophysical methods has increase in importance (Chopra, 2012). Prior to recovery a number of vertical wells (usually two or three) are drilled and fractured to determine if shale gas is present and can be extracted. After more wells (10 to 15) are drilled and fractured to: characterize the shale, examine how fractures will tend to propagate and establish if the shale could produce gas economically. Moreover, other wells are drilled (up to 30) to verify long term viability of the shale (Speight, 2013). Once the reservoir proprieties and contents have been defined, the drilling program and recovery operations will start.

DIRECTIONAL DRILLING Most wells drilled for oil and gas are vertical wells drilled straight into the earth. Drilling at an angle other than vertical can stimulate reservoir in a way that cannot be achieved using vertical wells. Horizontal wells begin at the surface as a vertical well. Drilling progresses until the drill bit is a few meters above the target. At that point the pipe is pulled from the well and a hydraulic motor is attached between the drill bit and the drill pipe. The drill bit can drill with the desired inclination. The hydraulic motor is powered by a flow of drilling mud down the drill pipe and drill it with the desired direction.

Directional and horizontal drilling have been used to reduce the footprint of oil and gas field and increase the length of the pay zone in a well. Horizontal trajectory length is up to 4 km (U.S DOE, 2013). This let the reservoir to be exposed as much as possible to the wellbore.

Horizontal wells are created to intersect a greater number of naturally existing fractures in the reservoir. The direction of the drill path is chosen based on the known fracture trends in each area.

Drilling for shale gas can penetrate the water table. A proper casing of shale gas wells is fundamental to preserve groundwater contaminations and to guarantee a proper functionality of the well during production. Casing and cementing is done to prevent leakage. The casing in the horizontal section is then perforated using explosive charges to enable the flow of hydraulic fracturing fluids out of the well into the shale and the flow of natural gas out of the shale into the well.

6 to 8 horizontal wells can be originate from the same pad. Typical well pad area is rectangular with an area of 15,000 to 20,000 m2 (Speight, 2013). This area is cleared leveled and cemented. Stacked wells and multilateral drilling can be used to increase area developed and production. Stacked horizontal wells are drilled when shale is sufficiently thick. One vertical wellbore can be used to produce gas from horizontal wells at different depths. Multilateral drilling involves the drilling of two or more horizontal wells from the same vertical wellbore. Horizontal wells access different areas of the shale at the same depth, but in different directions.

Figure 6. Well pad during drilling operation. Photo courtesy: Nick Price.

HYDRAULIC FRACTURING Low permeable rocks requires extensive fractures (natural or induced) to produce commercial quantities of gas and oil (Speight, 2013). Massive hydraulic fractures are created to effectively connect a huge reservoir area to the wellbore (Speight, 2013). Maximizing the total stimulated reservoir volume plays a major role in successful economic production.

Hydraulic fracturing is a procedure that pump liquids down a well into subsurface rock units under pressure that are high enough to fracture the rock. The goal is to create a network of interconnected fractures that will serve as pore spaces for the movement of oil and natural gas to the well bore. This method increase well production rates of oil and gas (Speight, 2013).

In this process water and sand are pumped at high pressure into the well. Water is the driving fluid used in the hydraulic fracturing process. Sand, named also proppant, is used to prop the fractured shale, prevent its closure and allow the flow of natural gas into the well. Chemicals are added improving hydraulic fracturing performances. The chemicals used are proprietary. The fissures created in the fracking process are held open by the sand particles so that natural gas in the shale can flow up through the well. Once released through the well, natural gas is captured, stored and transported to the relevant site processing unit.

Approximately 30 meters of wellbore is hydraulically fractured at a time, so each well must be hydraulically fractured in multiple stages, beginning at the furthest end of the wellbore. Cement plugs are used to isolate each hydraulic fracture stage and must be drilled out to enable the flow of natural gas up the well after all hydraulic fracturing is complete. When the water is pumped into the well the entire length of the well is not pressurized. Instead, plugs are inserted to isolate the portion of well where the fractures are desired. Only this section of the well receives the full force of pumping. As pressure builds up in this portion of the well, water opens fractures, and the driving pressure extends the fractures deep into the rock unit. Once pressure is released fluid flows back out through the top of the well. Flowback water not only contain proprietary blend but also chemicals naturally present in the reservoir. In many cases flowback water can be reused in subsequent hydraulic fracturing operations, this depends on the quality of the flowback and on the economies of other management alternatives (disposal). Disposal in injection wells place the flowback water in underground formations isolated from drinking water resources.

Re-fracturing the reservoir is an option that is becoming more and more commonplace (Cramer, 2008) and can yield additional recoverable reserves.

Following the hydraulic fracturing process the well is flowed back and tested using a controlled flaring process. In some areas a pipeline ready to take the gas to market will be in place and flaring will not be necessary (U.S DOE, 2013).

Figure 7. Well site during hydraulic fracturing stimulation. Source: Kansas Geological Society.

PRODUCTION After all of the wells on a pad have been drilled, completed (hydraulically fractured), and prepared for production (well heads, piping and surface equipment installed, flowback period completed) the wells are ready for production. The water and hydrocarbons (oil or condensate) produced along with the natural gas from multiple wells on a pad is separated and stored in tanks on the pad. Wells are connected to the production facilities and production can start. Most recoverable oil and gas is usually extracted after few years. Approximately 25% of a shale gas well’s gas production emerges in the first year and 50% within four years. After the output falls very slowly and wells are expected to continue supplying gas for about 30-50 years (Ridley, 2013).

Conventional well production last 30 years or more with a smooth production over time. Because shale gas production has been occurring recently, the production lifetime of shale gas wells is not fully established. Although it is observed that shale gas decline quicker than conventional natural gas production. Once a well no longer produces in an economic way, the well head is removed, wellbore is filled with cement to prevent leakage of gas into the air, surface is reclaimed and the site is abandoned to the holder of the land surface rights.

Figure 8. A shale gas well pad with multiple adjacent tress. Source: GoMarcellusshale.com

Figure 9. A shale oil well pad with a Pump Jack. Source: AllEagleFord.com

Figure 10. Schematic representation of well pad facilities. Source: Chesapeake Energy.

METHODS World shale deposits

Global maps of shale deposits were acquired from Advanced Resources International, Inc., who have developed the most updated internationally recognized geo-referenced dataset of the spatial extent of high quality shale areas (Kuustra, 2013). In the case of the United States the map of high quality shale areas came from the U.S. National Energy Technology Laboratory (U.S. NETL, 2016). These geo-databases provide information on 228 shale basins in 96 countries and classify shale deposits depending on their ability to yield oil, dry gas or wet gas. In this study we focus on high quality shale areas (or “shale plays”) that contain high quality shale gas and shale oil deposit, and therefore offer the most profitable opportunities for oil and natural gas extraction in the near future. Other lower quality and less explored areas within these basins are not included in this assessment.

Figure 11. World shale basins and high quality shale areas (shale play). Assessing water for shale development To identify the shale deposits in which oil and gas extraction could be limited by water availability and compete with water uses for agriculture and other uses, we overlaid maps of major shale basins (Kuustra, 2013) with the global distribution of water stress areas (Mekonnen and Hoekstra, 2016). We then calculated the water footprint of shale deposit development. The amount of water required to stimulate a horizontal well through hydraulic fracturing (WW) depends greatly on local geology, technology used and operational factors applied (Nicot, 2012; Scanlon, 2014; Gallegos, 2015). Chen et al. (2016) calculated that 80,047 wells were drilled in the U.S. from 2008 to 2014, accounting for an average water use (W) of 11,259 m3 per well, in agreement with previous work

(Jiang et al., 2014). Interestingly, shale oil and shale gas wells drilled in the same area use the same amount of water (Scanlon et al., 2014). In the calculation of the water footprint of shale deposit extraction we assume that only one stimulation is performed during the lifespan of a well. Therefore our estimate of the water footprint of shale gas and oil is conservative because in some cases, to increase production, hydraulic fracturing may be performed a few times over the lifetime of the well, provided that it is necessary and economically profitable (Gregory, 2011). Additional water is required to drill a well. We assume that 1000 m3 of water are used for drilling and “cementing” each well, based on previous studies (Scanlon et al. (2014), Clark et al. (2013)). Based on existing North American technology and the assessed recoverable shale resources, the number of potential wells (n) in each shale high quality area was calculated according to criteria developed by the U.S. Energy Information Administration (U.S. EIA, 2014) as:

푇푅푅푗 푛푗 = 퐸푈푅푗 where TRR is the technically recoverable resources of oil or natural gas (i.e. oil and natural gas that can be extracted, based on current technology, but without accounting for economic profitability) (Kuustra, 2013) and EUR is the average estimated ultimate recovery for oil or natural gas during the lifetime of a well, which is assumed to be equal to 30 years (U.S. EIA, 2014). The subscript j indicates whether the well is used to produce oil or natural gas. TRR values came from Kuustra et al., 2013. EUR values came from existing estimates for the United States in 2014, namely, 108.7∙103 bbl/well and 52.63∙106 m3/well for oil and gas, respectively (U.S. EIA, 2014). We assumed that wells are uniformly distributed within shale plays. Depending on the geology, the returning hydraulic fracturing fluid can be up to 70% of the injected water (Gregory, 2015). This water is returned as a brine rich of chemicals and heavy metals. It can be recycled and reused on site, transported to wastewater treatment facilities, or injected in disposal wells (Gregory, 2011; Warner, 2012; Mauter, 2014; Gregory, 2015). As part of the industry’s efforts to reduce freshwater use, an economical production of shale resources will require a management of this fracturing return fluid (Gregory, 2011). Here, three management scenarios for fracturing return fluid (rf) were analyzed: no recycling, 50% recycling, and 100% recycling. The number of drilled and stimulated wells depends also on the extent to which machinery and infrastructures needed for drilling and production (e.g., drilling rigs, trucks, pumps, water tanks, roads, and pipelines) are available (Kuustra, 2013). Therefore the wells are not drilled and stimulated all at once but within a timeframe of a few decades, here assumed to be 30 years. Thus every year the number of

drilled wells is assumed to be constant and equal to n/30. Thus the annual water footprint of well drilling and stimulation is: m3 W ∙ n ∙ rf WF ( ) = frac year 30 Water is usually withdrawn from nearby water sources (Gregory, 2011). We assumed that the water used for fracturing stimulations (WFfrac) is extracted within the same 0.5 degree grid cell where drilling takes place because oil and gas companies will likely try to minimize the transport costs of water (Vandecasteele, 2015). Generation of surface water stress map The global distribution of annual surface water stress was calculated as in Mekonnen and Hoekstra, 2016 but accounting also for the consumptive use of water for hydraulic fracturing in high quality shale areas. Surface water stress (WS) is defined as the ratio of the blue water footprint (WF) of human activities (i.e., municipal, agriculture, mining, and other industries) and the total blue water availability (WAtot) in a grid cell (Mekonnen, 2016). We calculate WS in grid cells at a 0.5° resolution (i.e., ~ 50 km at the equator), the local blue water footprint (WFloc) in each grid cell and total water availability (WAtot) in each grid cell.

The blue water availability in a grid cell (WAtot) is the sum of the local blue water generated in that cell (WAloc) and the net blue water availability from the upstream grid cells (WAnet,up).

WAtot,j = WA loc,j + WAnet,up,j where j denotes the cell under consideration. Upstream blue water availability is defined as the blue water generated in the upstream cells minus the blue water footprint in the upstream cells (WFup). n

WAnet,up,j = ∑(WAup,i − WFup,i) i=1 where the subscript i denotes the cells upstream from the cell j under consideration. The blue water footprint is the volume of surface and ground water that is withdrawn and not returned back to the environment as liquid water (i.e., consumptive use). This value is calculated by adding the blue water footprint by Mekonnen and Hoekstra, (2012) (aggregated at a 0.5° resolution) — which accounts for agricultural, industrial and municipal water uses — with the amount of water required for well drilling and hydraulic fracking (WFfrac). Runoff data at 0.5° resolution were obtained from the Composite Runoff V1.0 database of Fekete et al. (2002).

To account for environmental flow requirements, we assumed that 80% of the natural runoff is allocated to environmental flows and the remaining 20% is considered blue water available for human needs (Aldaya, 2012; Richter, 2012). To calculate the upstream water availability we used the flow direction raster (0.5° resolution) from the World Water Development Report II (http://www.grdc.sr.unh.edu/index.html) (Vorosmarty, 2000a; 2000b). Assessing other water related impacts Water used for shale gas and oil extraction can be withdrawn either from surface water bodies, or the groundwater. The latter can be more ubiquitous and therefore more likely available close to the production wells (Nicot, 2012). To identify areas in which hydraulic fracturing is expected to place additional pressure on stressed groundwater resources, we overlaid a groundwater stress map (Gleeson et al., 2012) with the global distribution of high quality shale areas. To identify shale deposits in which the extraction of oil and gas is expected to strongly compete with food production in the near future, we examined areas in which the increase in agricultural production by closing the yield gap of major crops (i.e., wheat, maize and rice) to 75% of attainable yield requires an increase in irrigation. To that end, we rely on the global assessment of irrigation-controlled yield gaps by Muller et al., (2012). The number of people living in affected areas was estimated using population distribution data taken from CIESIN’s Gridded Population of the World map (GPWv4) for the year 2010. The economic evaluation of water use for energy and food production was calculated using data of water use and average world commodities prices in 2015 (World Bank, 2016) (Table 2).

SHALE DEPOSITS IN WATER STRESSED REGIONS The global analysis of surface water stress potentially generated by shale deposit exploitation (Fig. 12) shows that 39% of the area of high quality shale deposits are located in areas affected by surface water stress and where 171 million people live. Some of the deposits in stressed areas occur in the south- central United States, Argentina, South Africa, Saharan Africa, China, India, and Australia. Water stress is particularly high in areas that are highly populated, irrigated, or with low water availability (Mekonnen, 2016). In these regions an increase in human appropriation of freshwater resources for shale gas or shale oil extraction would markedly increase competition with the existing water uses for agriculture and environmental flows. Water stress in Australia, Argentina (in the Vaca Muerta shale deposit) and Algeria is particularly severe and worse than in any other shale deposit around the world, due to a combination of high volumes of extractable shale resources and lower water availability.

Figure 12. Map of surface water stress within high quality shale deposits. Pixels with water stress indexes greater than 1are subjected to unsustainable water withdrawals (i.e., water use for human activities exceeds the limit imposed by environmental flow requirements). Bottom panels show some notable water stressed high quality shale deposits in which production has either started or is expected to take place in the near future. Nowadays, the United States, Canada, China and Argentina have commercial shale gas production. Development of shale gas resources is also expected in Algeria and

Mexico. Shale oil production is practiced in the United States and Canada, future commercial extraction is mainly expected in Russia, Argentina, Colombia, Mexico and Australia (U.S. E.I.A., 2016 a-b).

The extraction of shale deposits is expected to affect not only surface water resources but also groundwater (e.g., Freyman, 2014). Our analysis (Fig. 13) shows that 8% of high quality shale deposits are located in regions affected by groundwater stress across the United States, Argentina, South Africa, China, India and home to 118 million people. About 6% of the world’s high quality shale areas are affected by both surface and groundwater stresses (Fig. 14). To cope with strong limitations on both surface and groundwater resources, one such area - southern Texas - developed a water market. However, the recent emergence of hydraulic fracturing in the region has driven higher water prices (Nicot 2012) and enhanced the competition between water for food production through irrigation and water for oil and natural gas production from shale resources. Thus while water markets may offer an effective solution for allocating water rights within water-limited systems (Debaere et al. 2014), the ultimate result may be the displacement of agriculture if shale energy companies are willing to pay a higher price for water use.

Figure 13. Groundwater stress over high quality shale deposits. Groundwater stress is defined as the ratio of the groundwater footprint and the aquifer area (Gleeson, 2012). Pixels with water stress indexes greater than 1are subjected to unsustainable water withdrawals (i.e., water use for human activities exceeds the limit imposed by environmental flow requirements).

Figure 14. Surface and groundwater stress over high quality shale deposits. High quality shale areas (countries) facing these combined stresses are: Mississippian Lime, Niobrara, Permian and Eagle Ford (U.S.), Eagle Ford deposit (Mexico), Vaca Muerta and Aguada Bandera (Argentina), Collingham (South Africa), Tannezuft and Frasnian (Algeria), Tannezuft (Lybia), Khatatba (Egypt), Etropole (Romania), Wufeng/Gaobiajian, Longmaxi, Keuter and Qingshankou (China), Sembar (Pakistan) and Sembar, Cambay and Permian Triassic (India).

Our analysis shows that water stress levels are not sensitive to different degrees of water recycling in the hydraulic fracturing process. With the assumption to recycle 50% and 100% of water there is no major change in the spatial extent of water stressed areas. A similar finding was also reported by other authors who noticed that the geographic distribution and size of water stressed areas was not sensitive to changes in crop water footprint (Brauman et al., 2016). Interestingly, we find water stress to substantially vary within each high quality shale area, consistent with the observation of small-scale heterogeneity in the global distribution of water stress (Vorosmarty et al., 2005 and Perveen et al., 2011). The results of our analysis are sensitive to the uncertainty in water footprint for hydraulic fracturing, agriculture, and water availability.

SHALE DEPOSITS OVER IRRIGATED AREAS To better evaluate possible future competition for water resources between shale deposit development and agriculture, we examined the global distribution of areas in which irrigation is expected to increase to accommodate growing demand for food products (Figure 15). We find that 25% of high quality shale areas worldwide are in irrigated areas where about 220 million people live. About 7% of the high quality shale areas are located in regions where water use for irrigation has been projected to increase in order to close the crop yield gap – the difference between actual and attainable yields. Thus, competition for water use in these areas will not only increase due to shale energy production but also be exacerbated by a greater need for irrigation water.

Figure 15. Irrigated areas overlying high quality shale deposits. Projected increase in irrigation to close maize, rice and wheat yield gaps to 75% of attainable yields is expected to affect several shale deposits. Bottom panels show twelve high quality shale areas - Bakken (Canada), Bakken, Heath, Gammon, Mississippian Lime, Barnett, Eagle Ford (U.S.), Eagle Ford (Mexico), Lower Silurian (Morocco), Sembar (Pakistan), Sembar and Cambay Shale (India), Keuter and Quingshankou (China), Nam Duk Fm (Thailand) - where we predict the occurrence of future competition between water for shale resource development and food production. Other high quality shale areas threatening water- energy-food security are: Collingham (South Africa), Etropole (Romania and Bulgaria), Hamitabat (Turkey), Barren Measure (India), Longmaxi and Qiongzhusi (China), Carynginia (Australia).

By far the most substantial human appropriation of water, consumptive freshwater use for irrigation is 899 Gm3 annually (Hoekstra, 2012). We estimated that a total volume of 58.7 Gm3 of water is required to extract the global shale oil and shale gas reserves using current technology. This water is expected to be withdrawn from the local surface and groundwater reserves within a timeframe of a few decades. Detailed volumes of water required for the development of each high quality shale area are shown in Table 1 of the supplementary materials. These estimates are based on current technology and estimated size of extractable hydrocarbon deposits. Many of the assessed high quality shale areas, however, are not likely to be put under commercial production for economic, social and technical reasons. The countries in which commercial production is expected to occur (U.S. EIA, 2016 a-b) (Fig. 12) account for 92% and 57% of world technically recoverable shale gas and oil, respectively. In these countries, the extraction of shale gas and oil will require a total volume of 36.07 Gm3 of water. While our results show that large volumes of water will be required, it is worth noting that future technological development and water management improvements will likely minimize fresh water appropriation using brackish water - a globally abundant and underutilized resource - and maximizing the reuse of returning hydraulic fracturing water (Mauter, 2014; Nicot, 2012).

Presently, in the United States only 11% of returning hydraulic fracturing fluid water is recycled (Chen, 2016), while 89% is injected into disposal wells. This water is therefore removed from the water cycle and becomes unavailable for other uses on an annual scale (Gregory, 2011; Warner, 2012; Mauter, 2014; Gregory, 2015). Though the volume of water for shale oil and gas production is an order of magnitude less than that required for crop irrigation globally, we find that the effect of hydraulic fracturing on water resources could be profound at local scales (Nicot, 2012) with potentially stronger impacts on other societal uses and the environment than irrigated agriculture.

Table 1. List of world high quality shale areas and total volume of water required to extract technically recoverable resources of shale gas and shale oil. Country Basin High quality shale Total potential Total potential area shale gas water shale oil water use (Gm3) use (Gm3) Canada Horn River Muskwa/Otter Park 0.6195 0.0000 Evie/Klua 0.2542 0.0000 Cordova Muskwa/Otter Park 0.1335 0.0000

Liard Lower Besa River 1.0414 0.0000 Deep Basin Doig Phosphate 0.1660 0.0000 Alberta Basin Banff/Exshaw 0.0000 0.0106 East and West Duvernay 0.7450 0.1345 Shale Basin Deep Basin North Nordegg 0.0876 0.0265 NW Alberta Area Muskwa 0.0000 0.0711 Southern Alberta Colorado Group 0.2825 0.0000 Basin Williston Basin Bakken 0.0000 0.0528 Appalachian Fold Utica 0.2049 0.0000 Belt Windsor Basin Horton Bluff 0.0224 0.0000 Mexico Burgos Eagle Ford Shale 2.2609 0.2128 Tithonian Shales 0.3325 0.0000 Sabinas Eagle Ford Shale 0.6609 0.0000 Tithonian La Casita 0.1557 0.0000 Tampico Pimienta 0.1532 0.1852 Tuxpan Tamaulipas 0.0000 0.0170 Pimienta 0.0000 0.0155 Veracruz Maltrata 0.0228 0.0092 Australia Cooper Roseneath-Epsilon- 0.5887 0.0333 Murteree (Nappamerri) Roseneath-Epsilon- 0.0238 0.0148 Murteree (Patchawarra) Roseneath-Epsilon- 0.0000 0.0044 Murteree (Tenappera) Maryborough Goodwood/Cherwell 0.1264 0.0000 Mudstone

Perth Carynginia 0.1637 0.0000 Kockatea 0.0000 0.0181 Canning Goldwyer 1.5528 0.3273 Georgina L. Arthur Shale 0.0536 0.0039 (Dulcie Trough) L. Arthur Shale 0.0306 0.0289 (Toko Trough) Beetaloo M. Velkerri Shale 0.1460 0.0466 L. Kyalla Shale 0.0930 0.0446 Colombia Middle Magdalena La Luna/Tablazo 0.0000 0.1597 Valley Llanos Gacheta 0.0000 0.0211 Colombia/ Maracaibo Basin La Luna/Capacho 1.3313 0.4982 Venezuela Argentina Neuquen Los Molles 1.8164 0.1230 Vaca Muerta 2.0300 0.5450 San Jorge Basin Aguada Bandera 0.3353 0.0000 Pozo D-129 0.229 0.0168 Austral- L. Inoceramus- 0.8539 0.2203 Magallanes Basin Magnas Verdes Parana Basin Ponta Grossa 0.0211 0.0003 Brazil Parana Basin Ponta Grossa 0.5309 0.1439 Solimoes Basin Jandiatuba 0.4255 0.0095 Amazonas Basin Barreirinha 0.6591 0.0260 Paraguay Parana Basin Ponta Grossa 0.0541 0.0183 Uruguay Cordobes 0.0301 0.0191 Paraguay/ Chaco Basin Los Monos 0.6826 0.1261 Bolivia Chile Austral- Estratos con Favrella 0.3196 0.0788 Magallanes Basin

Poland Baltic Llandovery 0.6940 0.0413 Basin/Warsaw Trough Lublin Llandovery 0.0000 0.3022 Podlasie Llandovery 0.0663 0.0194 Fore Sudetic Carboniferous 0.1407 0.0000 Lithuania/ Baltic Basin Llandovery 0.0161 0.0485 Kaliningrad Russia West Siberian Bazhenov Central 0.0000 1.9442 Central West Siberian Bazhenov North 0.9300 0.5608 North Ukraine/ Carpathian L. Silurian 0.4781 0.0000 Romania Foreland Basin Ukraine Dniepr-Donets L. Carboniferous 0.5010 0.0384 Ukraine/ Moesian Platform L. Silurian 0.0637 0.0027 Romania Romania/ Etropole 0.0000 0.0133 Bulgaria UK N. UK Carboniferous Shale 0.1656 0.0000 Carboniferous Shale Region S. UK Jurassic Lias Shale 0.0000 0.0230 Shale Region Spain Cantabrian Jurassic 0.0000 0.0048 France Paris Basin Lias Shale 0.0000 0.0511 Permian- 0.8400 0.1067 Carboniferous Southeast Basin Lias Shale 0.0488 0.0000 Germany Lower Saxony Posidonia 0.1113 0.0177 Wealden 0.0000 0.0042

Netherlands West Netherlands Epen 0.0000 0.0790 Basin Geverik Member 0.0000 0.0106 Posidonia 0.0000 0.0089 Sweden Scandinavia Alum Shale - Sweden 0.0644 0.0000 Denmark Region Alum Shale - 0.2092 0.0000 Denmark Morocco (and Tindouf L. Silurian 0.1149 0.0079 Wetern Tadla L. Silurian 0.0202 0.0000 Sahara) Algeria Ghadames/Berkin Frasnian 0.6992 0.1311 e Tannezuft 1.1629 0.0159 Illizi Tannezuft 0.3675 0.0171 Mouydir Tannezuft 0.0627 0.0000 Ahnet Frasnian 0.0581 0.0000 Tannezuft 0.3372 0.0000 Timimoun Frasnian 0.6162 0.0000 Tannezuft 0.3898 0.0000 Reggane Frasnian 0.1070 0.0078 Tannezuft 0.6898 0.0107 Tindouf Tannezuft 0.1719 0.0023 Tunisia Ghadames Tannezuft 0.0699 0.0014 Frasnian 0.0799 0.0477 Libya Ghadames Tannezuft 0.2756 0.1745 Frasnian 0.0349 0.0434 Sirte Sirte/Rachmat Fms 0.0000 0.5452 Etel Fm 0.0000 0.0678 Murzuq Tannezuft 0.0000 0.0451 Egypt Shoushan/Matruh Khatatba 0.0000 0.0225 Abu Gharadig Khatatba 0.0000 0.0632 Alamein Khatatba 0.0000 0.0193 Natrun Khatatba 0.0000 0.0481

South Africa Karoo Basin Prince Albert 0.6354 0.0000 Whitehill 1.3940 0.0000 Collingham 0.5407 0.0000 China Sichuan Basin Qiongzhusi 0.8239 0.0000 Longmaxi 1.8899 0.0000 Permian 1.4151 0.0000 Yangtze Platform L. Cambrian 0.2984 0.0000 L. Silurian 0.0000 0.0000 Jianghan Basin Niutitang/Shuijintuo 0.0754 0.0000 Longmaxi 0.0434 0.0013 Qixia/Maokou 0.0644 0.0083 Greater Subei Mufushan 0.0478 0.0000 Wufeng/Gaobiajian 0.2373 0.0075 U. Permian 0.0127 0.0016 Tarim Basin L. Cambrian 0.2901 0.0000 L. Ordovician 0.6224 0.0000 M.-U. Ordovician 0.4047 0.0522 Ketuer 0.0000 0.2174 Junggar Basin Pingdiquan/Lucaogou 0.0000 0.1827 Triassic 0.0000 0.2251 Songliao Basin Qingshankou 0.0000 0.3848 Mongolia East Gobi Tsagaantsav 0.0000 0.0573 Tamtsag Tsagaantsav 0.0000 0.0573 Thailand Khorat Basin Nam Duk Fm 0.0359 0.0000 Indonesia C. Sumatra Brown Shale 0.0000 0.0931 S. Sumatra Talang Akar 0.0000 0.1372 Tarakan Naintupo 0.0000 0.1679 Meliat 0.0247 0.0000 Tabul 0.0000 0.0106 Kutei Balikpapan 0.0000 0.0226 Bintuni Aifam Group 0.1884 0.0000

India Cambay Basin Cambay Shale 0.1947 0.0910 Krishna-Godavari Permian-Triassic 0.3751 0.0203 Cauvery Basin Sattapadi- 0.0000 0.0076 Andimadam Damodar Valley Barren Measure 0.0000 0.0070 Pakistan Lower Indus Sembar 0.6647 0.1951 Ranikot 0.0289 0.1097 Turkey SE Anatolian Dadas 0.0000 0.1534 Thrace Hamitabat 0.0427 0.0032 Jordan Hamad Batra 0.0441 0.0000 Wadi Sirhan Batra 0.0008 0.0000 USA Michigan Antrim 0.1266 0.0000 Permian Avalon-Bone Spring 0.0593 0.0974 Williston Bakken 0.0798 0.7624 Ft. Worth Barnett 0.1154 0.0067 Marfa Barnett-Woodford 0.0098 0.0235 Permian Barnett-Woodford 0.0817 0.0000 Greater Green Baxter 0.0006 0.0302 River Greater Green Baxter/HIllard 0.0026 0.0201 River Powder River Baxter/HIllard 0.0158 0.0705 Appalachian Big Sandy 0.0817 0.0134 Black Warrior Chattanooga 0.0105 0.0000 Appalachian Chattanooga 0.0290 0.0000 Appalachian Cleveland 0.0065 0.0033 Permian Cline 0.0718 0.0302 Valley and Ridge Conasauga 0.0283 0.0000 (APB) Appalachian Devonian (Ohio) 0.1563 0.0000 Burgos Eagle Ford 0.3654 0.3459

Western Gulf Eagle Ford Gas Play 0.1688 0.1679 Western Gulf Eagle Ford Oil Play 0.1965 0.1780 Cherokee Excello-Mulky 0.0098 0.0235 Arkoma Fayetteville 0.17611 0.0000 Black Warrior Floyd-Chattanooga 0.0105 0.0000 Black Warrior Floyd-Neal 0.0211 0.0201 Williston Gammon 0.0000 0.1108 TX-LA-MS Salt Haynesville-Bossier 1.4023 0.0268 (APB) Williston Heath 0.0481 0.0033 Pardox Hermosa 0.0065 0.0268 San Juan Lewis 0.1352 0.0000 Montana Thrust Lombard 0.0006 0.0000 Belt Uinta Mancos 0.2796 0.0235 Uinta-Piceance Mancos 0.0639 0.0067 Uinta Manning Canyon 0.0072 0.0033 Appalachian Marcellus 0.9808 0.0100 Anadarko Mississippian Lime 0.2711 0.0302 Los Angeles Monterey 0.01978 0.0201 Santa Maria Monterey 0.0804 0.0000 Powder River Mowry 0.0699 0.0302 Illinois New Albany 0.1919 0.0000 North Park Niobrara 0.0178 0.0134 Denver Niobrara 0.0758 0.0033 Uinta-Piceance Niobrara 0.2658 0.0302 Maverick Pearsall-Eagle Ford 0.0323 0.0000 Raton Pierre - Niobrara 0.0039 0.0134 TX-LA-MS Salt Tuscaloosa 0.0936 0.2351 Appalachian Utica 0.3601 0.0302 Permian Wolfcamp/Wolfbone 0.1642 0.2048

Ardmore Woodford 0.0639 0.0201 Anadarko Woodford 0.1517 0.0302 Arkoma Woodford-Caney 0.1517 0.0000

WATER FOR FRACKING OR WATER FOR IRRIGATION? The increasing food and energy needs of humanity (e.g., Suweis, 2013) and the possible local decline in water availability as an effect of climate change (IPCC, 2013) are expected to increase pressure on freshwater resources (Muller, 2012; Davis, 2014). As a result, regions affected by increasing water stress will face not only environmental and social challenges, but also see the emergence of financial obstacles both for the food and energy industries. In some water constrained areas where shale resources are present a rush for water appropriation by oil and gas companies has surged, leaving the agricultural sector with limited water supply (Nicot, 2012; Mauter, 2014). This pattern is expected to occur in many other water stressed agricultural regions in which shale deposits are going to be developed because of the higher profits of water use for energy than for food production. In fact, despite the current low price of oil and natural gas, the use of water for energy production generates greater profits than agriculture (Table 2). Oil production from hydraulic fracturing is also less water intensive than oil from oil sands and conventional oil through secondary recovery (i.e., water flooding of the reservoir) and therefore more economically convenient. Only conventional oil from primary recovery (i.e., using lift pump) and conventional gas, which has a zero water footprint in extraction, exhibits a higher economic yield of water use than shale oil and shale gas. Interestingly, bioethanol and biodiesel are less profitable - in terms of economic yield of water - than fossil oil and gas, but are more economical than the production of certain food crops. Moreover, our analysis shows that, despite their similar water requirements per unit of energy produced, shale oil and shale gas strongly differ in the economic yields of the water used in their production processes. Because of the higher price of shale oil with respect to shale gas, the economic yield of water used in shale oil extraction is higher than in the case of shale gas. Oil and gas prices are currently driving investors to target the more lucrative high quality shale areas where the more profitable oil and natural gas liquids (i.e., ethane, propane, butane and isobutene) are known to exist (U.S. DOE, 2013).

Table 2. Economic yield of water for energy and food production. 1 bbl (barrel) = 0.159 m3 Water footprint of Water per unit Average global price Economic major food and fuel of energy (year 2015) ¶¶ yield of water products (L/GJ) ($/m3) Shale gas 0.20-0.55 L/m3 * 5.69-15.54 2.47-6.88 $/GJ 159-434 Shale oil 57.08-65.86 L/bbl * 9.85-11.37 50.75 $/bbl 771-889 Conventional gas 0 ** 0 2.47-6.88 $/GJ - Conventional oil: 31 L/bbl ** 5.31 50.75 $/bbl 1651 primary recovery Conventional oil: 1361 L/bbl ** 235 50.75 $/bbl 37 secondary recovery Oil from oil sands 95-906 L/bbl † 16.15-154.05 50.75 $/bbl 56-532 Maize 1222 L/kg ¶ 800157 0.169 $/kg 0.14 Rice 1830 L/kg ¶ 336459 0.351-0.386 $/kg 0.14 Wheat 2497 L/kg ¶ 176043 0.205 $/kg 0.11 Bioethanol 2107-2854 L/L ¶ 51000-121000 0.38 $/L 0.19-0.32 Biodiesel 11400 L/L ¶ 343000 0.76 $/L 0.07 * Scanlon, 2014; ** Mielke, 2010; † this work; ¶ Mekonnen, 2011; ¶¶ The World Bank, 2016

BALANCING ENERGY AND FOOD SECURITY A number of shale deposits are situated in countries lacking conventional hydrocarbon deposits such as South Africa, Jordan, Chile, and European countries (Kuustra, 2013). Their potential future exploitation will create a new geography of oil and natural gas production, with important implications for the global geopolitical landscape (IEA, 2016 a-b). Shale resources are an opportunity for these countries to increase their energy security, while reducing costs of fossil fuel imports and potentially changing their import-export balance (Vidic, 2013; Mauter, 2014). Should they choose to exploit their shale deposits, these countries will need to develop responsible water management plans to ensure that other sectors are not impacted.

Nowadays, the United States, Canada, China and Argentina have commercial shale gas production. Development of shale gas resources is also expected in Algeria and Mexico. Shale oil production is practiced in the United States and Canada, future commercial extraction is mainly expected in Russia,

Argentina, Colombia, Mexico and Australia (U.S. E.I.A., 2016 a-b). Shale oil production is expected to more than double in the near future from 4% of the world’s oil demand in 2015 to about 10% by 2040 (U.S. EIA, 2016 a-b; British Petroleum, 2016; Exxon Mobil, 2016); likewise, shale gas production is predicted to surge from 10% of the global natural gas supply in 2014 to 30% by 2040 (U.S. EIA, 2016 a-b; British Petroleum, 2016; Exxon Mobil, 2016). The shale revolution has created new jobs and economic benefits in North America, supporting economic growth also in some rural and less developed areas. For example, in the U.S. the shale industry employed about 1.7 million people in 2015, a trend that is expected to grow to 3.5 million jobs by 2035 (U.S Chamber of commerce, 2013). Thus, with adequate policies and regulations, shale extraction has the potential to enhance the economic growth and energy security of some regions. Despite these benefits, in many regions of the world shale deposits development will be problematic because of water limitations and will likely exacerbate a competition with water for food and other human needs. Particularly critical appears to be the case of some high quality shale areas in water stressed regions of the United States, Mexico, South Africa, China, South Asia, and Australia (Figure 15). In some of these regions oil and gas production from shale rocks is expected to threaten the local water and food security. In these water stressed areas where shale resources are present adequate policies need to be put in place in order to avert social, economic, and ecological consequences.

REFERENCES Aldaya, Maite M., et al. The water footprint assessment manual: Setting the global standard. Routledge, 2012. Al-Jubori, Ahmed, et al. "Coalbed methane: clean energy for the world." Oilfield Review 21.2 (2009): 4-13.

Annual energy outlook 2016 (a U.S. Energy Information Administration, 2016); https://www.eia.gov/forecasts/aeo/index.cfm Ayensu, Edward, et al. "International ecosystem assessment." Science286.5440 (1999): 685-686. Bazilian, Morgan, et al. "Considering the energy, water and food nexus: Towards an integrated modelling approach." Energy Policy 39.12 (2011): 7896-7906. Biggs, Eloise M., et al. "Sustainable development and the water–energy–food nexus: A perspective on livelihoods." Environmental Science & Policy 54 (2015): 389-397. Brauman, Kate A., et al. "Water depletion: An improved metric for incorporating seasonal and dry-year water scarcity into water risk assessments." Elementa: Science of the Anthropocene 4.1 (2016): 000083.

Brittingham, Margaret C., et al. "Ecological risks of shale oil and gas development to wildlife, aquatic resources and their habitats." Environmental Science & Technology 48.19 (2014): 11034- 11047.

Bunch, A. G., et al. "Evaluation of impact of shale gas operations in the Barnett Shale region on volatile organic compounds in air and potential human health risks." Science of the Total Environment 468 (2014): 832-842. Center for International Earth Science Information Network - CIESIN – Columbia University. 2015. Gridded Population of the World, Version 4 (GPWv4): Population Count. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). http://dx.doi.org/10.7927/H4BG2KXS. Accessed May 2016. Change, IPCC Climate. "The physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change." K., Tignor, M., Allen, SK, Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, PM, Eds (2013): 1535.

Charpentier, Alex D., Joule A. Bergerson, and Heather L. MacLean. "Understanding the Canadian oil sands industry’s greenhouse gas emissions. "Environmental research letters 4.1 (2009): 014005.

Chen, Huan, and Kimberly E. Carter. "Water usage for natural gas production through hydraulic fracturing in the United States from 2008 to 2014." Journal of environmental management 170 (2016): 152-159. Chew, Kenneth J. "The future of oil: unconventional fossil fuels." Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences 372.2006 (2014): 20120324.

Chopra, S., Sharma, R. K., Keay, J., & Marfurt, K. J. (2012, November). Shale gas reservoir characterization workflows. In 2012 SEG Annual Meeting. Society of Exploration Geophysicists. Clark, Corrie E., Robert M. Horner, and Christopher B. Harto. "Life cycle water consumption for shale gas and conventional natural gas." Environmental science & technology 47.20 (2013): 11829- 11836.

Commodity price data: updated August, 2016 (The Wolrd Bank, 2016); http://www.worldbank.org/en/research/commodity-markets

Davis, Kyle F., Paolo D'Odorico, and Maria Cristina Rulli. "Moderating diets to feed the future." Earth's Future 2.10 (2014): 559-565.

Debaere, Peter, et al. "Water markets as a response to scarcity." Water Policy16.4 (2014): 625- 649.

Denney, Dennis, ed., 2009, Health, safety, and environment— Evaluating implications of hydraulic fracturing in shale-gas reservoirs: Journal of Petroleum Technology, v. 61, no.8. Accessed on July 10, 2012 at http://www.spe.org/jpt/print/ archives/2009/08

DOE, US. "Modern shale gas development in the United States: An Update."Office of Fossil Energy and National Energy Technology Laboratory, United States Department of Energy (2013).

DOE, US. "Modern shale gas development in the United States: An Update."Office of Fossil Energy and National Energy Technology Laboratory, United States Department of Energy (2013).

Drohan, P. J., et al. "Early trends in landcover change and forest fragmentation due to shale-gas development in Pennsylvania: a potential outcome for the Northcentral Appalachians." Environmental management 49.5 (2012): 1061-1075.

Energy, U. N. "The energy challenge for achieving the Millennium Development Goals." New York (2005).

Fekete, Balázs M., Charles J. Vörösmarty, and Wolfgang Grabs. "High‐resolution fields of global runoff combining observed river discharge and simulated water balances." Global Biogeochemical Cycles 16.3 (2002). Fiksel, Joseph. "Sustainability and resilience: toward a systems approach."Sustainability: Science, Practice, & Policy 2.2 (2006). Freyman, Monika. "Hydraulic fracturing & water stress: Water demand by the numbers." Ceres (2014): 49-50.

Gale, J. F., Reed, R. M., & Holder, J. (2007). Natural fractures in the Barnett Shale and their importance for hydraulic fracture treatments. AAPG bulletin, 91(4), 603-622. Gallegos, Tanya J., et al. "Hydraulic fracturing water use variability in the United States and potential environmental implications." Water resources research 51.7 (2015): 5839-5845.

Gleeson, Tom, et al. "Water balance of global aquifers revealed by groundwater footprint." Nature 488.7410 (2012): 197-200. Godfray, H. Charles J., et al. "Food security: the challenge of feeding 9 billion people." science 327.5967 (2010): 812-818.

Goodwin, Stephen, et al. "Water intensity assessment of shale gas resources in the Wattenberg field in northeastern Colorado." Environmental science & technology 48.10 (2014): 5991-5995.

Gordon, Deborah. "Understanding unconventional oil." (2012).

Gregory, Kelvin B., Radisav D. Vidic, and David A. Dzombak. "Water management challenges associated with the production of shale gas by hydraulic fracturing." Elements 7.3 (2011): 181-186.

Haddeland, Ingjerd, et al. "Global water resources affected by human interventions and climate change." Proceedings of the National Academy of Sciences 111.9 (2014): 3251-3256.

Hoekstra, Arjen Y. "Water scarcity challenges to business." Nature climate change 4 (2014): 318- 320.

Hoekstra, Arjen Y., and Thomas O. Wiedmann. "Humanity’s unsustainable environmental footprint." Science 344.6188 (2014): 1114-1117.

Holditch, Stephen A., K. Perry, and J. Lee. "Unconventional Gas Reservoirs–Tight Gas, Coal Seams, and Shales." Working Document of the National Petroleum Council on Global Oil and Gas Study (2007).

Howells, Mark, et al. "Integrated analysis of climate change, land-use, energy and water strategies." Nature Climate Change 3.7 (2013): 621-626.

Howells, Mark, et al. "Integrated analysis of climate change, land-use, energy and water strategies." Nature Climate Change 3.7 (2013): 621-626.

Howells, Mark, et al. "Integrated analysis of climate change, land-use, energy and water strategies." Nature Climate Change 3.7 (2013): 621-626. Humphries, Marc. "North American oil sands: history of development, prospects for the future." LIBRARY OF CONGRESS WASHINGTON DC CONGRESSIONAL RESEARCH SERVICE, 2008.

International Energy Outlook 2016 (b U.S. Energy Information Administration, 2016); http://www.eia.gov/forecasts/ieo/ Jiang, Mohan, Chris T. Hendrickson, and Jeanne M. VanBriesen. "Life cycle water consumption and wastewater generation impacts of a Marcellus shale gas well." Environmental science & technology 48.3 (2014): 1911-1920. Kargbo DM, Wilhelm RG, Campbell DJ (2010) Natural gas plays in the Marcellus Shale: Challenges and potential opportunities. Environmental Science & Technology 44:5679-5684 Kelly, Erin N., et al. "Oil sands development contributes elements toxic at low concentrations to the Athabasca River and its tributaries." Proceedings of the National Academy of Sciences 107.37 (2010): 16178-16183.

Kiviat, Erik. "Risks to biodiversity from hydraulic fracturing for natural gas in the Marcellus and Utica shales." Annals of the New York Academy of Sciences 1286.1 (2013): 1-14.

Kuuskraa, Vello, Scott H. Stevens, and Keith D. Moodhe. Technically recoverable shale oil and shale gas resources: an assessment of 137 shale formations in 41 countries outside the United States. 2013. Mauter, Meagan S., et al. "Regional variation in water-related impacts of shale gas development and implications for emerging international plays."Environmental science & technology 48.15 (2014): 8298-8306.

McKenzie, Lisa M., et al. "Human health risk assessment of air emissions from development of unconventional natural gas resources." Science of the Total Environment 424 (2012): 79-87. Mekonnen, Mesfin M., and Arjen Y. Hoekstra. "A global assessment of the water footprint of farm animal products." Ecosystems 15.3 (2012): 401-415. Mekonnen, Mesfin M., and Arjen Y. Hoekstra. "Four billion people facing severe water scarcity." Science advances 2.2 (2016): e1500323. Mekonnen, Mesfin M., and Arjen Y. Hoekstra. "The green, blue and grey water footprint of crops and derived crop products." Hydrology and Earth System Sciences 15.5 (2011): 1577-1600.

Mielke, Erik, Laura Diaz Anadon, and Venkatesh Narayanamurti. "Water consumption of energy resource extraction, processing, and conversion."Belfer Center for Science and International Affairs (2010).

Mobil, Exxon. "The Outlook for Energy: A view to 2040." ExxonMobil: Irving, TX, USA (2016). Moniz, Ernest J., et al. "The future of natural gas." Cambridge, MA: Massachusetts Institute of Technology (2011).

Mossop, Grant D. "Geology of the Athabasca oil sands." Science 207.4427 (1980): 145-152.

Mueller, Nathaniel D., et al. "Closing yield gaps through nutrient and water management." Nature 490.7419 (2012): 254-257. Nicot, Jean-Philippe, and Bridget R. Scanlon. "Water use for shale-gas production in Texas, US." Environmental science & technology 46.6 (2012): 3580-3586. Oil and Gas Supply Module of the National Energy Modeling System: Model Documentation 2014 (Energy Information Administration, 2014). Oil and Gas Supply Module of the National Energy Modeling System: Model Documentation 2014. Accessed July, 2016.

Oil sands mining operators- Water allocation and use by year (Alberta Environment and Parks, 2015); http://osip.alberta.ca/library/Dataset/Details/51. Accessed March, 2016.

Osborn, Stephen G., et al. "Methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing." proceedings of the National Academy of Sciences 108.20 (2011): 8172-8176.

Pacala, Stephen, and Robert Socolow. "Stabilization wedges: solving the climate problem for the next 50 years with current technologies." science305.5686 (2004): 968-972.

Perveen, Shama, and L. Allan James. "Scale invariance of water stress and scarcity indicators: Facilitating cross-scale comparisons of water resources vulnerability." Applied Geography 31.1 (2011): 321-328.

Petroleum, British. "BP energy outlook 2035." London: British Petroleum (2016). Reig, P., T. Luo, and J. N. Proctor. "Global shale gas development: water availability and business risks. Washington (DC): World Resources Institute." (2014).

Resources to reserves (IEA, International Energy Agency, 2013); https://www.iea.org/publications/freepublications/publication/Resources2013.pdf

Richter, Brian D., et al. "A presumptive standard for environmental flow protection." River Research and Applications 28.8 (2012): 1312-1321. Ridley, M. (2011). The shale gas shock. Global Warming Policy Foundation.

Rogner, Hans-Holger. "An assessment of world hydrocarbon resources."Annual review of energy and the environment 22.1 (1997): 217-262.

Rulli, Maria Cristina, Antonio Saviori, and Paolo D’Odorico. "Global land and water grabbing." Proceedings of the National Academy of Sciences 110.3 (2013): 892-897.

Rulli, Maria Cristina, et al. "The water-land-food nexus of first-generation biofuels." Scientific reports 6 (2016).

Rutqvist, Jonny, et al. "Modeling of fault reactivation and induced seismicity during hydraulic fracturing of shale-gas reservoirs." Journal of Petroleum Science and Engineering 107 (2013): 31-44.

Scanlon, Bridget R., Robert C. Reedy, and J-P. Nicot. "Comparison of water use for hydraulic fracturing for unconventional oil and gas versus conventional oil." Environmental science & technology 48.20 (2014): 12386-12393.

Schneider, Richard, and Simon Dyer. Death by a thousand cuts: impacts of in situ oil sands development on Alberta's boreal forest. Calgary, AB, Canada: Pembina Institute, 2006.

Seventeen Sustainable Goals (United Nations, 2016). http://www.un.org/sustainabledevelopment/sustainable-development-goals/

Souther, Sara, et al. "Biotic impacts of energy development from shale: research priorities and knowledge gaps." Frontiers in Ecology and the Environment 12.6 (2014): 330-338.

Speight, J. G. (2013). Shale gas production processes. Elsevier, Wyoming, USA

Stamford, Laurence, and Adisa Azapagic. "Life cycle environmental impacts of UK shale gas." Applied energy 134 (2014): 506-518.

Stokstad, Erik. "Will fracking put too much fizz in your water?." Science344.6191 (2014): 1468- 1471.

Suweis, Samir, et al. "Water-controlled wealth of nations." Proceedings of the National Academy of Sciences 110.11 (2013): 4230-4233.

T., Gregory P., et al. "Energy consumption in desalinating produced water from shale oil and gas extraction." Desalination 366 (2015): 94-112. U.S. Chamber’s fracking job boom: Behind the numbers (U.S Chamber of commerce, 2016); http://www.energyxxi.org/us-chamber%E2%80%99s-fracking-job-boom-behind-numbers

The Water–Energy–Food Nexus: A New Approach in Support of Food Security and Sustainable Agriculture Food and Agriculture Organisation of the United Nations, Rome (2014) Unconventional Resources Atlas (U.S. National Energy Technology Laboratory, 2016). http://www.unconventionalenergyresources.com/. Accessed July, 2016 Vandecasteele, Ine, et al. "Impact of shale gas development on water resources: a case study in Northern Poland." Environmental management 55.6 (2015): 1285-1299. Vidic, Radisav D., et al. "Impact of shale gas development on regional water quality." Science 340.6134 (2013): 1235009.

Vörösmarty, C. J., et al. "Geomorphometric attributes of the global system of rivers at 30-minute spatial resolution." Journal of Hydrology 237.1 (2000): 17-39. Vörösmarty, C. J., et al. "Global system of rivers: Its role in organizing continental land mass and defining land‐to‐ocean linkages." Global Biogeochemical Cycles 14.2 (2000): 599-621. Vörösmarty, Charles J., et al. "Geospatial indicators of emerging water stress: an application to Africa." AMBIO: A journal of the Human Environment 34.3 (2005): 230-236.

Warner, Nathaniel R., et al. "Geochemical evidence for possible natural migration of Marcellus Formation brine to shallow aquifers in Pennsylvania." Proceedings of the National Academy of Sciences 109.30 (2012): 11961-11966.

World Economic Forum, Global Risks 2015, 10th Edition (World Economic Forum, Geneva, Switzerland, 2015).

World Energy Outlook 2012 (IEA, International Energy Agency, 2012) Paris, France.

World Water Assessment Programme (2012), The United Nations World Water Development Report 4: Managing Water under Uncertainty and Risk, United Nations Educational Scientific and Cultural Organization, Paris.

Page intentionally left blank

PART 2: OIL SANDS

Also known as tar sands, oil sand deposits differ from conventional oil fields in two ways. First, oil sands are orders of magnitude larger respect to conventional oil pools. Second, oil sands have different physical characteristics from conventional crude oil (Mossop, 1980). Oil sands are a natural occurring mixture of sand, clay, water, other minerals and bitumen (Dai, 1996). Oil sands contain a highly viscous crude oil in an unconsolidated sand matrix.

Bitumen is a heavy oil with API gravity − a measure of how heavy a petroleum is compared to water (U.S. EIA, 2016) − less than 10° (Meyer, 2007). Bitumen is too heavy and viscous to flow or be pumped in natural conditions, but undergoes a substantial reduction in viscosity when subjected to elevated temperatures (Dai, 1996). Bitumen must be treated before it can be used by refineries to produce usable fuels such as gasoline and diesel. Bitumen chemical composition has a low hydrogen to carbon ratio and is composed by carbon, hydrogen, oxygen, nitrogen, sulfur1 (Alberta Energy, 2016).

Figure 1. Composition of oil sands.

1 Chemical composition of bitumen varies among different deposits.

Oil sand deposits have been discovered in a few region around the world and account for about 30% of the proved reserves of global crude oil. Because of their oil sand deposits, Venezuela and Canada have the first and third largest recoverable oil reserves in the world, respectively, while Saudi Arabia is at the second place with its endowment of conventional oil (Figure 2) (U.S. Central Intelligence Agency, 2016). Presently, oil sand deposits are intensively exploited only in Alberta and Venezuela (in the “Orinoco’s Heavy Oil Belt”, which encompasses both oil sand and heavy oil deposits). In Madagascar and Utah commercial extraction has just started, while the development of other deposits around the world is still at the exploration or planning stage.

Crude Oil proved reserves 350 300 250 200 150

100 BillionBarrels 50 0

Figure 11. World crude oil proved reserves. Source: U.S. Central Intelligence Agency, 2016.

The Athabasca deposit in Alberta is the most developed in the world. In Alberta, best available technologies are used in the exploration and production processes. Although oil sands are found in many countries, the availability of water in Alberta, make water-based bitumen extraction feasible. Furthermore, Alberta’s oil sands are considered unique since are water-wet, meaning that a layer of water cover the sand particles and bitumen. This water facilitates the separation of bitumen from sand (Oil sands magazine, 2015). While, oil sands in Venezuela and USA (Utah) are oil-wet and therefore cannot be separated in a water-based process (Oil sands magazine, 2015). Deposits in Venezuela and Utah require solvents and chemicals to separate bitumen from the sand.

Figure 12. Water-wet and oil-wet oil sands. Source: U.S. Oil Sands, 2015 Historically, oil sands was incorrectly referred to as tar sands. The name tar derives from the now outdated and largely ineffective practice of using oil sands for roofing and paving tar. Oil sands and tar appear to be visibly similar, but they are different.

 Oil sands are a naturally occurring petrochemical that can be upgraded into crude oil and other petroleum products.  Tar is synthetically produced from coal, wood, petroleum or peat through destructive distillation. Tar is generally used to seal against moisture.

Furthermore, oil sands can be refined to make oil, while tar cannot. Tar has historically been used to seal wood and rope against moisture (Alberta Energy, 2016).

Figure 4. Tar (left) and oil sands (right). Source: Tar sands basic.

CANADA Canada has about 173 billion barrels of oil that can be recovered with today’s technology (Canada Association of Petroleum Producers, 2016). Of that number, the 97 per cent are located in the oil sands (Canada Association of Petroleum Producers, 2016). However, with new technologies this reserve estimates could be significantly increased. In fact, Canada’s total oil sands reserves are estimated at 1.8 trillion barrels (Alberta Oil sands industry, 2016). Today more than a half of Canada’s oil production comes from the oil sands. Canada is the largest supplier of oil to USA (Alberta’s oil sands, 2016).

United States

Figure 5. Location and extensions of Alberta’s oil sand deposits (red). Oil sands are located in three main regions within the province of Alberta: the Athabasca, Cold Lake and Peace River regions (Alberta Energy, 2016).

Formation of oil sands

Alberta’s oil sands were formed in Early Cretaceous age, about 110 millions of years ago when the province was covered by a warm tropical sea (Government of Alberta, 2016). Oil began forming in southern Alberta when tiny marine creatures died and were drifted to the seafloor. Over time their body were compressed by heat and pressure and formed liquid rock oil.

Meanwhile, in the north, rivers flowing away from the sea deposited sand and sediment. When tectonic plates shifted to form the Rocky Mountains, the pressure squeezed the oil northward causing it to seep into the sand and transporting with it oxygenated water (Alberta Energy, 2016). Aerobic microbe bacteria fed off the lighter hydrocarbon molecules in a biodegradation process. Heavy and complex hydrocarbons were left behind to form Alberta’s oil sands (Oil sands magazine, 2015).

History of oil sands Bitumen appears in the first writings about the Athabasca region that immediately identifying it as a resource. The primary use was not related to energy, but was used to waterproof canoes mixing bitumen and spruce gum.

Efforts to tap the oil sands resources began in the early 20th century. The production of oil from oil sands has been conducted commercially for almost five decades. Canada’s first large scale oil sands mine started in 1967. Initially oil sands were primarily accessed through large open pit mining operation. While oil sands mining is on the commercial scale since the 1980s, in situ drilling is a more recent technology. In situ technologies have played a growing role in oil sands production (Alberta Energy, 2016). Since 2012 in situ oil sands production exceed mined oil sands production in Alberta (Alberta Oil sands industry 2016).

Canada oil sands production 900 800 700 600 500 400 300

200 Production [million barrels] 100 0

Mining bitumen Thermal in situ bitumen

Figure 6. Bitumen production from Canada’s oil sands. Source: Alberta Environment and Parks, 2016.

RECOVERING THE OIL Oil sands are recovered using two main methods: surface mining and in situ drilling. The method used depends on how deep the reserves are located. Bitumen that is close to the surface is mined. While, bitumen that is deep within the ground is extracted in situ using specialized techniques. Approximately 80 per cent of Alberta’s oil sands are recoverable through in situ production, with only 20 per cent recoverable using mine extraction (Alberta Energy, 2016).

SURFACE MINING Oil sands mines are one of the largest earth moving operation in the world (Oil sands magazine, 2015). The principal requirements for a viable mine are that there is a sufficient reserves to support a plant life for at least 20 years and that there are oil sands with more than 8% bitumen by weight (Mossop, 1980).

Surface mining is economically feasible when oil sands are close to the surface. Surface mining collects deposits less than 75 m deep. It is practiced in the Athabasca region, near Fort McMurray, adjacent to the Athabasca River (Oil sands magazine, 2015).

Figure 7. Oil sands mineable and in situ drilling areas. Source: Peak oil, 2016. A typical oil sands mining facility has the following units (Oil sands magazine, 2015):

 A surface mine;  A bitumen production plant, where bitumen is separated from sand and water;  A tailings storage facility, known as tailings pond;  An utility plant, which supplies steam, power and hot water;  A storage facility for bitumen and diluents.

Anatomy of an oil sands deposit A layer of topsoil, containing trees, shrubs normally covers the deposit. An oil sands deposit consist of four distinct layers (Oil sands magazine, 2015):

 Muskeg: A wet top layer of peatland, muskeg has a limited use as construction material. Once removed is typically sent to a waste dump.  Overburden: a layer that consists of glacial drift composed by sand, gravel and a little amount of bitumen. Sections of deposit containing less than 7% bitumen are considered overburden as defined by Alberta Energy Regulator. Overburden has a high sand content and is used as building material. The overburden average thickness is 15 meters.

 Oil sands: a zone of bitumen rich sand. Oil sands deposit are an unconsolidated deposit of mainly sand, bitumen and water. The oil sands layer is typically 50 meter in thickness.  Rock: a layer of rock that sits below the oil sands.

Figure 8. Anatomy of an oil sands deposit. Source: Oil sands magazine, 2015.

 Clearing of the forest: black spruce2 is cleared during winter period, when frost penetration into the underlying muskeg is sufficient to allow the movement of heavy trucks. The trees are harvested and sold to forestry companies.  Remove the muskeg layer: the muskeg layer is a peatland that needs hydraulic remediation. Drainage trenches are excavated and after muskeg is dewatered, it is removed and it is stored in a waste dump for subsequent use in future reclamation of the mines.  Dewater the mining area: some areas can be very wet. Water is collected and pumped to a water collection pond in order to lower the groundwater table.  Remove the overburden: overburden is removed and is used on site to build dikes for the containment of tailings3 and muskeg or is used as building material.  Excavate the opening cut: the opening cut is the very first excavation of the oil sands deposit. The techniques uses shovels and trucks4. Mining shovels dig into sand and load it into trucks. The mined material is transferred to the Bitumen Production processing plant. It is very common for

2 Black spruce, Picea Mariana, is a North American species of spruce tree in the pine family. It is frequent part of the biome known as taiga or boreal forest. 3 Tailings are a combination of water, sand, silt and clay that are a by-product of the bitumen extraction from the oil sands. 4 The oil sands mining truck is the Caterpillar 797. This truck is one of the world’s largest trucks with the capacity to haul up to 400 tons per load.

new oil sand mines to have a below average performance for the first few months of operation due to the low grade of bitumen in the oil sands.

Figure 9. Overview of an oil sands mines during the opening cut phase. Source: Oil sands magazine, 2015.

 Mine out the oil sands following the mine plan: the sequence and sections of mine to be excavated are outlined in a mine plan provided by geologists in order to ensure constant bitumen input and optimal performances of the plant. Mining continues until the pit is depleted or become uneconomical.  Backfill the mined out pit: once the pit is mined out, the empty pit is backfilled with tailings, overburden and waste material that was excavated out of the mine.  Reclaim the mine site: once the mine is backfilled, it needs by law to be reclaimed. Reclamation is completed when the area is capable of sustaining wildlife or vegetation.

Figure 10. Former oil sands mine backfilled and reclaimed. Source: Oil sands magazine, 2015.

Ore Preparation Plant Oil sands is haul from the mine to the Ore Preparation Plant (OPP). OPP is the first step in any Bitumen Production facility. Here mined oil sands ore are crushed and mixed with hot water to form a slurry that is pumped to the processing plant. A correct functionality of the OPP is of fundamental importance for bitumen extraction (Oil sands magazine, 2015):

 Large debris are removed in the OPP stage. These materials can damage downstream equipment.  The oil sands are crushed to make more effective the blending with water.  The mixing of hot water with oil sands ore allows the slurry mixture to aerate and entrain air bubbles. These air bubbles then attach on the surface of the liberate bitumen particles allowing the separation by gravity with sand. In addition to hot water, it is also used caustic soda in order to increase the PH of the slurry facilitating the following separation process.

Extraction The slurry is transported through pipelines to the Bitumen Extraction facility. Here, primary separation occurs. The slurries goes through a separation frothing process. Injection air forms tiny bubbles that separate bitumen from sand. Sand and water are settled at the bottom, while tiny air bubbles trapped in the bitumen cause it to form into froth and rise to the surface, where it is skimmed out and diluted.

Figure 11. Gravity separation of oil sands slurry. Source: Oil sands magazine, 2015.

The objectives of the Bitumen Extraction facility are (Oil sands magazine, 2015):

 Maximize the recovery of bitumen;  Produce a good quality bitumen, minimizing fines and water content;  Sent as much of the solids as possible to the tailings plant.

The froth separated has a lot of impurities. The composition is only 50 to 60% bitumen with water up to 30% and 10% of fine solids (Oil sands magazine, 2015). This low quality froth is further processed. Water and fine solids are removed using a solvent. The solvent used is a light hydrocarbon that reduces the viscosity of the bitumen and enables a gravity separation of the various phases in the froth. The final product is a good quality bitumen that can be sent to an upgrader for conversion to Synthetic Crude Oil and then sent to refineries to obtain marketable products.

Bitumen is too viscous to flow through pipelines. Pipeline users dilute bitumen with a hydrocarbon solvent (usually refinery naphtha) before it could be transported. Once mixed with a hydrocarbon diluent is sent to upgrading (Oil sands magazine, 2015).

Tailings ponds Tailing ponds or tailing storage facilities are made up of:

 water and sand from the Bitumen Extraction facilities;  residual hydrocarbons and fine clays from froth separation;  unrecovered bitumen;

Tailings streams are normally about 50% water (Oil sands magazine, 2015). Tailings are the leftover liquid mixture of mostly water and clay, some sand and residual oil.

Tailings ponds are large engineered dam and dyke systems designed for store tailings. These large volume of water cannot be readily returned to the environment because of suspended clay particles, heavy metals and other contaminants. While sand can be easily separated by gravity and recycled for backfill the mined out areas. Tailings ponds are harmful, since they contain chemicals that are present in the oil sands and solvents used in the extraction process. There is also always some residual oil floating the surface of the pond. For these reasons tailings ponds are dangerous for wildlife, and have to be provided by wildlife deterrent systems.

Tailings ponds act as a settling basin. Once the solids have settled, clarified water is recovered and recycled back to the processing plant for reuse. Tailings ponds are considered temporary storage facilities. The solids in the ponds are returned to the mine site to rebuild the mined out areas.

Tailings ponds are built using compacted clay. Clay has low permeability rate which helps prevent water leachate into the groundwater. Furthermore, vertical water pumps are installed around the tailings dyke. These wells serve to monitor the quality of the groundwater. The leachate are pumped back into the tailings pond (Oil sands magazine, 2015).

Figure 12. Oil sands water storage facilities. Source: Oil sands magazine, 2015.

Reclamation

By law all mines must be fully reclaimed. Reclamation process begin in mined out areas while mining continues in other areas. The replaced topsoil has muskeg peat and plant seeds, so that the re-growth can begin. Native plant and grasses are used to recreate productive landscapes. The first government reclamation certificate for an oil sands mine was issued in 2008 (Canada’s oil sands, 2016). Furthermore, tailings ponds have to be reclaimed. Re-vegetation of tailing ponds has proved difficulties. The sand is sterile having been subjected to boiling and with a high concentration of chemicals and solvent.

Figure 13. Oil sands mining trucks and shovels. Photo courtesy: Caterpillar

IN SITU DRILLING

80% of Alberta’s oil sand reserves are too deep to be mined, so they are recovered in situ (in place) by drilling wells (Oil sands magazine, 2016). This techniques recover oil from unconsolidated oil sands within the earth. All of the in situ technologies are based on concepts originally developed for recovery in conventional oil field. An in situ technology methods is not suitable for all geological conditions. What works in a deposit, may not work in another deposit.

An in situ commercial facilities consists of the following operations (Oil sands magazine, 2015):

 A series of well are drilled throughout the oil sands deposit;  A steam and power generation plant , which provides power for the facility and high-pressure steam for injection into the wells;  A central processing plant, where bitumen and water emulsion produced at the injection wells are separated;  A water treatment plant , where the recovered water is cleaned and recycled back into the process;  A bitumen storage facilities, where bitumen is diluted with a hydrocarbon solvent for transportation via pipeline to the upgrading and/or refining plant.

A commercial project operates almost 20 years. From a geological point of view obstacles are from the geometry of the reservoir sand bodies. Internal discontinuities and permeability barriers in the reservoir control the pattern of bitumen. In situ methods depend on the resolution of two main issues:

 Reducing the viscosity of the bitumen;

 Recovering the bitumen from deep within the earth.

These challenges are overcame using energy and water. In situ methods are more expensive respect to mining, but are smaller and simpler plant than a mining bitumen production plant. In situ techniques are different depending on the flowing capacity of bitumen. Recovering methods are divided in thermal recovery and non-thermal recovery. In situ production involves only the mobilization of the target product, while sand remains in the place.

In thermal recovery, indirect heating is employed for winning bitumen from formations of oil sands. Thermal recovery methods require to heat hydrocarbons so as to reduce their viscosity sufficiently to cause mobility. Thus, a heat transfer from a heat source to the heavy oils is required. The principal and more efficient method of heat transfer is convection (U.S. Patent No 3,338,306, 1967). Convection has economic efficiency in transferring heat energy underground from a source to the heavy oil to reduce its viscosity. As temperature is raised to 150° C, bitumen is fluid enough to flow through the pore holes in the earth and be pumped to the surface through recovery wells.

Projects that have been conducted to date are based on steam, hot water or combustion. Advances in technology, such as directional drilling, enable in situ operations to drill multiple wells (sometimes more than 20) from a single location, further reducing surface disturbance.

Thermal in situ recovery are:

 Steam Assisted Gravity Drainage (SAGD): the majority of the in situ operations use SAGD recovery. It is a well-recognized and proven commercial in situ thermal recovery process. Two horizontal wells are drilled five meter apart, one above the other. SAGD uses directional drilling technology, which let to drill wells also in the horizontal direction. Well depth can be up to 450 meters, while the horizontal length of a well can be 1000 meters long. Steam at high pressure is injected into the upper well and into the oil sands through the steam injection well. Heat mobilizes the bitumen allowing it to flow to the production well located beneath the steam injection well. Bitumen along with condensed water from the steam then flows to the surface. The bitumen is sent to a central process facility by an above ground pipeline, where water gas and impurities are removed before the product is shipped to the upgrader. More than the 90 per cent of water is recycled and reused in the process (Oil sands magazine, 2015).

Figure 14. SAGD recovery methods. Source: RAMP, 2016.

 Cyclic Steam Stimulation (CSS): CSS is also known as steam soak. It is a form of well stimulation that heats the reservoir by periodically injecting steam into a production well for a period of weeks to months. The high pressure steam injected fractures the oil sands deposit. Furthermore, heat is injected to reduce oil viscosity. This injection continues until the reservoir is fully saturated. Then soak phase starts, the well is closed allowing the heat to penetrate a considerable zone around the well. Following the well is put on production to pump off the bitumen that has become mobile. Each cycle of this process can take from four months to four years, and several cycles can be completed in a formation. Water is treated and recycled back into the process at the processing plant. Bitumen is sent to an upgrader for further processing. Recently directional drilling has been applied also to CSS allowing a less surface disturbance (Oil sands magazine, 2015).

Figure 15. CSS recovery stages. Source: Canada’s oil sands, 2016.

 Toe to Hell Air Injection (THAI): THAI is still in its infancy, but very promising. This method is a new combustion process that combines a vertical air injection well with a horizontal production well. First, steam is pumped down to the vertical well to heat the bitumen beneath the surface. When the reservoir reaches a certain temperature, air is injected down through the vertical well. Here, bitumen auto ignites. Air is supplied continuously to keep the combustion going. A portion of bitumen is consumed in the combustion zone. Bitumen outside the combustion zone is mobilized and is moved toward the recovery well. The recovery well pump to the surface bitumen. By burning the bitumen underground, THAI creates an upgraded crude oil. It also saves on steam and on energy use as noted by the company Petrobank Energy and Resources Ltd. According to Petrobank THAI can be used in area where CSS and SAGD cannot (The Oil Drum, 2007).

Figure 16. THAI well. Source: Petrobank Energy and resources Ltd.

A small number of in situ methods are based on non-thermal processes. Non-thermal recovery is applied when the flux of bitumen does not require heat.

 Cold production: cold production can be used in areas of the reservoir where the heavy oil is mobile enough to be pumped to the surface without the use of steam. The cold production concept uses long horizontal wells to pump the product to the surface without the use of heat.

 BEST-Bitumen Extraction Solvent Technology: reduction of viscosity of bitumen is achieved using an addition of diluents or emulsifier. Solvent can also be used for reservoir preconditioning before a thermal process.

Once extracted the bitumen is diluted with a hydrocarbon solvent and is sent to an upgrading facilities.

Figure 17. In situ oil sands plants. Source: The oil sands project, 2016.

Open pit mine vs. in situ drilling

Advantages from in situ drilling are (Oil sands magazine, 2016):

 Smaller land use: the surface area occupied by wellheads is smaller respect to the mine areas.  Less water use: in situ facilities use water for steam production, which is mostly recovered and recycled in the waste water processing plant. Open pit mine require large amount of fresh water for the gravity water-based separation process.  No tailings ponds: in situ operations do not require tailings ponds, since much of the sand is left in the ground.

Disadvantages from in situ drilling are (Oil sands magazine, 2016):

 Lower bitumen recovery rates: recovery rate percentages of bitumen depend on the method of extraction:

i. 20-40% from average world recovery of conventional oil (Muggeridge, 2014); ii. Up to 35-40% bitumen using cyclic steam simulation; iii. Up to 50-60% using steam assisted gravity drainage; iv. Up to 75-80% using Toe to Hell Air Injection 5; v. Up to 90% of bitumen from mining;  Great uncertainty: there is a degree of uncertainty in the positioning of the wellheads and expected recovery rates.  Intensive exploration activities for the positioning of the extraction wells.  More GHG emissions: in situ extraction requires large volume of steam to heat the bitumen in the ground. This steam is produced mainly burning natural gas. In situ extraction generates more GHG emissions per barrel of bitumen produced respect to mining.

UPGRADING

Bitumen recovered from the oil sands is a mixture of heavy hydrocarbons and contains impurities of water and solid particles and trace metals. These compounds cannot be handled by conventional refineries (Oil sands magazine, 2015). Bitumen before needs to be upgraded.

Upgrading is the process of separating the components of bitumen into different petroleum products for further processing. Synthetic Crude Oil (SCO) is the main product of upgrading. SCO is usually low sulfur and contains no residue or very heavy components and can be refined into commercial products. Upgrading uses temperature, pressure and catalyst to crack the big molecules of bitumen into smaller ones. Adding hydrogen and removing carbon from oil creates hydrocarbon molecules like those in light oil. Upgraded oil is used as a replacement for conventional crude oil to make marketable products such as gasoline, fuel oil, ethylene and propylene. Different combinations or variations of these processes are used by different companies to produce the desired end product. Upgrading is usually divided in two main stage processes (Oil sands magazine, 2015):

5 Nowadays THAI technology is not at the commercial scale, it is present only in pilot plants.

 Coking or hydrocracking: the bitumen is separated from the diluent-naphtha, which is sent back to the extraction facilities and reused. While, bitumen is transported to the cooker unit. Here, bitumen is heated to high temperatures. Heat breaks the long chain of carbon-carbon bonds. Smaller molecules are formed. Hydrogen is added to stabilize the hydrocarbon molecules. The superheated hydrocarbon vapors from the cooker units are sent to the fractionating tower where, using different boiling points, vapor condenses into naphtha, kerosene and gas oil. Excess of carbon, in the form of petroleum coke is a byproduct that is used as a fuel to generate the heat needed for this process.  Hydrotreating: this process stabilize the oil adding hydrogen at high temperatures to the unsaturated molecules. Moreover, hydrotreating remove impurities such as sulfur, nitrogen and trace metals. The output is a mixture called Synthetic Crude Oil.

Currently about 50% of Alberta’s bitumen is upgraded to crude oil, the remaining 50% is diluted and is sold directly to the market (Oil sands magazine, 2015). Bitumen from oil sands mine has to be upgraded, while bitumen from in situ technology can also be sold to the market without upgrading. SCO is sent to refineries through pipelines across the USA and Canada to be refined.

OIL SANDS IN THE WORLD Alberta is today the main producer of bitumen from oil sands. But, investments of different corporations are expanding also in other countries.

The viability of a commercial oil sands operation does not only require the presence of an exploitable deposit, but also a network of infrastructures. This network of infrastructures could be a challenge in some areas. In general an oil sands project requires (Unconventional series, 2016):

 Water;  Natural gas feed or an alternative energy source for the production of steam, hot water and hydrogen;  Diluents and solvent for piping bitumen;  Pipes transportation system;  Large trucks, in the case bitumen is mined out.

VENEZUELA

Venezuela has the largest oil deposit on the planet. The vast Orinoco Heavy Oil Belt contains 300·109 barrels of recoverable oil with actual technology (USGS, 2012). In Venezuela operations for recovery bitumen from oil sands are not developed as in Alberta. However, there are many similarities between Alberta and Venezuela’s oil sands. There are not barriers for the implementation of Canadian technology to Orinoco oil sands. Venezuelan deposits are located in a more favorable climate area without harsh winters. The oil in Orinoco oil belt is more mobile, with thicker reservoir and higher permeability values respect to Alberta’s oil sands. 88-92% of oil sands deposits is too deep to be mined out (Wykes, 2010). The other 8-12% is mineable. For this reason mining is not implemented at a commercial scale, but could be done in the future. Deposits in Venezuela are at higher temperature respect to Alberta. The higher temperature improves the viscosity of oil and this means that it is easier to extract bitumen. Venezuela unlike Alberta has not had resources to invest in oil sands extraction due to lack of capital and technical skills.

The reservoir contains heavy oil with a range of API gravity from 4 to 16 degrees (USGS, 2010). This oil has a high sulfur content and needs to be upgraded before refining. Oil sands in the Orinoco Belt are largely unconsolidated sandstone. Unconsolidated means that reservoir have high porosity with no significant grain to grain cementification. Nowadays, Venezuela extract heavy-oil with a technology

similar to in situ drilling for oil sands. This oil is the more profitable respect to extraction of bitumen from oil sands, since requires less technology and less operating costs. However, investment in exploration and in the creation of a network for oil sands extraction is ongoing.

Figure 18. Venezuela’s Orinoco heavy oil belt and concessions to different corporations. Source: Caracaschronicles.com

USA

USA oil sands deposits are present in the following states: Alaska, Utah, Alabama, California, Texas, Colorado, Wyoming, and Oklahoma (USGS, 1988).

Table 1. Location of the oil sands deposits by United States of America. The reserves are expressed in known reserves, while the recoverable reserves with the actual technology is unknown. As a rule of thumb 10% of the reserves could be recoverable using Canadian technology. Source: USGS, 1988.

State Proved reserves (Barrels) Utah 18.7·109 Alaska 15·109 Alabama 6.4·109 Texas 4.9·109 California 4.5·109 Kentucky 3.4·109 Oklahoma 0.8·109 New Mexico 0.35·109 Wyoming 0.15·109

Diffusion of oil sand industries in USA have many obstacles including:

 Remote and difficult topography of oil sand deposits;  USA oil sand deposits are more scattered and smaller respect to Alberta or Venezuela’s deposits;  Deposits are present mainly in regions with lack of water.

All these have resulted in an uneconomic oil resource. However, many companies are significant investing in R&D to bring oil sands extraction at the commercial scale in the USA (Humphries, 2008). Any recovery of oil sands in USA can be mainly done using mine technique (USGS, 1988).

Nowadays, exploration is done almost in every of these state. While, extraction is done in California and Utah. There are some in situ extraction facilities of extra-heavy-oil in California in the Pleasant Valley. The production is of about 1000 barrels per day (Try Valley Corp, 2016). In the state of Utah there are more than half of the USA oil sand deposits (U.S. Oil Sands, 2015). The deposits are located in the area

of Uinta Basin in the Eastern Utah. This region rich of oil and natural gas has been exploited for decades also for the extraction of conventional oil and natural gas. Hence, the implementation of oil sands extraction at a commercial scale will be supported by the network of infrastructures constructed from conventional oil and gas fields.

The Uinta Basin was formed about 66 millions of year ago. Oil shale and oil sands nowadays present were formed with the following geological process. Sediments eroded from high mountains that at that time surrounded the Uinta Basin, flowed into Lake Uinta. This formed a sequence of organic-rich shale and sandstone. Bitumen resources are present in Green River formation which is composed by oil sands and oil shale. Here, bitumen is not present as a continuous layer, but is present with lenticular beds with a bitumen saturation up to 16% w/w (U.S. Oil Sands, 2016). Utah’s bitumen is of higher quality respect to Alberta’s bitumen, since it is lighter and low sulfur content. The main advantage of Utah’s bitumen is the proximity to the markets and the operating environment is not as complicated as in Alberta.

While Alberta’s oil sands are water-wet, Utah’s oil sands are oil-wet. This is of fundamental importance for the separation process. Water-wet means that the sand grains are surrounded by water and then by oil. In Utah, the sand grains are directly surrounded by oil. Hot water separation process is not effective since gives low recovery rates of bitumen and leave a high amount of bitumen leftover. Bitumen separation is achieved using a solvent.

Figure 19. An oil sands seep at a Utah’s oil sands mine. Photo credit: U.S. Argonne National Laboratory.

A Canadian company, U.S. Oil Sands, in 2005 has leased an area of 130 km2 in the Uinta Basin. Since 2005 this company invested in developing a new extraction process suitable for Utah’s environmental conditions in order to extract bitumen from mining operations. The deepest oil sands excavated is 200 meters. If it is too deep too much solids have to be mined. Another factor to make viable the extraction is the bitumen content. In order to have an economical viable production, 14% w/w of bitumen content is required (U.S. Oil Sands, 2015). U.S. Oil Sands patented an extraction process which doesn’t require large amounts of water (U.S. Patent No. US20130062258 A1, 2014). The process produce bitumen and dry tailings. The extraction process is based on the use of a solvent. The solvent is citrus base. 95% of water and 98% of solvent are recycled and reused (U.S. Oil Sands, 2016). Nowadays, the production is at the first stage of the commercial scale with a production of 2000 barrels of bitumen per day (U.S. Oil Sands, 2016). This plant requires 1.5 barrels of water per barrels of bitumen produced (U.S. Oil Sands, 2015). Water used come from deep wells. However, this value of water does not account for water use for crop citrus and for solvent production.

TRINIDAD & TOBAGO

Oil sands are located in South-west of Trinidad & Tobago. Deposit are too shallow for in situ recovery. Deposits outcrops from the surface to 200 meters in depth (Petrotrin, 2011). Mining is required for the vast majority of deposits. Bitumen has been mined on a small scale quarry for pavement, but no conversion to oil is done at a commercial scale (Tar sands in T&T, 2010). Exploration and studies about economic feasibility are done. Corporations are investing in order to create a network of infrastructures needed for bitumen extraction, upgrading and refining (Tar sands in T&T, 2010). T&T is a region with lack of freshwater. Water for processing will be provided by a desalination plant in construction, which will convert water from ocean into fresh water. Furthermore, construction of an upgrader has been proposed.

NIGERIA

Nigeria is well known for its conventional oil and natural gas fields. This country has also deposits of oil sands. Starting from the 20th experiments has been done on Nigeria’s oil sands. In 2000 a Canadian corporation operated exploration of Nigerian oil sands. Deposits are shallow. Hence, extraction could be mainly done using mining operations.

REPUBLIC OF CONGO

There are two main oil sands deposits on Congo Basin near Brazzaville. Plan for extraction has been proposed by Italian corporation ENI. The resources are deep in the range of 100 to 200 meters. Extraction requires in situ technology to develop (Wykes, 2010).

DEMOCRATIC REPUBLIC OF CONGO

Deposit are present near Brazzaville along Congo River and the East part of the country near Uganda. This country has also great reserves of shale oil. Mining of oil sands in the east part of the country has taken place on a small scale for decades to produce asphalt. No extraction plans for oil sands or shale oil are currently present. However, corporations are exploring these deposits (Wykes, 2010)

MAGADASCAR

There are two deposit of oil sands in Madagascar: Tsimiroro and Bemolanga. While Tsimiroro’s oil sands can be recovered via in situ drilling, Bemolanga requires mining extraction. Bitumen is present with an average content of 6% (Andrianasolo, 1987). Small level of production of bitumen are active. The aim is to establish a bitumen production of 15000 barrels/day but the potentiality of a massive commercial operation can be 100000 barrels/day (Andrianasolo, 1987). Recovery is done using hot water extraction Canadian process. Bemolanga is one of the most arid part of the country. Tsimoro deposit is too deep for mining. An in situ extraction facility is being built to achieve a production of 31500 barrels/day.

RUSSIA

Russia has four main deposits of oil sands. The largest is Tunguska in Siberia. This deposit has extensive resources, the location in Siberia let this resources unrecoverable in the near future. The three other small deposits are in the Caucasus: Volga-Ural, Timan-Pechora, and North Caucasus-Mangyshlak. Corporations are experimenting feasibility of extraction both with in situ and mining techniques (Wykes, 2010).

METHODS Description and data We assessed current and potential impacts of oil sand extraction in terms of forest loss, water use, GHG emissions, and population potentially affected for the five countries – Canada, Venezuela, the United States, Republic of Congo, and Madagascar – whose deposits account for 93% of the oil sand deposits discovered worldwide. Alberta has the most extensive oil sand deposit, followed by Venezuela (Table 2). However, Venezuela’s deposits are larger (in terms of recoverable heavy oil) than those of Alberta (Table 3). These deposits have an area that is about 16% and 14% of the land areas of Alberta and Venezuela, respectively. Madagascar, the Republic of Congo and Utah have deposit and concession areas that are much smaller than those of Venezuela and Alberta. The oil sand deposits of Alberta and the Republic of Congo occur in major primary forests that so far have undergone relatively low rates of forest clearing (Potapov, 2008). In Venezuela heavy oil deposits are found in a region where forest vegetation has been more heavily affected by human activities and land use change (e.g., forestry and agriculture) (Vera, 2006), while the forest loss observed in Utah since the year 2000 was due to forest fire (UtahFireInfo, 2016). In Madagascar no forest cover is present over concessions areas. Table 2. Areas of oil sand deposits and concession areas. Forested area in year 2000 and net forest changes over the five countries studied.

Area Forested Net forest deposit area 2000 change 2000- (km2) (km2) 2014 (%) Canada 107,680 75,540 10 Venezuela 49,963 6,293 37 Rep. of Congo 1,800 1,732 8 Madagascar 352 no forest cover according to the classification used in this study USA 2,953 292 26

Table 3. Oil sand deposits in the world. Actual production and projected production from surface mining and in situ extraction (Alberta Oil sands industry, 2016), remaining oil in place and proved oil reserves. Zero values are used for areas where no extraction is expected to occur; missing data are reported as “-”. The deposit in Venezuela contains both oil sands and heavy oil, no geological surveys have been done to define the boundaries of oil sand deposits. Remaining oil in-place is the volume of oil within a formation before the start of production. Proved reserves are volumes of oil that geologic and engineering data demonstrate with reasonable certainty to be recoverable in future years under existing economic conditions and technology.

Projected Projected Remaining Proved Production Production production production oil in-place reserves surface mining in-situ (2025) surface (2025) in- (billion (billion (barrels/day) (barrels/day) mining situ barrels) barrels) (barrels/day) (barrels/day) Canada 1.38∙106 1.43∙106 1.84∙106 2.96∙106 315 175 Dem. Rep. of Congo 0 0 - - 0.30 0.03 Madagascar 1.5∙104 1.5∙104 1.8∙105 2.0∙105 200 25 Nigeria 0 0 - - 30 4 Rep. of Congo 0 0 0 3.00∙104 25 2.5 Russia 0 0 0 1.00∙105 245 34 Trinidad & Tobago 0 0 4.00∙104 0 2 1.50 USA 5.00∙103 0 3.50∙104 0 55 10 Venezuela 0 6.00∙105 0 2.10∙106 2500 300

Georeferenced data on the spatial extent of oil sand deposits and existing concession areas for oil sands extraction were acquired from various government and private institutions and agencies. Data on tree cover in the year 2000, annual forest loss between 2000 and 2015, and cumulative forest gain came from a recent high resolution (30m) satellite-based dataset produced by Hansen et al. (2013). Forest loss and fragmentation Following Hansen et al. (2013), we defined forested pixels as areas with at least 50% vegetation cover of 5 meters or higher. For oil sands areas where exploration and production have already started, historical net changes in forest cover were calculated simply as the difference between cumulative forest loss and cumulative forest gain. Potential forest losses were then calculated over entire deposit and concession areas, assuming that the entire area is cleared in mining operations and that the clear area needed for in situ extraction is the same as in the case of deposits that are currently under production in Alberta (i.e., ≈ 6 % of the land area). This is because the removal of forest vegetation associated with extraction technologies, exploration activities, and infrastructure construction (e.g., roads, pipelines) is likely to be similar across regions in order for the enterprise to be globally competitive. Net forest change

(i.e., ((forest loss + gain)/forested area in year 2000) × 100), in Alberta was calculated separately for operational mines and in-situ facilities. Because forest fires cause substantial forest loss in this region, the effect of fire on deforestation was removed using Alberta’s historical spatial wildfire data (Alberta Agriculture and Forestry, 2015), assuming that forest loss by fire would occur even in the absence of oil sand extraction. Hansen forest maps was validated using the methods from Carlson et al. (2013) by randomly selecting x points for each country or region considered. Points were classified as forested or not forested area based on land cover maps. We generated confusion matrices to calculate accuracy (po), the proportion of the total number of predictions that were correct, and the kappa coefficient (k), which takes into account the agreement occurring by chance, from a total of y validation points. The comparison yielded to the results shown in Table 4. Table 4. Validation of forest map for Alberta, Venezuela, Republic of Congo and Utah. Using x randomly selected points for each country or state we calculated accuracy (po) and kappa coefficient (k) from a total of y validation points. Validation for Utah has been done removing tree classes of Junipers and Pinyon since they are less than five meters high. For Alberta the kappa value is not so high since the land cover map used has as definition of forest a treed areas with at least a 10% ground cover of trees. Source of land cover map used po k x y Alberta A.B.M.I., 2000 0.66 0.42 1500 732 Venezuela Arino, 2010 0.85 0.69 1500 761 Rep. of Congo Arino, 2010 0.72 0.45 1500 726 Utah USGS, 2004 0.95 0.67 1500 142

Forest fragmentation was assessed over the deposits and concessions designated as in situ exploration and extraction areas (Alberta, Venezuela, and Republic of Congo) following the approach by Vogt et al. (2007). Using binary land cover maps, this methodology classified each pixel of the area of interest into one of four forest cover categories - core, patch, perforation, or edge. Cores were defined as forested pixels having all adjacent pixels as forested. Core pixels were also further classified into 3 sub-categories, based on the size of the contiguous forest core they belong to (<100 ha, 100 ha to 200 ha, or >200 ha). Patches were defined as forested pixels not containing core forest pixels. Edges were defined as forested pixels having at least one adjacent non-forested pixel. Perforations were defined as edge pixels surrounding a non-forested area with a maximum width of 100 m.

To summarize the extent of fragmentation within an area, we used a composite fragmentation index (CFI), defined as the ratio between the sum of number of pixels classified as “edges”, “perforated”,

“patches”, or smaller core areas (i.e., those <200 ha), and the total number of pixels in that area. CFI varies between 0 and 1; CFI=1 in areas with extremely fragmented forest cover, while CFI=0 in areas with no fragmented forest cores or no forest cover at all. Changes in fragmentation between 2000 and 2014 were expressed in terms of two indices, namely, CCFI (changes in CFI) and CPFI (changes in the number of patch pixels).

Water footprint of extraction

To calculate the amount of water used to extract and treat one liter of mined bitumen, we considered the amount of water required to attain the two process target densities starting from mined oil sand ore. Details about the calculations of water use for extraction and preparation can be found in Figure 20 and Figure 21. The typical composition of mined oil sand ore is reported in Table 5. Table 5. Typical composition and density of mined oil sands. Source: Oil sands magazine, 2016. Composition Density 3 [% v/v] [kg/m ]

Water 5 1000 Sand 85 2650 Bitumen 10 1050

3 We calculated the recoverable bitumen (푏푟푒푐%) from 1 m of mined oil sands ore according to the Alberta Energy Regulator Directive 082 formula (Alberta Energy Regulator, 2013) (Alberta Energy Regulator, 2013): 2 푏푟푒푐(%) = −202.7 + 54.1 ∙ 푤푡% − 2.5 ∙ 푤푡% where wt% is the average bitumen content, that ranges from 8% to 12% (Oil sands magazine, 2015), of the mined oil sands ore reported as weight percent. In Alberta 78% of mined oil sands were also upgraded in year 2015, a process requiring additional water. Data on upgrading water requirements are sparse, but detailed values have been reported by one of the upgraders – Scotford Upgrader – involved in the treatment of bitumen from Alberta’s oil sands (Canada Patent number 2004352). In this study we assumed that all other upgraders used roughly same amount of water per liter of Synthetic Crude Oil (SCO), the product of upgrading from bitumen. Estimates of the total amount of water needed to treat mined oil sands are based on data of production from mines and upgrading facilities in 2015 (Table 6, 7).

Table 6. Barrels of bitumen produced per day from Alberta’s oil sands mine facilities in 2015. Source: Oil sands magazine, 2016.

Project name (Company) BarrelsBitumen/day Horizon (Canadian Natural Resources LTD) 1.52·105 Kearl (Imperial Oil) 1.10·105 Kearl Expansion (Imperial oil) 1.10·105 Muskeg (Shell Canada) 1.55·105 Jackpine (Shell Canada) 1.00·105 Steepbank (Suncor Energy) 1.50·105 Millenium (Suncor Energy) 1.80·105 Mildred Lake (Syncrude) 2.00·105 Aurora North (Syncrude) 2.25·105 Total 1.38·106

Table 7. Barrels of Synthetic Crude Oil produced from Alberta’s upgraders in 2015. Source: Alberta Oil Sands Industry, 2016.

Upgrader (Company) BarrelsSCO/day Horizon (Canadian Natural Resources) 1.27·105 Millennium (Suncor Energy) 3.57·105 Mildred Lake (Syncrude) 3.50·105 Scotford Upgrader (Shell) 2.55·105 Total 1.09·106

The volume of water used to produce one liter of bitumen from in situ operations was calculated using data from annual facility performances expressed in terms of steam to oil ratio (m3 water / m3 produced bitumen) (Alberta Energy Regulator, 2015; Oil sands magazine, 2016)(Table 8). In situ plant performances are measured through steam to oil ratio (SOR). SOR is a measure of how efficiently energy is used in bitumen recovery from oil sands. SOR is the volume of water transformed into steam required to produce one cubic meter of bitumen (Alberta Energy, 2016). The lower the SOR value the higher the efficiency of water and energy usage.

Table 8. Data for each in situ operating project in terms of barrel of bitumen produced per day, Steam to Oil Ratio and technology of extraction used in year 2015. Commercial operating facilities use Steam Assisted Gravity Drainage (SAGD) and Cyclic Steam Simulation (CSS) extraction technologies. In situ plant performances are measured through Steam to Oil Ratio (SOR). SOR is a measure of how efficiently energy is used in bitumen recovery from oil sands. SOR is the volume of water transformed into steam required to produce one cubic meter of bitumen (Alberta Energy, 2016). The lower the SOR value the higher the efficiency of water and energy usage. Source: Alberta Energy Regulator, 2015; Oil sands magazine, 2016.

Production SOR 3 3 Project name (company) (BarrelsBitumen/day) (m Water/m bitumen) Techology Cold Lake (Imperial Oil) 1.80·105 3.40 CSS Wolf lake (Canadian Natural Resources) 0.13·105 3.90 CSS Primrose (Canadian Natural Resources) 1.07·105 3.90 CSS Christina lake (Cenovus) 2.10·105 1.70 SAGD Foster creek (Cenovus) 1.50·105 2.50 SAGD Long lake (CNOOC) 0.92·105 4.10 SAGD Great Divide ( Connacher) 0.20·105 4.50 SAGD Surmont (ConocoPhillips) 1.49·105 2.50 SAGD Jackfish (Devon Energy) 1.05·105 2.00 SAGD Sunrise (Husky Energy) 0.60·105 3.00 SAGD Kirby South (Canadian Natural Resources) 0.40·105 2.50 SAGD Christina lake (Meg Energy) 0.60·105 2.50 SAGD Firebag (Suncor) 1.80·105 2.70 SAGD MacKay river (Suncor) 0.38·105 3.00 SAGD Leismer (Statoil) 0.20·105 2.97 SAGD Total 1.43·106

20% of the water used in the processes is new fresh water, while the other 80% is recycled (Wu, 2009; Jacobs Consultancy, 2012; Oil sands magazine, 2016). To convert from barrels of bitumen to cubic meters of bitumen, we considered an API gravity of bitumen equal to 8°.

Similar to our calculations for potential forest loss, the potential water use for oil sands extraction and processing in deposits that are not yet under production was estimated assuming that the technology currently used in Alberta will be adopted also in the other regions. Rates of actual and projected production of major oil sand deposits in the world are in Table 3. In the calculation of water use the projected production rates are multiplied by the water footprint of bitumen extraction and processing determined for Alberta.

Net water consumption from mining and in situ was assessed considering that 80% of the total water footprint is recycled, and adding the water required to upgrade and refine bitumen. Assessing other impacts GHG emissions from oil sands extraction were calculated considering projected production rates of bitumen from in situ or mine operations (Table 3), assuming an 80% bitumen yield to SCO (Canada Patent number CN 105339469 A) and using for mine and in situ the mean values of GHG emissions provided by a previous study (Charpentier, 2009). The number of people potentially affected was assessed using data on population distribution taken from CIESIN’s Gridded Population of the World map (GPWv4) for the year 2010 and calculating the number of people living within the perimeters of oil sand deposits and concessions. Water stress areas were assessed overlaying the five oil sands deposit and concessions areas with a water depletion map Brauman et al. (2016). Water depletion was calculated as the fraction of renewable water consumptively used for human activities.

ENVIRONMENTAL IMPACTS OF OIL SANDS PRODUCTION The total water footprint of bitumen extraction and processing from oil sand deposits in Alberta differed greatly depending on the extraction method. We found that 2.8 liters of water were required to obtain a liter of bitumen using in situ drilling.

Figure 20. Volume of water required to extract 1 barrel of bitumen from in-situ drilling. 1 barrel =0.159 m3 Bitumen that is too deep to be mined is extracted using thermal in-situ technologies. In-situ facilities use horizontal drilling technology. High pressure and temperature steam is injected into the reservoir to reduce bitumen viscosity. An emulsion of bitumen and condensed steam (produced water) is pumped out to the ground surface. Approximately 10% of the injected steam is retained into the reservoir and another 10% of steam is disposed or lost in disposal and blowdown processes at the cogeneration power plant (Jacobs Consultancy, 2012). Make-up fresh water is required to compensate these losses and maintain the water balance required by this process. This water is usually withdrawn from fresh or saline groundwater reservoirs. Water from saline aquifers has to be transformed into fresh water (desalination) before it can be used in this process.

Conversely, the water footprint of surface mining was 28.5 l H2O per l bitumen. Net fresh water consumed is the 20% of the total water footprint, since the 80% was recycled in both extraction methods recycled (Wu, 2009; Jacobs Consultancy, 2012; Oil sands magazine, 2016). Where information was available, these values agreed well with previous studies that determined the net water footprint of bitumen extraction (Wu, 2009; World Energy Outlook 2012, 2012).

Figure 21. Volume of water required to extract 1 barrel of bitumen from mined oil sands in Canada. 1 barrel =0.159 m3 Oil sands close to the ground surface (<75m in depth) are mined (Oil sands magazine, 2016). Bitumen is extracted from mined oil sands using a water intensive process known as hot water extraction (Oil sands magazine, 2016). First, at the slurry preparation plant water is added to the mined oil sand ore to obtain a slurry with a density of 1500 kg/m3 (Oil sands magazine, 2016). Second, the produced slurry is sent to the bitumen extraction plant, where more water is added to reach a density of 1400 kg/m3 (Oil sands magazine, 2016). Bitumen is separated by gravity from tailings. Bitumen is usually upgraded into Synthetic Crude Oil (SCO) before being sent to refineries. Tailings are one of the by- products from mining operations. They are made of water, sand, clay, left-over bitumen and trace amounts of chemicals used in the extraction process. Tailings are stored in large dykes named tailing ponds from 3 to 5 years (Government of Alberta, 2011). In 2013 the tailing pond area in Alberta was about 220 km2 with a volume of fine fluid tailings stored of 0.975 km (Alberta Environment and Parks, 2015). Approximately 80% of water used in the extraction process is recycled. The other 20% of water is freshwater withdrawn from Athabasca River. Where information was available, these values agreed well with previous studies that determined the net water footprint of bitumen extraction (Wu, 2009; IEA, 2012). At current rates of production, we estimate 3 -1 that extraction from Canadian oil sands requires 0.49 km H2O yr .

Figure 22. Forest fire in Alberta from 2000 to 2014. The boxed area in the main map are shown in the corresponding local map to the right. About 1476 km2 of Alberta’s forests – 15% of forests covering Alberta’s oil sands concession areas in year 2000 – have been removed since the start of the century. In addition 6% of “natural” forest loss is due to fire (unrelated to extraction activities) in oil sand concession areas since the beginning of the century. Forest fragmentation has also increased – both in terms of cumulative fragmentation index (CFI, the areal fraction of sites located at the edges of forested areas, small forest patches, and smaller forest cores; see Methods) and in the number of forest patches. This is especially true for in situ operations where CFI and the number of forest patches increased by 7% and 81%, respectively, since the year 2000. It should also be noted that our estimate of forest loss for in situ concessions is likely conservative as typical exploration lines are 4 meters wide– narrower than the 30m resolution of the dataset (Hansen, 2013) – with 60 meters spacing between lines (Figure 23, 24).

Figure 23. Exploration seismic lines used for in situ extraction in Alberta. It is also visible a typical configuration of an in situ extraction facilities. Source: Google Earth, 2016.

Figure 24. Exploration seismic lines used for in situ extraction in Venezuela over Orinoco Heavy Oil Belt area. Source: Google Earth, 2016. While multiple environmental impacts are apparent, there have however been substantial economic benefits with affordable and secure energy to U.S. and Canadian market, more than 478,000

jobs created in Canada in year 2012 (3% of all jobs in the country) (PetroLMI, 2016), and tax revenues and royalties paid to the governments (Exxon Mobil, 2016). Thus there are clear and ongoing tradeoffs between economic development, energy, and the environment. The effect of oil sand extraction from undeveloped deposits in Alberta and other study countries was estimated assuming that the water use and the land footprint (i.e., fraction of deposit area that needs to be clear of vegetation) of mining or drilling operations were the same as those calculated for operational concessions in Alberta (Table 9). Table 9. Impacts from oil sands extraction and processing in Alberta: in-situ versus surface mining impacts on water, forest loss, GHG emissions (in mass of equivalent CO2 per barrel of Synthetic Crude Oil, SCO), and number of jobs directly created by the oil sand industry. Surface mining In-situ drilling Source Water (l water / l bitumen) 28.52 2.77 This study Net forest loss (%) 100 6 This study

GHG emissions (kg CO2eq/barrel SCO) 113 138 Charpentier (2009) Direct job creation (#workers) 14,750 11,274 PetroLMI (2016)

The impact on water resources and forest cover was then evaluated using site specific water stress and forest cover data (see Methods Section). We estimate that full exploitation of oil sands deposits will have profound environmental consequences, with the greatest impacts by far expected in Canada and Venezuela (Table 10).

Table 10. Future potential impacts from extraction and processing of oil sands. Only relatively shallow deposits can be mined, while bitumen extraction from deeper deposits requires in situ technology. In Alberta only 4.5% of the deposit area is mineable, while in Venezuela mining can potentially take place in 12% of the deposit area. Deposits in Rep. of Congo are too deep to be mined, while Utah has only shallow deposits. In Madagascar 50% of the deposit is mineable and the other 50% is suitable for in situ extraction. Actual water stress over oil sand deposits expressed as in Brauman et al. (2016). Water depletion is calculated as the fraction of renewable water consumptively used for human activities.

Country GHG emissions Freshwater use Deforestation Population Actual water

-1 3 -1 2 (source) (Mtonne CO2 eq yr ) (km yr ) (km ) potentially affected stress

Canada actual 102.81 0.486 1,476 696 <5% (Alberta Environment and Parks, 2015) Canada potential 179.35 0.680 7,482 162,000 <5% (Alberta Energy Regulator, 2016; ArcGIS, 2014) Venezuela potential 84.32 0.065 871 455,916 <5% (acknowledgements) Rep. of Congo potential 1.20 0.001 97 20,476 <5% (Global Forest Watch, 2016) Madagascar potential 13.97 0.064 - 1,796 <5% (acknowledgements) USA (Utah) potential 1.15 0.011 215 59 Dry year (USGS, 2005)

Specifically, cumulative forest loss across these studied oil sand deposits (Figure 25, 26, 27, 30) may eventually reach 8,665 km2, an area equal to 5% of global forest loss in the year 2014 (Hansen, 3 -1 2013). In addition, fresh water use may nearly triple to 1.31 km H2O yr with important impacts on the local freshwater resources, as the case of Utah where the deposits are located in a potentially water- -1 stressed areas. Moreover, projected annual GHG emissions (383 Mtonne CO2 eq yr ) – due in large part to growth in energy-intensive in situ production – would be commensurate with those from land use and land cover change for all of Indonesia (Carlson, 2013), and as many as many as 640,000 people could potentially be affected by the complete development of these five world’s oil sands deposits (Table 10).

Figure 25. Map of forest cover, forest loss, oil sand concession boundaries in year 2013 and deposits in Alberta. The boxed area in the main map are shown in the corresponding local map to the right. Map 1 shows the typical forest loss over an oil sand mining area; map 2 shows forest loss and fragmentation over situ drilling extraction areas. In-situ concession areas exhibit also a massive network of exploration lines (with a typical width of 4 meters) that cannot be detected by the 30m resolution forest cover dataset used in this study. Source: Alberta Environment and Parks, 2015; Alberta Energy Regulator, 2016; ArcGIS, 2014.

Figure 26. Map of forest cover, forest loss and oil sand deposits in Venezuela. The boxed area shows the typical trend of exploration lines over an oil sands area suitable for in situ drilling extraction.

Figure 27. Map of forest cover, forest loss and oil sand concessions in Republic of Congo. Source: Global Forest Watch, 2016

Figure 28 Oil sand concessions in Madagascar. No forest is present inside the concession areas according to the methods used in this study.

Figure 29. Madagascar's concessions areas.

Figure 30. Map of forest cover, forest loss over oil sand deposits in Utah. Source: USGS, 2005.

TRADEOFFS BETWEEN ECONOMY, ENERGY, AND THE ENVIRONMENT In Alberta the amount of water required to extract and process bitumen from oil sands is of the same order of magnitude as the province’s water consumption for irrigated agriculture. Thus, it can be argued that in this state the oil sand industry was able to develop because the region is relatively rich of water resources. The projected growth by 2025 in oil sand production is expected to lead to a 40% increase in freshwater appropriations by oil sand operations. This could threaten Alberta’s water resources, with important ecological and societal impacts, especially in the Athabasca watershed, where most of the deposits and production are located. The development of oil sand extraction in countries with drier climates is expected to exert a stronger pressure on the local water resources, likely competing with other water uses such as crop production and flows for aquatic ecosystems. Thus the increasing reliance on oil sand extraction may raise new water security concerns in countries where deposits are located in water stressed watersheds (Utah), thereby further sustaining ongoing debates on competing water uses within the context of the water-energy nexus (Rulli, 2016). This study demonstrates that mining operations and in situ drilling differ significantly in their environmental impacts. While in-situ technology is more energy intensive, the process requires less water, does not produce tailings, and has less extensive impacts on overall forest cover. For in situ concessions, forest loss tends to be more scattered, limited to areas of intensive exploration activities for the positioning of the extraction wells, and construction of infrastructures (e.g., roads, pipelines). Even though in-situ drilling entails a smaller net change in forest cover than surface mining, its impact on wildlife habitat and landscape fragmentation should not be underappreciated. Conversely, forest vegetation is completely cleared within mining operations. In addition, while oil sand extraction is not expected to displace a large number of people – as these deposits are located in remote forested areas – these operations are likely to have important impacts on carbon sequestration, because of carbon emissions associated with land use change and deforestation. Overall, in situ drilling is more expensive (Muggeridge, 2014) and has less environmental impacts in terms of water use and forest loss; however it requires more energy and has a bitumen recovery rate lower than mining (up to 60% for in situ vs. 90% for mining) (Oil sands magazine, 2015). For this reason the current trend, with low oil price, is to increase mining production instead of in situ extraction (PetroLMI, 2016). This choice, likely driven by economic factors, is in disagreement with the previous trend of increasing investments in situ production (PetroLMI, 2016) and does not account for the big difference in environmental impacts highlighted by this study.

Ongoing technological innovation is trying to reduce the environmental impacts of oil sands production. GHG emissions per barrel have already decreased by 25-40% in the last 25 years (The Oxford Institute for Energy studies, 2016). Nevertheless, oil sand extraction and processing still requires a large amount of energy, which corresponds to emitting about three times more GHGs than the production of conventional oil. The water required to upgrade bitumen has decreased by 40% since 2005 (Table 11). Likewise, the amount of water needed to extract bitumen using in situ techniques has, while we found that more water is required to win bitumen from mined oil sands than what was reported by previous studies. Our results show that the gasoline produced from secondary conventional oil recovery techniques (i.e., water flooding) requires almost the same amount of water used to produce gasoline from mined oil sands. Progress in technology, proximity to the markets and high prices of crude oil have favored the proliferation of the oil sand industry in Alberta. Though this study has focused on the major environmental impacts of oil sand extraction, this source of energy also offers some advantages compared to conventional oil. First, the recovery rate of bitumen from oil sands is greater than for conventional oil. The average recovery rates for conventional oil range from 20 to 40% of the deposit, while the recovery rate of bitumen from oil sand extraction varies from 50-60% (in-situ drilling) to 90% (surface mining) (Oil sands magazine, 2016; Muggeridge, 2014)). Second, oil sand deposits decline more slowly (4% decline per year) than those of conventional oil (20% per year), allowing these deposits to last longer (e.g., 30 years in the case of Alberta) (Oil sands magazine, 2016). Because of their long lifespan, oil sand deposits are drawing increasing interest and investments from oil and gas corporations. Moreover, oil sand industries have led to considerable job creation. For example, in 2012, the oil sands industry in Canada and U.S. employed – directly and indirectly – almost 558,000 people (80,000 in U.S.) (IHS CERA, 2014; Canadian Energy Research Institute, 2011). Employment is expected to grow in the most positive scenario of oil sands development to 2.2 million jobs in the U.S. and Canada by 2035 (Canadian Energy Research Institute, 2011). By 2020 direct job creation in the oil sand industry in Alberta is expected to add about 5,170 new jobs to the existing workforce for a total of 35,070 (PetroLMI, 2016). As a result of this oil sands ‘boom’, Alberta has become one of the richest regions in the world, with a GDP per capita in year 2014 (91,000 U.S. dollar) among the top five countries in the world (Alberta Government, 2016; The World Bank, 2014). Thus, pairing oil sands extraction with responsible policies can ultimately enhance economic growth and create employment opportunities. Yet despite these obvious benefits, our study clearly demonstrates that environmental considerations need to be incorporated into decision-making surrounding oil sands.

Table 11. Net water consumption and GHG emissions from extraction and processing of petroleum products. GHG emissions are obtained from a review of seven publicly available studies that account for: recovery & extraction, upgrading, electricity supply chain, natural gas supply chain, venting & flaring, fugitive leaks and fugitive tailings ponds (Charpentier, 2009). Net water consumed Source GHG emissions from well- Source to-tank Bitumen upgrading to SCO 0.6 l water/ l SCO This study

36.9 – 66.6 kg CO2eq./ bbl Nimana, 2015 Bitumen upgrading to SCO 1.0 l water/ l SCO Peachey, 2005 BITUMEN

Oil refining ( U.S. average) 1.4 l water/ l refined Wu, 2009 46.3 – 92.1 kg CO2eq./ bbl Nimana, 2015

REFINED

Gasoline (Alberta oil sands 7.7 l water/ l gasoline This study 62 – 164 kg CO2eq./ bbl SCO Charpentier, mining) 2009

Gasoline (Alberta oil sands 5.2 l water/ l gasoline Wu, 2009 mining)

Gasoline (Alberta oil sands 2.0 l water/ l gasoline This study 99 – 176 kg CO / bbl Charpentier, in situ)a 2eq. SCO 2009

Gasoline (Alberta oil sands 2.6 – 6.2 l water/ l Wu, 2009 in situ) gasoline

Gasoline (U.S. conventional 0.2 l water/ l gasoline Wu, 2009 crude-primary recovery) Gasoline (U.S. conventional 3.4 – 6.6 l water/ l 27 – 58 kg CO2eq./ bbl Wu, 2009 Charpentier, crude-secondary recovery) b gasoline GASOLINE 2009 Gasoline (Saudi 2.8 – 5.8 l water/ l Wu, 2009 conventional crude) gasoline

a Bitumen extracted from in situ drilling is not upgraded to SCO, but it is sent directly to refineries. b Secondary recovery via water flooding.

Most global oil sand reserves are expected to be put under production and exploited in the near future (U.S. EIA, 2016). While low oil prices (in year 2014 and 2015) have presently slowed extraction from oil sand deposits, investments are ongoing, with massive extraction for commercial production likely to take place as soon as oil prices become higher again. Production is expected to increase from the current 3.5 million barrels a day to 7.5 million barrels a day by 2025. The environmental impacts of oil sands extraction are already apparent in places where production is already occurring. Thus areas

containing oil sands can expect marked changes in terms of land cover and freshwater appropriations in the near future. This expected escalation in oil sand extraction may therefore alter the existing equilibria in the water-energy system and reshape patterns of water allocations, governance strategies, and associated institutional arrangements.

REFERENCES Alberta land cover (Alberta Biodiversity Monitoring Institute, 2000); http://archive.abmi.ca/abmi/rawdata/geospatial/gisdownload.jsp?categoryId=2&subcategoryId=4. Accessed: May, 2016 Alberta mineable oil sands plant statistics (Alberta Energy Regulator, 2014); http://www.aer.ca/documents/sts/ST39-2014.pdf.

Alberta oil sands area (Alberta Energy (2016); http://www.energy.alberta.ca/LandAccess/pdfs/OSAagreeStats.pdf. Accessed April, 2016.

Alberta Oil sands industry (2016) Quarterly update, fall 2015; http://www.albertacanada.com/business/statistics/oil-sands-quarterly.aspx. Accessed March, 2016

Arino O., J. Ramos, V. Kalogirou, P. Defourny and F. Achard. GlobCover 2009. ESA Living Planet Symposium, 27 June. 2 July 2010, Bergen, Norway Brauman, K. A., Richter, B. D., Postel, S., Malsy, M., & Flörke, M. (2016). Water depletion: An improved metric for incorporating seasonal and dry-year water scarcity into water risk assessments. Elementa: Science of the Anthropocene, 4(1), 000083. Canada oil production expect to increase despite lower prices (EIA, Energy Information Agency, 2016); http://www.eia.gov/todayinenergy/detail.cfm?id=25112. Accessed May, 2016

Canada’s oil sands (Exxon Mobil Corporation, 2016); http://corporate.exxonmobil.com/en/current-issues/oil-sands/canadian-oil- sands/overview?parentId=c3ebc0ca-65e0-4116-9506-3c2ba8c4a568. Accessed June, 2016

Carlson, Kimberly M., et al. "Carbon emissions from forest conversion by Kalimantan oil palm plantations." Nature Climate Change 3.3 (2013): 283-287. Center for International Earth Science Information Network - CIESIN – Columbia University. 2015. Gridded Population of the World, Version 4 (GPWv4): Population Count. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). http://dx.doi.org/10.7927/H4BG2KXS. Accessed May 2016. Charpentier, Alex D., Joule A. Bergerson, and Heather L. MacLean. "Understanding the Canadian oil sands industry’s greenhouse gas emissions. "Environmental research letters 4.1 (2009): 014005.

Chew, Kenneth J. "The future of oil: unconventional fossil fuels." Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences 372.2006 (2014): 20120324.

Dai, Qi, and Keng H. Chung. "Hot water extraction process mechanism using model oil sands." Fuel 75.2 (1996): 220-226.

Directive 082 (Alberta Energy Regulator, 2013); https://www.aer.ca/documents/directives/Directive082.pdf. Economic Impacts of Staged Development of Oil Sands Projects in Alberta (2010-2035) (Canadian Energy Research Institute, 2011); http://static1.squarespace.com/static/557705f1e4b0c73f726133e1/t/570e74f7f8baf3a82bb4132a/146 0565241762/2011-08-24+CERI+Study+125+Section+1.pdf. Accessed May, 2016 Extracting bitumen from mined oil sands (Oil sands magazine, 2015); http://www.oilsandsmagazine.com/oil-sands-mining. Accessed March, 2016.

GDP per capita (The World Bank, 2014); http://data.worldbank.org/indicator/NY.GDP.PCAP.CD?page=1. Accessed June, 2016 Google Earth Pro 7.1. 57°16'42.66"N, 111°04'07.24"W, elevation 5 km. 3D. [Viewed 9 July 2016].

Google Earth Pro 7.1. 8°47'46.34"N, 63°20'34.43"W, elevation 10 km. 3D. [Viewed 9 July 2016].

Gordon, Deborah. "Understanding unconventional oil." (2012).

Gosselin, Pierre, et al. "Environmental and health impacts of Canada’s oil sands industry." Royal Society of Canada Expert panel report, Ottawa, ON (2010).

Green River total petroleum system (USGS, 2005); https://www.sciencebase.gov/catalog/item/4f4e4799e4b07f02db48faa7. Accessed: March, 2016

Hansen, Matthew C., et al. "High-resolution global maps of 21st-century forest cover change." Science 342.6160 (2013): 850-853.

Highlights of the Alberta economy (Alberta Government, 2016); https://www.albertacanada.com/files/albertacanada/SP-EH_highlightsABEconomyPresentation.pdf. Accessed June, 2016. Hot water extraction process and apparatus (Kohur, S., Ray, P., Stevenson, C. Git, L.,Jan, Hrouda, 1995). Canada Patent number 2004352.

Humphries, Marc. "North American oil sands: history of development, prospects for the future." LIBRARY OF CONGRESS WASHINGTON DC CONGRESSIONAL RESEARCH SERVICE, 2008.

In situ performance presentation (Alberta Energy Regulator, 2015); http://www.aer.ca/data-and- publications/activity-and-data/in-situ-performance-presentations. Accessed April, 2016 Kelly, Erin N., et al. "Oil sands development contributes elements toxic at low concentrations to the Athabasca River and its tributaries." Proceedings of the National Academy of Sciences 107.37 (2010): 16178-16183.

Labor demand outlook to 2020 (PetroLMI, 2016); http://www.careersinoilandgas.com/media/243763/oil_sands_outlook_online.pdf. Accessed April, 2016. Mining operations (Oil sands magazine, 2016); http://www.oilsandsmagazine.com/projects/oilsands-mining. Accessed March, 2016.

Mossop, Grant D. "Geology of the Athabasca oil sands." Science 207.4427 (1980): 145-152.

Muggeridge, Ann, et al. "Recovery rates, enhanced oil recovery and technological limits." Phil. Trans. R. Soc. A 372.2006 (2014): 20120320.

National Inventory Report, 1990–2010, and Natural Resources Canada (Environment Canada, 2012).

Nimana, Balwinder, Christina Canter, and Amit Kumar. "Energy consumption and greenhouse gas emissions in upgrading and refining of Canada's oil sands products." Energy 83 (2015): 65-79.

Oil sands (Alberta Energy Regulator, 2016); http://www1.aer.ca/ProductCatalogue/OILS.html. Accessed April, 2016.

Oil sands chemical & physical proprieties (Oil sands magazine (2016); http://www.oilsandsmagazine.com/oil-sands-deposit. Accessed March, 2016. Oil sands chemical and physical proprieties (Oil sands magazine, 2016); http://www.oilsandsmagazine.com/oil-sands-deposit. Accessed March, 2016. Oil Sands Economic Benefits: Today and in the Future (IHS CERA, 2014); http://www.ontarioenergyboard.ca/html/oebenergyeast/documents/IHS- CERA_Oil_Sands_Economic_Benefits_Jan2014.pdf. Accessed March, 2016. Accessed April, 2016.

Oil sands mining operators- Water allocation and use by year (Alberta Environment and Parks, 2015); http://osip.alberta.ca/library/Dataset/Details/51. Accessed March, 2016.

Oil sands production profile: 2004-2014 (Alberta Energy, 2016); http://www.energy.alberta.ca/Org/pdfs/InitiativeOSPP.pdf. Accessed April, 2016.

Ore Preparation Plant Explained (Oil sands magazine, 2016); http://www.oilsandsmagazine.com/oil-sands-surface-mining-ore-preparation-opp-slurry- preparation-plant-spp. Accessed March, 2016. Partial upgrading process for heavy oil and bitumen (Kresnvak, S., 2014). Canada Patent number CN 105339469 A.

Peachey, Bruce, and P. Eng. "Strategic Needs for Energy Related Water Use Technologies: Water and the EnergyINet." New Paradigm Engineering (2005): 34-36.

Potapov, Peter, et al. "Mapping the world’s intact forest landscapes by remote sensing." Ecology and Society 13.2 (2008): 51.

Potential extent of Athabasca oil sands (ArcGIS, 2014); http://www.arcgis.com/home/item.html?id=4b7863a60c8844a3824a7d58f0238afire. Accessed April, 2016.

Repubblica Bolivariana de Venezuela, country pasture/forage profile (Vera R., 2006); http://www.fao.org/ag/agp/agpc/doc/counprof/venezuela/venezuela.htm. Accessed May, 2016

Republic of the Congo mining permits (Global Forest Watch, 2016); http://www.globalforestwatch.org/map/6/0.66/14.88/COG/grayscale/forestgain/640?tab=countries- tab&threshold=30&dont_analyze=true. Accessed April, 2016.

Resources to reserves (IEA, International Energy Agency, 2013); https://www.iea.org/publications/freepublications/publication/Resources2013.pdf

Rogner, Hans-Holger. "An assessment of world hydrocarbon resources."Annual review of energy and the environment 22.1 (1997): 217-262.

Rulli, Maria Cristina, et al. "The water-land-food nexus of first-generation biofuels." Scientific reports 6 (2016).

Shapefiles for oil sands project boundaries and centroid points: 2013 (Alberta Environment and Parks, 2015); http://osip.alberta.ca/library/Dataset/Details/670. Accessed April, 2016.

Spatial wildfire data (Alberta Agriculture and Forestry, 2015); http://wildfire.alberta.ca/wildfire- maps/historical-wildfire-information/spatial-wildfire-data.aspx. Accessed April, 2016

Speight, James G. Shale gas production processes. Gulf Professional Publishing, 2013.

Tailings management (Government of Alberta, 2011); http://www.oilsands.alberta.ca/factsheets/tailings_management.pdf. Accessed April, 2016

The future of Canadian oil sands (The Oxford Institute for Energy studies, 2016); https://www.oxfordenergy.org/wpcms/wp-content/uploads/2016/02/The-Future-of-the-Canadian- Oil-Sands-WPM-64.pdf. Accessed June, 2016 The Outlook for Energy: A View to 2040 (Exxon Mobil Corporation, 2016), Corporation, Irving, TX. The world factbook (Central Intelligence Agency, 2016), Washington D.C. https://www.cia.gov/library/publications/the-world-factbook/rankorder/2244rank.html. Accessed March, 2016.

Thermal in situ water conservation study- Final Report (Jacobs Consultancy, 2012). Thermal in-situ bitumen extraction (Oil Sands Magazine, 2015); http://www.oilsandsmagazine.com/oil-sands-in-situ/. Accessed March, 2016

Thermal in-situ facilities (Oil sands magazine, 2016); http://www.oilsandsmagazine.com/projects/thermal-in-situ. Accessed March, 2016 Total volume of the oil sands Tailings Ponds over time (Alberta Environment and Parks, 2015); http://osip.alberta.ca/library/Dataset/Details/545. Accessed: March, 2016.

USGS National Gap Analysis Program. 2004. Provisional Digital Land Cover Map for the Southwestern United States. Version 1.0. RS/GIS Laboratory, College of Natural Resources, Utah State University

UtahFireInfo (2016); http://www.utahfireinfo.gov/. Accessed May, 2016.

Vogt, Peter, et al. "Mapping spatial patterns with morphological image processing." Landscape ecology 22.2 (2007): 171-177.

Water usage in the oil sands (Oil sands magazine, 2016); http://www.oilsandsmagazine.com/oil- sands-water-usage. Accessed March, 2016.

Water use history for oil sands projects (Alberta Environment and Parks, 2015); http://osip.alberta.ca/library/Dataset/Details/56.

Wu, May, et al. "Water consumption in the production of ethanol and petroleum gasoline." Environmental management 44.5 (2009): 981-997.

Wykes, Sarah, and Steven Heywood. "Tar sands Fuelling the climate crisis, undermining EU energy security and damaging development objectives."Friends of the Earth Europe, Friends of the Earth France, CEE Bankwatch, Milieudefensie (2010): 9.

CONCLUSIONS 39% of the area of world high quality shale deposits is located in areas affected by surface water stress and where 171 million people live. Water stress is particularly high in areas that are highly populated, irrigated, or with low water availability (Mekonnen, 2016). In these regions an increase in human appropriation of freshwater resources for shale gas or shale oil extraction would markedly increase competition with the existing water uses for agriculture and environmental flows. The extraction of shale deposits is expected to affect not only surface water resources but also groundwater (e.g., Freyman, 2014). Our analysis shows that 8% of high quality shale deposits are located in regions affected by groundwater stress. About 6% of the world’s high quality shale areas are affected by both surface and groundwater stresses. In these water constrained areas a water market could be developed increasing water prices and enhancing the competition between water for food production through irrigation and water for oil and natural gas production from shale resources. Thus while water markets may offer an effective solution for allocating water rights within water-limited systems (Debaere et al. 2014), the ultimate result may be the displacement of agriculture if shale energy companies are willing to pay a higher price for water use. We find that 25% of high quality shale areas worldwide are in irrigated areas where about 220 million people live. About 7% of the high quality shale areas are located in regions where water use for irrigation has been projected to increase in order to close the crop yield gap – the difference between actual and attainable yields. Thus, competition for water use in these areas will not only increase due to shale energy production but also be exacerbated by a greater need for irrigation water.

The increasing food and energy needs of humanity (e.g., Suweis, 2013) and the possible local decline in water availability as an effect of climate change (IPCC, 2013) are expected to increase pressure on freshwater resources (Muller, 2012; Davis, 2014). As a result, regions affected by increasing water stress will face not only environmental and social challenges, but also see the emergence of financial obstacles both for the food and energy industries. In some water constrained areas where shale resources are present a rush for water appropriation by oil and gas companies has surged, leaving the agricultural sector with limited water supply (Nicot, 2012; Mauter, 2014). This pattern is expected to occur in many water stressed agricultural regions in which shale deposits are going to be developed because of the higher profits of water use for energy than for food production. In fact, despite the current low price of oil and natural gas, the use of water for energy production generates greater profits than agriculture. Oil production from hydraulic fracturing is also less water intensive than oil from oil sands and conventional

oil through secondary recovery (i.e., water flooding of the reservoir) and therefore more economically convenient. Only conventional oil from primary recovery (i.e., using lift pump) and conventional gas, which has a zero water footprint in extraction, exhibits a higher economic yield of water use than shale oil and shale gas. Interestingly, bioethanol and biodiesel are less profitable - in terms of economic yield of water - than fossil oil and gas, but are more economical than the production of certain food crops. Moreover, our analysis shows that, despite their similar water requirements per unit of energy produced, shale oil and shale gas strongly differ in the economic yields of the water used in their production processes.

The total water footprint of bitumen extraction and processing from oil sand deposits in Alberta differed greatly depending on the extraction method. We found that 2.8 liters of water were required to obtain a liter of bitumen using in situ drilling. Conversely, the water footprint of surface mining was 28.5 l H2O per l bitumen. At current rates of production, we estimate that extraction from Canadian oil sands 3 -1 requires 0.49 km H2O yr . Substantial GHG emissions (from well-to- refinery entrance gate) have also been associated with this production. Currently, oil sands production accounts for 103 Mtonne CO2 eq yr-1 (14% of Canada’s GHG emission in year 2014).

About 1476 km2 of Alberta’s forests – 15% of forests covering Alberta’s oil sands concession areas in year 2000 – have been removed since the start of the century. Forest fragmentation has also increased. We estimate that full exploitation of oil sands deposits will have profound environmental consequences, with the greatest impacts by far expected in Canada and Venezuela. Specifically, cumulative forest loss across these studied oil sand deposits may eventually reach 8,665 km2, an area equal to 5% of global 3 -1 forest loss in the year 2014. In addition, fresh water use may nearly triple to 1.31 km H2O yr with important impacts on the local freshwater resources, as the case of Utah where the deposits are located in a potentially water-stressed areas. Moreover, projected annual GHG emissions (383 Mtonne CO2 eq yr-1) – due in large part to growth in energy-intensive in situ production – would be commensurate with those from land use and land cover change for all of Indonesia, and as many as many as 640,000 people could potentially be affected by the complete development of these five world’s oil sands deposits. While multiple environmental impacts are apparent, there have however been substantial economic benefits with affordable and secure energy to U.S. and Canadian market, more than 478,000 jobs created in Canada in year 2012 (3% of all jobs in the country) (PetroLMI, 2016), and tax revenues

and royalties paid to the governments. Thus there are clear and ongoing tradeoffs between economic development, energy, and the environment. The development of oil sand extraction in countries with dry climates is expected to exert a stronger pressure on the local water resources, likely competing with other water uses such as crop production and flows for aquatic ecosystems. Thus the increasing reliance on oil sand extraction may raise new water security concerns in countries where deposits are located in water stressed watersheds (Utah), thereby further sustaining ongoing debates on competing water uses within the context of the water-energy nexus (Rulli, 2016).

This study demonstrates that mining operations and in situ drilling differ significantly in their environmental impacts. While in-situ technology is more energy intensive, the process requires less water, does not produce tailings, and has less extensive impacts on overall forest cover. For in situ concessions, forest loss tends to be more scattered, limited to areas of intensive exploration activities for the positioning of the extraction wells, and construction of infrastructures (e.g., roads, pipelines). Even though in-situ drilling entails a smaller net change in forest cover than surface mining, its impact on wildlife habitat and landscape fragmentation should not be underappreciated. Conversely, forest vegetation is completely cleared within mining operations. In addition, while oil sand extraction is not expected to displace a large number of people – as these deposits are located in remote forested areas – these operations are likely to have important impacts on carbon sequestration, because of carbon emissions associated with land use change and deforestation. Progress in technology, proximity to the markets and high prices of crude oil have favored the proliferation of the oil sand industry in Alberta. Though this study has focused on the major environmental impacts of oil sand extraction, this source of energy also offers some advantages compared to conventional oil. First, the recovery rate of bitumen from oil sands is greater than for conventional oil. Second, oil sand deposits decline more slowly (4% decline per year) than those of conventional oil (20% per year), allowing these deposits to last longer (e.g., 30 years in the case of Alberta) (Oil sands magazine, 2016). Because of their long lifespan, oil sand deposits are drawing increasing interest and investments from oil and gas corporations. Moreover, oil sand industries have led to considerable job creation. For example, in 2012, the oil sands industry in Canada and U.S. employed – directly and indirectly – almost 558,000 people (80,000 in U.S.) (IHS CERA, 2014; Canadian Energy Research Institute, 2011). Employment is expected to grow in the most positive scenario of oil sands development to 2.2 million jobs in the U.S. and Canada by 2035 (Canadian Energy Research Institute, 2011As a result of this oil sands ‘boom’, Alberta has become one of the richest regions in the world, with a GDP per capita in year

2014 among the top five countries in the world (Alberta Government, 2016; The World Bank, 2014). Thus, pairing oil sands extraction with responsible policies can ultimately enhance economic growth and create employment opportunities. Yet despite these obvious benefits, our study clearly demonstrates that environmental considerations need to be incorporated into decision-making surrounding oil sands.

DISCUSSION Potential exploitation of unconventional fossil fuels will create a new geography of oil and natural gas production, with important implications for the global geopolitical landscape. Their projected growth by 2040 will meet 30% of world natural gas demand and 18% of global oil production. Shale resources and oil sands are an opportunity for some countries to increase their energy security, while reducing costs of fossil fuel imports and potentially changing their import-export balance. The high amount of water required to extract these resources requires that countries that decide to exploit their unconventional resources will need to develop responsible water management plans to ensure that other sectors are not impacted.

The shale revolution and oil sands boom have created million of new jobs and economic benefits in North America, supporting economic growth also in some rural and less developed areas. Thus, with adequate policies and regulations, unconventional fossil fuels extraction has the potential to enhance the economic growth and energy security of some regions.

Despite these benefits, in many regions of the world unconventional fossil fuels development will be problematic because of water limitations and will likely exacerbate a competition with water for food and other human needs. Particularly critical appears to be the case of some high quality shale areas in water stressed regions of the United States, Mexico, South Africa, China, South Asia, and Australia and in the United States (Utah deposit) for the case of oil sands. In some of these regions oil and gas production from unconventional fossil fuels is expected to threaten the local water and food security. In these water stressed areas where shale resources are present adequate policies need to be put in place in order to avert social, economic, and ecological consequences.