Contents

• Introduction
• Scope and Purpose of This Report
• Life-Cycle Assessment Methodology
• Key Factors in the Assessment of Canadian Oil Sands Crudes
• Results of Selected Life-Cycle Emissions Assessments
• Life-Cycle Assessments of Canadian Oil Sands Crudes
• Findings
• Design Factors and Input Assumptions for Life-Cycle Assessments of Canadian Oil Sands Crudes
• Life-Cycle Assessments of Canadian Oil Sands Crudes versus Other Reference Crudes
• Findings
• Design Factors and Input Assumptions for Life-Cycle Assessments of Reference Crudes
• Life-Cycle Assessments of Canadian Oil Sands Crudes versus Other Fuel Resources
• Further Considerations

Summary

Canadian Oil Sands and Climate Change

Recent congressional interest in U.S. energy policy has focused in part on ways through which the United States could secure more economical and reliable petroleum resources both domestically and internationally. Many forecasters identify petroleum products refined from Canadian oil sands as one possible solution. Increased production from Canadian oil sands, however, is not without controversy, as many have expressed concern over the potential environmental impacts. These impacts include emissions of greenhouse gases (GHG) during resource extraction and processing. A number of key studies in recent literature have expressed findings that GHG emissions per unit of energy produced from Canadian oil sands crudes are higher than those of other crudes imported, refined, and consumed in the United States. The studies identify two main reasons for the difference: (1) oil sands are heavier and more viscous than lighter crude oil types on average, and thus require more energy- and resource-intensive activities to extract; and (2) oil sands are chemically deficient in hydrogen, and have a higher carbon, sulfur, and heavy metal content than lighter crude oil types on average, and thus require more processing to yield consumable fuels by U.S. standards.

Selected Findings from the Primary Published Studies

CRS surveyed the published literature, including the U.S. State Department-commissioned studies for the Keystone XL pipeline project in both the 2011 Final Environmental Impact Statement and the 2014 Final Supplementary Environmental Impact Statement. The primary literature reveals the following:

• Canadian oil sands crudes are generally more GHG emission-intensive than other crudes they may displace in U.S. refineries, and emit an estimated 17% more GHGs on a life-cycle basis than the average barrel of crude oil refined in the United States;
• compared to selected imports, Well-to-Wheels GHG emissions for Canadian oil sands crudes range from 9% to 19% more emission-intensive than Middle Eastern Sour, 5% to 13% more emission-intensive than Mexican Maya, and 2% to 18% more emission-intensive than various Venezuelan crudes;
• compared to selected energy- and resource-intensive crudes, Well-to-Wheels GHG emissions for Canadian oil sands crudes are within range of heavier crudes such as Venezuelan Bachaquero and Californian Kern River, as well as lighter crudes that are produced from operations that flare associated gas (e.g., Nigerian Bonny Light);
• discounting the final consumption phase of the life-cycle assessment (which can contribute up to 70%-80% of Well-to-Wheels emissions), Well-to-Tank (i.e., "production") GHG emissions for Canadian oil sands crudes are 9%-102% higher than for selected imports;
• the estimated effect of the Keystone XL pipeline on global GHG emissions remains uncertain, as some speculate that its construction would encourage an expansion of oil sands investment and development, while others suggest that the project would not substantially influence either the rate or magnitude of oil extraction activities in Canada or the overall volume of crude oil transported to and refined in the United States.

Scope and Purpose of This Report

Congressional interest in the GHG emissions attributable to Canadian oil sands crudes has encompassed both a broad understanding of the resource as well as a specific assessment of the proposed Keystone XL pipeline. This report focuses on the broader resource. It discusses the methodology of life-cycle assessment and compares several of the publicly available studies of GHG emissions data for Canadian oil sands crudes against each other and against those of other global reference crudes. For a detailed analysis of the GHG emissions attributable to the proposed Keystone XL pipeline, and the findings from the State Department's Final Environmental Impact Statement, see CRS Report R43415, Keystone XL: Greenhouse Gas Emissions Assessments in the Final Environmental Impact Statement.

Introduction

Recent congressional interest in U.S. energy policy has focused in part on ways through which the United States could secure more economical and reliable petroleum resources both domestically and internationally. Many forecasters identify petroleum products refined from Canadian oil sands¹ as one possible solution. Canadian oil sands account for about 54% of Canada's total crude oil production, and that number is expected to rise from its current level of 1.9 million barrels per day (mbd) in 2013 to 4.8 mbd by 2030.² Further, the infrastructure to produce, upgrade, refine, and transport the resource from Canadian oil sands reserves to the United States is in place, and additional infrastructure projects—such as the Keystone XL pipeline—have been proposed.³ Increased production from Canadian oil sands, however, is not without controversy, as many have expressed concern over the potential environmental impacts. These impacts may include increased water and natural gas use, disturbance of mined land, effects on wildlife and water quality, trans-boundary air pollution, and emissions of greenhouse gases (GHG) during extraction and processing.⁴

A number of key studies in recent literature have expressed findings that GHG emissions per unit of energy produced from Canadian oil sands crudes are higher than those of other crudes imported, refined, and consumed in the United States.⁵ While GHG emissions and other air quality issues originating in the upstream sectors of Canada's petroleum industry may not directly impact U.S. National Emissions Inventories or U.S. GHG reporting per se, many environmental stakeholders and policy makers have noted that the increased use of more emission-intensive resources in the United States may have negative consequences for both U.S. and global energy and environmental policy.

The U.S. Department of State (DOS), in response to comments on the 2010 Draft Supplementary Environmental Impact Statement (2010 Draft EIS)⁶ for the Keystone XL pipeline project (which would connect oil sands production facilities in the Western Canadian Sedimentary Basin with refinery facilities in the United States), commissioned a contractor to analyze the life-cycle GHG emissions associated with these resources in comparison to other reference crudes.⁷ DOS presented this analysis in the 2011 Final Environmental Impact Statement (2011 Final EIS) released on August 26, 2011, as a "matter of policy," but noted that neither the National Environmental Policy Act (NEPA) nor DOS regulations (22 C.F.R. 161.12) nor Executive Orders 13337 and 12114 (Environmental Effects Abroad of Major Federal Activities) legally require that an EIS include an assessment of environmental activities outside the United States.

The initial permit for the Keystone XL Project was denied due to insufficient time to prepare a rigorous, thorough, and transparent review of the pipeline's proposed routes through Nebraska. In May 2012, Keystone filed a new permit application for a revised route, implementing a new national interest determination. In accordance with this process, DOS released a revised Draft Supplementary EIS (Draft EIS) for the revised project on March 1, 2013, and a revised Final Supplementary EIS (Final EIS) on January 31, 2014, including an assessment of the indirect cumulative impacts and life-cycle GHG emissions of Canadian oil sands crudes.⁸ While DOS commissioned a different contractor to assist with the EIS,⁹ the data used to determine the GHG life-cycle emissions associated with the resource, as well as the market analysis used for supply and demand projections, remained largely unchanged. Hence, the 2014 Final EIS made similar findings to the 2011 Final EIS, including the following:

• 1. Canadian oil sands crudes "are more GHG-intensive than the other heavy crudes they would replace or displace in U.S. refineries, and emit an estimated 17% more GHGs on a life-cycle basis than the average barrel of crude oil refined in the United States in 2005,"¹⁰ and
• 2. "Approval or denial of any one crude oil transport project, including the proposed Project, is unlikely to significantly impact the rate of extraction in the oil sands or the continued demand for heavy crude oil at refineries in the United States."¹¹

Opponents of the pipeline, however, are critical of this impact assessment. They contend that the lack of transport infrastructure and the price discount it occasions has already affected production of the oil sands and, if continued, would further depress investment and development in the region.¹²

Scope and Purpose of This Report

This report presents a summary of life-cycle emissions assessments of Canadian oil sands crudes and provides an analysis of their respective findings. The first section of the report, "Life-Cycle Assessment Methodology," discusses the basic methodology of life-cycle assessments and examines the choice of boundaries, design features, and input assumptions. The second section of the report, "Results of Selected Life-Cycle Emissions Assessments," compares several of the publicly available assessments of life-cycle GHG emissions data for Canadian oil sands crudes against each other, against those of other global reference crudes, and against those of other fossil fuel resources. The report concludes with a discussion of some tools for policy makers who are interested in using these assessments to investigate the potential impacts of U.S. energy policy choices on the environment. For a specific analysis of the GHG emissions attributable to the proposed Keystone XL pipeline, see CRS Report R43415, Keystone XL: Greenhouse Gas Emissions Assessments in the Final Environmental Impact Statement, by [author name scrubbed].

Life-Cycle Assessment Methodology

Life-cycle assessment (LCA) is an analytic method used for evaluating and comparing the environmental impacts of various products (in this case, the climate change implications of hydrocarbon resources). LCAs can be used in this way to identify, quantify, and track emissions of carbon dioxide and other GHG emissions arising from the development of these hydrocarbon resources, and to express them in a single, universal metric of carbon dioxide equivalent (CO₂e) GHG emissions per unit of fuel or fuel use.¹³ This figure is commonly referred to as the "emissions intensity" of the fuel. The results of an LCA can be used to evaluate the GHG emissions intensity of various stages of the fuel's life cycle, as well as to compare the emissions intensity of one type of fuel or method of production to another.

GHG emissions profiles modeled by most LCAs are based on a set of boundaries commonly referred to as "cradle-to-grave," or, in the case of transportation fuels such as petroleum, "Well-to-Wheels" (WTW). WTW assessments for petroleum-based transportation fuels focus on the emissions associated with the entire life cycle of the fuel, from extraction, transport, and refining of crude oil; to the distribution of refined product (e.g., gasoline, diesel, jet fuel) to retail markets; to the combustion of the fuel in end-use vehicles. Other LCAs (e.g., Well-to-Tank [WTT] or Well-to-Refinery Gate [WTR]) establish different (i.e., more specific) life-cycle boundaries to evaluate emissions (see Figure 1). Inclusion of the final combustion phase allows for the most complete picture of petroleum's impact on GHG emissions, as this phase can contribute up to 70%-80% of WTW emissions. However, other LCAs can be used to highlight the differences in emissions associated with particular stages as well as experiment with certain boundary assumptions. The choice of boundaries is an important component to any LCA and can lead to vastly differing reported results.¹⁴

Figure 1. Crude Oil Life-Cycle Schematic

Source: CRS.

Key Factors in the Assessment of Canadian Oil Sands Crudes

Because of the complex life cycle of hydrocarbon fuels and the large number of analytical design features that are needed to model their emissions, LCAs must negotiate many variables and uncertainties in available data. Key factors that influence the results of an LCA include (1) composition of the resource that is modeled, (2) extraction process of the resource that is modeled, (3) design factors chosen for the assessment, and (4) assumptions made in the input data for the assessment. Some of these factors with respect to Canadian oil sands crudes are as follows:

Crude Oil Types. Oil sands are a type of unconventional petroleum deposit. They are commonly formations of loose sand or consolidated sandstone containing naturally occurring mixtures of sand, clay, and water, as well as a dense and extremely viscous form of petroleum technically referred to as bitumen.¹⁵ Most LCAs do not include an assessment of raw bitumen, because it is near solid at ambient temperature and cannot be transported in pipelines or processed in conventional refineries. Thus, bitumen is often diluted with liquid hydrocarbons or converted into a synthetic light crude oil to produce the resource known as "oil sands-derived crude" or simply "oil sands crude." Several kinds of crude-like products can be generated from bitumen, and their properties differ in some respects from conventional light crude. They include the following:

• Upgraded Bitumen, or Synthetic Crude Oil (SCO). SCO is produced from bitumen through an upgrading process that turns the very heavy hydrocarbons into lighter fractions. Since the upgrading process begins at the production facility for SCO, the allocation of GHG emissions is weighted more heavily upstream than other crude types.
• Diluted Bitumen (Dilbit). Dilbit is bitumen mixed with diluents—typically natural gas liquids such as condensate—to create a lighter, less viscous, and more easily transportable product. Mixing bitumen with less carbon-intensive diluents lessens the GHG emissions impact per barrel of dilbit in relation to bitumen or SCO. Some refineries need modifications to process large quantities of dilbit feedstock, since it requires more heavy oil conversion capacity than conventional crudes. Increased processing in refineries shifts GHG emissions downstream, potentially intensifying the downstream GHG emission impact of dilbit in relation to SCO or other crudes (e.g., if dilbit is transported from Canada to the United States via a pipeline, the need for increased refining downstream would shift the potential for emissions to the United States).
• Synthetic Bitumen (Synbit). Synbit is typically a combination of bitumen and SCO. The properties of each kind of synbit blend vary significantly, but blending the lighter SCO with the heavier bitumen results in a product that more closely resembles conventional crude oil. Refining emissions from synbit occur both upstream and downstream, depending upon a variety of factors.

Extraction Process. Two types of methods for extracting bitumen from the reservoir are currently used in the Canadian oil sands. They include the following:

• Mining. Oil sands deposits that are less than approximately 75 meters below the surface can readily be removed using conventional strip-mining methods. An estimated 20% of currently recoverable reserves are close enough to be mined. The strip-mining process includes removal of the overburden (i.e., primary soils and vegetation), excavation of the resource, and transportation to a processing facility. Higher intensities of GHG emissions may result from increased land use changes during strip-mining. Mining accounts for slightly more than 50% of current production, and is expected to remain between 40% and 50% through 2030.¹⁶ Currently, a significant portion of mined bitumen is upgraded to SCO.
• In Situ. Oil sands deposits that are deeper than approximately 75 meters are recovered using in situ methods. Most in situ recovery methods currently in operation involve injecting steam into an oil sands reservoir to heat—and thus decrease the viscosity of—the bitumen, enabling it to flow out of the reservoir to collection wells. Steam is injected using cyclic steam stimulation (CSS), where the same well cycles both the steam and the bitumen, or by steam-assisted gravity drainage (SAGD), where a top well is used for steam injection and the bottom well is used for bitumen recovery. Because significant amounts of energy are currently required to create steam, in situ methods are generally more GHG-intensive than conventional mining (excluding land use impacts). With more than 80% of recoverable reserves situated too deep for conventional mining techniques, it is assumed that the industry will eventually move toward an increased use of the in situ extraction process in some form.

Study Design Factors. Design factors relate to how the GHG comparison is structured in each study and which parameters are included. These factors may include

• overall purpose and goal of the study,
• time frame for the inputs and the results,
• life-cycle boundaries that are established for comparison,
• units and metrics used for comparison,
• GHG global-warming potential used for comparison,¹⁷
• treatment of co-products during refining (e.g., asphalt, petroleum coke, liquid gases, lubricants),
• treatment of secondary emission flows (e.g., capital infrastructure, land-use changes),¹⁸
• treatment of power co-generation at the facilities, and
• treatment of flaring, venting, and fugitive emissions.

Input Assumptions. Input assumptions can impact life-cycle results at each stage of the assessment. Studies often use simplified assumptions to model GHG emissions due to limited data availability and the complexity of and variability in the practices used to extract, process, refine, and transport crude oil, diluted crude, or refined product. Key input assumptions for Canadian oil sands crudes may include

• percentage contribution of each type of crude and each type of extraction process in the final transported product;
• type of upgrading or refining processes;
• amount of petroleum coke produced, stored, combusted, or sold;
• ratios for bitumen-to-diluents, steam-to-oil, gas-to-oil, water-to-oil; and
• energy efficiencies for steam generation and other production processes.

Results of Selected Life-Cycle Emissions Assessments

Life-Cycle Assessments of Canadian Oil Sands Crudes

Greenhouse gases, primarily in the form of carbon dioxide and methane, are emitted during a variety of stages in oil sands production (see text box below).¹⁹ A number of published and publicly available studies have attempted to assess the life-cycle GHG emissions data for Canadian oil sands crudes. This report examines the life-cycle assessments analyzed by the U.S. Department of State (DOS)—in conjunction with the consultancy firm ICF International LLC (ICF)—in the Keystone XL Project's August 2011 Final Environmental Impact Statement (2011 Final EIS). The studies were selected by ICF using several criteria: (1) they evaluated Canadian oil sands crudes in comparison to other reference crude oils, (2) they focused on GHG emissions impacts throughout the entire crude oil life-cycle, (3) they were published within the past 10 years, and (4) they represented the perspectives of a range of stakeholders. The use of these studies was replicated in the 2014 Final Environmental Impact Statement (2014 Final EIS) conducted by DOS and the contractor Environmental Resources Management.

Table 1 provides a list of the studies referenced by the DOS analysis. While the type, boundaries, and design features vary across all studies, DOS determined the data and results from AERI/Jacobs 2009, AERI/TIAX 2009, NETL 2008, and NETL 2009 to be sufficiently robust for inclusion in the 2011 Final EIS as well as the 2014 Final EIS. Reasons against the inclusion of the remaining studies are presented briefly in the table, and outlined in more detail in the EIS.

The 2014 Final EIS mentioned several other studies published after the release of the 2011 Final EIS. These studies include Jacobs Consultancy, EU Pathway Study: Life Cycle Assessment of Crude Oils in a European Context, 2012; IHS CERA, Oil Sands, Greenhouse Gases, and U.S. Oil Supply Getting the Numbers Right—2012 Update; Adam Brandt, Upstream GHG Emissions from Canadian Oil Sands as a Feedstock for European Refineries, 2011; and Joule Bergerson et al., Life Cycle Greenhouse Gas Emissions of Current Oil Sands Technologies: Surface Mining and In Situ Applications, 2012. The Final EIS, however, retained a focus on the data and results from AERI/Jacobs 2009, AERI/TIAX 2009, NETL 2008, and NETL 2009.

Findings

The primary studies—as well as the DOS/ICF meta-analysis—report the following findings:

• Comparisons across the published studies of GHG life-cycle emissions intensities for fuels derived from different sources are sensitive to each study's choice of boundaries and input parameters.
• As reported in the studies, Well-to-Wheels GHG emissions for the full range of Canadian oil sands crudes and production processes are valued between 101-120 gCO₂e/MJ lower heating value (LHV)²⁰ gasoline.
• As reported in NETL 2008, Well-to-Wheels GHG emissions for a select range of Canadian oil sands crudes and production processes are valued between 101-110 gCO₂e/MJ LHV gasoline.
• As reported in NETL 2008, Well-to-Wheels GHG emissions for the weighted average²¹ of transportation fuels sold or distributed in the United States (in reference year 2005) are valued at 91 gCO₂e/megajoule (MJ) LHV gasoline.²²
• As reported in NETL 2008, Canadian oil sands crudes emit an estimated 17% more GHGs on a life-cycle basis than the weighted average of transportation fuels sold or distributed in the United States (in reference year 2005).
• As reported in NETL 2008, discounting the final consumption phase of the life-cycle assessment (which can contribute up to 70%-80% of Well-to-Wheels emissions), Canadian oil sands emit an estimated 80% more GHGs on a Well-to-Tank (i.e., "production") basis than the weighted average of transportation fuels sold or distributed in the United States (in reference year 2005).

These numbers serve as averages, and are intended to reflect the range of estimates from the primary studies. Conversely, individual estimates reported by each of the studies listed in Table 1—both primary and secondary—for various Canadian oil sands crude types and production processes can be found in Figure 2 and Table 2.

Figure 2 illustrates the WTW GHG emissions estimates as reported by each of the studies for various Canadian oil sands crude types and production processes. Table 2 summarizes and compares each study's emissions estimates, data, and relevant input assumptions. Variability among the estimates is, in part, the result of each study's differing design and input assumptions. A discussion of these assumptions—and their estimated impacts on GHG emissions—follows in the next section.

Several life-cycle GHG emissions assessments have been published since the release of the 2011 Final EIS. These studies include Jacobs 2012, IHS CERA 2012, Brandt 2011, and Bergerson 2012, among others. IHS CERA 2012 found that transportation fuels produced from oil sands crudes result in average WTW GHG emissions that are 14% higher than the average crude refined in the United States (results range from 5%-23% higher). Jacobs 2012 found that WTW GHG intensities of transportation fuels produced from oil sands crudes are within 7%-12% of the "upper range" of the WTW intensity of conventional crudes. Bergerson 2012 reported that "although a high degree of variability exists in Well-to-Wheels emissions due to differences in technologies employed, operating conditions, and product characteristics, the surface mining dilbit and the in situ SCO pathways have the lowest and highest emissions, 88 and 120 g CO2eq/MJ reformulated gasoline," and that the lower values for certain oil sands production activities "overlap with emissions in literature for conventional crude oil."

Figure 2. Well-to-Wheels GHG Emissions Estimates for Canadian Oil Sands Crudes

Source: CRS, from studies outlined in Table 1. Average U.S. petroleum baseline for 2005 provided by U.S. Environmental Protection Agency (U.S. EPA), Renewable Fuel Standard Program (RFS2): Regulatory Impact Analysis, February 2010, EPA-420-R-10-006, with data sourced from DOE/NETL, Development of Baseline Data and Analysis of Life Cycle GHG Emissions of Petroleum Based Fuels, November 2008. Notes: See section "Life-Cycle Assessment Methodology" for key to crude oil types and production processes. U.S. EPA 2005 (U.S. Average) assesses "the average life cycle GHG profile for transportation fuels sold or distributed in the United States in 2005 [and] is determined based on the weighted average of fuels produced in the U.S. plus fuels imported into the U.S. minus fuels produced in the U.S. but exported to other countries for use" (NETL 2008, p. ES-5). This baseline includes Canadian oil sands crudes, but does not include emissions from some of the most carbon-intensive imported crude oils (e.g., Venezuelan Heavy) due to modeling uncertainties (NETL 2008, p. ES-7; NETL 2009, p. ES-2). The baseline number is internally consistent only with the other NETL findings reported in the figure.

Design Factors and Input Assumptions for Life-Cycle Assessments of Canadian Oil Sands Crudes

Most published and publicly available studies on the life-cycle GHG emissions data for Canadian oil sands crudes identify two main factors contributing to the difference in emissions intensity relative to other reference crudes:

• 1. oil sands are heavier and more viscous than lighter crude oil types on average, and thus require more energy- and resource-intensive activities to extract; and
• 2. oil sands are chemically deficient in hydrogen, and have a higher carbon, sulfur, and heavy metal content than lighter crude oil types on average, and thus require more processing to yield consumable fuels by U.S. standards.

While most studies agree that Canadian oil sands crudes are, on average, more GHG-intensive than the crudes they may displace in the U.S. refineries, the range of the reported increase varies among assessments. Key design and input assumptions can significantly influence results. These factors include, but are not limited to, the following:

• Metrics. Comparing results from various studies is complicated by each study's choice of functional units. While GHG emissions have been normalized by most studies and reported as CO₂-equivalents, the units they are expressed "over" vary greatly. Some evaluate GHG emissions on the basis of a particular final fuel product (e.g., gasoline, diesel, or jet fuel). Others evaluate emissions by an averaged barrel of refined product. Some studies report emissions per unit of volume (e.g., millions of barrels [mbl]), and others by unit of energy produced (e.g., British Thermal Units [Btus] or megajoules [MJ]). For example, NETL 2008, Jacobs 2009, and TIEX 2009 use functional units for energy produced across the final products—MMBtus or MJs for gasoline, diesel, and/or jet fuel. IHS CERA 2010 expresses GHG emissions "per barrel of refined product produced"; while others, like Charpentier 2009 (not included in the reported findings), by "kilometers driven," among others. The choice affects how the results are presented and makes it challenging to compare across studies if the data or conversion values are not fully published or transparent.
• Extraction Process. GHG emissions vary by the type of extraction process used to recover bitumen. Due to the high energy demands of steam production, in situ methods are generally assumed to be more GHG-intensive than mining operations. However, not all studies assess the difference to be the same. IHS CERA 2010 estimates the increase of WTW GHG emissions from in situ extraction to be, on average, 7% greater than mining. NRDC 2010 estimates 9%. Specific estimates in Jacobs 2009 show a 4% increase (for SAGD dilbit over mining dilbit) and in NRCan 2008 an increase of 9% (for SAGD SCO over mining SCO).
• In Situ Steam-to-Oil (SOR) Ratio. The amount of steam injected into a reservoir during in situ processes to extract a unit volume of bitumen varies across reservoirs and across extraction facilities. The resulting energy consumption and GHG emissions estimates vary accordingly. Thus, the figure used in LCAs to express this ratio may significantly impact GHG estimates. NRCan 2008 reports SOR values from 2.5 to 5.0 across SAGD operations in Canadian oil sands. NRDC 2010 reports a range from 1.94 to 7.26. IHS CERA cites an industry average of 3. Charpentier 2009 demonstrates that GHG emissions at the production phase are very sensitive to SOR, estimating that every 0.5 increase in the ratio corresponds to an increase of 10 kgCO₂e GHG emissions per barrel of bitumen produced.
• Upgrading Process. Bitumen needs pre-processing in order to lower its viscosity and remove impurities before it is fit for conventional refineries. This pre-processing is called "upgrading," the key components of which include (1) removal of water, sand, physical waste, and lighter products; (2) catalytic purification (i.e., the process of removing excess sulfur, oxygen, nitrogen, and metals); and (3) hydrogenation through either carbon rejection or catalytic hydrocracking (i.e., the process of removing or breaking down the heaviest fraction of the oil residuum by either vacuum distillation and precipitation or by adding hydrogen in a "hydrocracking process that breaks long-chain hydrocarbons into shorter, more useful ones). The residuum can be further refined in a "coking" process to produce gasoline, distillate, and petroleum coke. The resulting product is synthetic crude oil (SCO) and numerous co-products, including water, sand, waste, sulfur, oxygen, nitrogen, distillate, and petroleum coke, among others. Some of the co-products from the upgrading process contain carbon and other potential GHG emission sources. Thus, a consistent and comprehensive accounting of the GHG emission from all co-products would be necessary for a full life-cycle assessment of oil sands crudes—or any hydrocarbon—production.
• Treatment of Petroleum Coke. Petroleum coke (a source of excess carbon) is a co-product of bitumen production at both the upgrader and the refinery. Roughly 5%-10% of a barrel of crude ends up as coke; and the heavier the crude, the greater the percentage of coke. Bitumen refining can produce about 50% more coke than the average conventional crude. The treatment of coke is a primary driver behind the results of any WTW GHG oil sands crudes assessment. If coke is combusted (i.e., for process heat, electricity, or hydrogen production at the upgrader in lieu of natural gas combustion), WTW GHG emissions may increase anywhere from 14% (TIAX 2009) to 50% (McCulloch 2006) over lighter crudes. If it is stored, sold, and/or combusted elsewhere, its potential emissions may not be factored into the LCA. The main concern for modeling is ensuring that coke produced at the upgrader (for SCO) is treated consistently with coke produced at the refinery (for dilbit or other imported crudes). Based on the studies analyzed in this report, petroleum coke at the upgrader is either (1) consumed (for process heat, electricity, or hydrogen production); (2) stored; or (3) sold as a fuel for combustion. In contrast, the studies assume that petroleum coke produced at the refinery that is not consumed by the refinery itself is either (1) used to back out coal combustion for electricity generation; or (2) allocated outside of the LCA.²³ These inconsistent methodologies make comparisons problematic. Coke produced at U.S. refineries has a low domestic demand, and is therefore often shipped to overseas markets for use as a replacement fuel for coal combustion or steel production (most studies include neither the overseas transportation nor the incremental combustion emissions of coke in WTW GHG emissions assessments).
• Cogeneration. Cogeneration facilities use both steam and electricity generated from the steam to achieve higher energy efficiencies. In situ extraction facilities often have steam requirements much greater than electricity requirements, thus leaving excess capacity for electricity generation that can be exported back into the grid for use elsewhere. Offset credits given to exported electricity in LCAs can have a substantial impact on WTW GHG emissions. Cogeneration assumptions vary across the studies of Canadian oil sands crudes in two ways: (1) whether cogeneration credits are included, and (2) if so, what source of electricity is offset (e.g., coal-fired generation, oil, or natural gas). Some estimates show that applying credits from oil sands facilities to offset coal-fired electricity generation could reduce WTW GHG emission to within the range of conventional crudes. Many studies currently do not consider offset credits because the practice is not in widespread use among producers.
• Upgrading and Refinery Emissions. Because SCO delivered to a refinery has already been processed at the upgrader, the energy consumption at the refinery—and therefore the GHG emissions at the refinery—may be lower than the refinery emissions of dilbit or other crudes. Accounting for the reduced emissions from SCO has a modest effect on WTW GHG emissions, as refinery emissions are commonly around 5%-15% of the total. Many studies do not mention this accounting, and it is unclear if the reductions for SCO at the refinery are incorporated into many of the LCAs.
• Diluents. Because the viscosity of raw bitumen is too high to be transported via pipeline, diluting bitumen with lighter hydrocarbons to assist in its transport has become a common industry practice. Accounting for the effects of diluting bitumen is an important component in emission estimates, because producing and refining the diluents into finished products may result in a lower WTW GHG emissions estimate per barrel of dilbit in comparison to a barrel of raw bitumen. LCAs that report emissions for dilbit on a per barrel of refined product basis are thus reporting the emissions from a combination of both oil sands bitumen and the supplemental hydrocarbons. Additionally, diluting raw bitumen with light hydrocarbons creates a crude product that is more difficult and energy-intensive to refine than other crude oils, thus producing less premium refined product per barrel after the refinery stage.²⁴ The extent to which this difference in yield is accounted for across the various studies is unclear. The IHS CERA 2010 estimates for crude production of SAGD dilbit do not show an adjustment for the difference. TIAX 2009 and Jacobs 2009 both show slightly higher refinery emissions for dilbit compared to other crudes, but the reasons for the increase are not specified.
• Upstream Production Fuels. Some studies include the GHG emissions associated with the upstream production of purchased electricity that is imported to provide process heat and to power machinery throughout crude production. The upstream GHG emissions for natural gas fuel and electricity generation used in the production of oil sands crudes can be significant. Jacobs 2009 demonstrates that the GHG emissions associated with the upstream fuel cycle account for roughly 4%-5% of the total WTW GHG emissions for average Canadian oil sands crudes. IHS CERA 2010 indicates that although its study excludes upstream fuel and electricity GHG emissions, the inclusion of them would add 3% to WTW GHG emissions per barrel of refined product.
• Flared, Vented, and Fugitive Emissions. Emissions associated with flaring and venting can be a significant source of GHG emissions. The TIAX 2009 study indicates that including venting and flaring emissions associated with oil sands production (particularly for mining extraction techniques) may contribute up to 4% of total WTW GHG emissions. Further, methane emissions from fugitive leaks throughout the oil sands production process can potentially contribute up to 1% of GHG emissions.²⁵ Methane emissions from oil sands mining and tailings ponds may have an even larger impact, contributing from 0% to 9% of total GHG emissions.²⁶ TIAX 2009, McCulloch 2006, and NRCan 2008 state that they include emissions from these sources. IHS CERA 2010 excludes emissions from methane released from tailings ponds but recognizes there is considerable uncertainty and variance in quantifying these emissions. Other studies do not specify.
• Infrastructure/Construction Emissions. None of the existing studies include the GHG impacts associated with capital equipment and the construction of facilities, machinery, and infrastructure needed to produce oil sands. Charpentier 2009 discusses the need to more fully investigate and include these potentially significant supply chain infrastructure GHG emissions in future life-cycle studies of oil sands crudes.
• Local and Indirect Land-Use Change Emissions. Emissions associated with changes in biological carbon stocks from the removal of vegetation, trees, and soil during oil sands mining operations may be significant, albeit temporary in some cases, and highly dependent upon the reclamation activities employed after use. Yeh 2010 estimates that surface mining of oil sands results in a 0.9%-2.5% increase in the WTW emissions versus the baseline (2005 U.S. average). The range was dependent on the type of lands displaced, with the removal of peatland having the largest impact and certain in situ facilities having the least impact. None of the life-cycle studies reviewed, however, includes land-use change GHG emissions in the WTW life-cycle assessment.²⁷ Some recent studies, including the 2014 Final EIS, have begun to assess the effects.²⁸

Life-Cycle Assessments of Canadian Oil Sands Crudes versus Other Reference Crudes

To compare the life-cycle GHG emissions intensities from Canadian oil sands crudes against those of other crude oils imported into the United States, many of the published studies conduct reference assessments of other global resources.

Findings

Figure 3 presents the results of one of the more comprehensive studies (NETL 2009), which compared Well-to-Wheels GHG emissions of reformulated gasoline across various crude oil feedstocks (a review of the NETL 2009 input assumptions is included in the figure's "Notes" section). NETL 2009 reported the following:

• Well-to-Wheels GHG emissions from gasoline produced from a weighted average of Canadian oil sands crudes imported to the United States are approximately 17% higher than those from gasoline derived from the weighted average of transportation fuels sold or distributed in the United States in the reference year 2005. This corresponds to an increase in Well-to-Tank (i.e., "production") GHG emissions of 80% over the weighted average of transportation fuels sold or distributed in the United States in the reference year 2005 (18 gCO₂e/MJ).
• Compared to a few selected imports, Well-to-Wheels GHG emissions from gasoline produced from a weighted average of Canadian oil sands crudes are 19%, 12%, and 18% higher than the life-cycle emissions from Middle Eastern Sour, Mexican Maya, and Venezuelan Conventional crudes, respectively.²⁹ This corresponds to an increase in Well-to-Tank (i.e., "production") GHG emissions of 102%, 53%, and 92% higher than the production emissions from Middle Eastern Sour, Mexican Maya, and Venezuelan Conventional crudes, respectively.
• Compared to selected energy- and resource-intensive crudes, Well-to-Wheels GHG emissions from gasoline produced from a weighted average of Canadian oil sands crudes are found to be "within range" of those produced from heavier crudes such as Venezuelan Bachaquero and Californian Kern River, as well as lighter crudes that are produced from operations that flare associated gas (e.g., Nigerian Bonny Light).

Individual estimates of WTW GHG emissions from Canadian oil sands crudes from the primary studies listed in Table 1 range from 9% to 19% more GHG-intensive than Middle Eastern Sour, 5% to 13% more GHG-intensive than Mexican Maya, and 2% to 18% more GHG-intensive than various Venezuelan crudes (including both Venezuelan Conventional and Bachaquero).

Design Factors and Input Assumptions for Life-Cycle Assessments of Reference Crudes

Similar to LCAs conducted on Canadian oil sands crudes, assessments of other global crude resources confront many variables and uncertainties in available data. Likewise, these assessments are bounded by specific design factors and input assumptions that can affect the quality of the results. Conditions that may impact the results include the following:

• Choice of Reference Crudes Studied. Crude oil resources around the world vary significantly in regard to resource quality and production methods. Thus, GHG emissions intensities may also vary significantly. The results of comparisons between Canadian oil sands crudes and other global crudes may depend on which crudes are used as a reference and/or which crudes are evaluated to determine a baseline. Some studies suggest that GHG emissions intensities of Canadian oil sands crudes should be measured against a global average in order to assess the full environmental impacts of the resource. Others believe they should be measured against an average of all crudes consumed in a given marketplace (e.g., a particular country or region, like the United States or the European Union). Still others argue that Canadian oil sands crudes should be measured against a representative basket of crudes they may displace in production (e.g., crudes potentially displaced by Canadian oil sands crudes at U.S. PADD III refineries; or, more specifically, only "heavy" crudes potentially displaced at U.S. PADD III refineries).³⁰
• Sensitivity to Water-Oil and Gas-Oil Ratios. Due to the complex nature of crude oil production systems and resource reservoirs, studies often use ratios to describe the fraction of the flow from a well that is oil, water, and gas. The use of ratios simplifies the relationship between energy use and GHG emissions and may fail to accurately report the variability across differing resources. Further, assumptions regarding venting or flaring of associated gas, and fugitive emissions from produced water, may further impact GHG emissions intensities.
• Transportation Emissions. Assumptions regarding how LCAs account for the contribution of transportation may impact WTW GHG emissions estimates to a small degree. These include the distance and mode of transportation from oil field to export terminal, and from producer to refiner, as well as the final transportation emissions of all co-products.
• Uncertainty Analysis. Accurately measuring GHG emissions intensities is highly uncertain. Few of the studies listed in Table 1 fully consider associated uncertainty, and none of them rigorously treat underlying uncertainties in data inputs and models. Most calculate averages from a wide range of values and develop point estimates without providing statistical bounds. These bounds may prove to be important if their ranges are shown to overlap with other results.
• Data Transparency. The quality of the data and the transparency in presentation vary considerably by study. Most studies do not provide complete transparency in their methodologies, assumptions, or data sources. This is partially a function of the difficulty in accessing necessary data elements from the field. Data on the Canadian oil sands are more robust than some global resources and less robust than others. Lack of transparency impedes the ability to make meaningful comparisons of the results for oil sands crudes and reference crudes.

Figure 3. Well-to-Wheels GHG Emissions Estimates for Global Crude Resources

Source: CRS, from NETL, An Evaluation of the Extraction, Transport and Refining of Imported Crude Oils and the Impact of Life Cycle Greenhouse Gas Emissions, National Energy Technology Laboratory, March 27, 2009. Notes: U.S. EPA 2005 (U.S. Average) assesses "the average life cycle GHG profile for transportation fuels sold or distributed in the United States in 2005 [and] is determined based on the weighted average of fuels produced in the U.S. plus fuels imported into the U.S. minus fuels produced in the U.S. but exported to other countries for use" (NETL 2008, p. ES-5). This baseline includes Canadian oil sands crudes, but does not include emissions from some of the most carbon-intensive imported crude oils (e.g., Venezuelan Heavy) due to modeling uncertainties (NETL 2008, p. ES-7; NETL 2009, p. ES-2). NETL values converted from kgCO2e/MMBtu using conversion factors of 1,055 MJ/MMBtu and 1,000 g/kg. NETL input assumptions are as follows: (1) assumes a weighted average of Canadian oil sands extraction at 43% raw bitumen (not accounting for blending with diluents to form dilbit) from CSS in situ production and 57% SCO from mining production in the years 2005 and 2006; (2) allocates refinery emissions from co-products other than the gasoline, diesel, and jet fuel to the co-products themselves, including petroleum coke, and thus outside the boundaries of the LCA (unless combusted at refinery); (3) uses linear relationships to relate GHG emissions from refining operations based on API gravity and sulfur content, thus failing to fully account for the various produced residuum ranges of bitumen blends and SCO; (4) does not fully evaluate the impact of pre-refining SCO at the upgrader prior to the refinery; (5) does not account for the transportation emissions of co-products; and (6) bounds the GHG emissions estimates for Venezuela's ultra-heavy oil/bitumen using uncertainty analysis due to the limited availability of public data. Further, as noted in Table 2, NETL 2009 study assumptions do not state SOR, do not include upstream fuel production, do not include infrastructure or land-use changes, and do not specify cogeneration, but do include emissions from venting, flaring, and fugitives.

Life-Cycle Assessments of Canadian Oil Sands Crudes versus Other Fuel Resources

Figure 4 offers a comparison of the life-cycle GHG emissions intensities of petroleum products from Canadian oil sands crudes with estimates from other unconventional petroleum products, natural gas, and coal. These data are drawn from several different studies employing many different design features and input assumptions, not the least of which are different methods of combusting the final fuel products. Further, it should be noted that different and non-substitutable end uses for the fuel products (e.g., the different end uses for coal and petroleum combustion) make a full comparison of their emissions impacts problematic. The figure presents an average value for each fuel; the original source materials provide a full description of each study's design characteristics as well as a presentation of each estimate's uncertainty analysis.

Figure 4. Life-Cycle GHG Emissions Estimates for Selected Fuel Resources

Source: CRS, from NETL, Development of Baseline Data and Assessment of Life Cycle Greenhouse Gas Emissions of Petroleum-Based Fuels, National Energy Technology Laboratory, November 26, 2008; Brandt, A.R. and A.E. Farrell, "Scraping the Bottom of the Barrel: Greenhouse Gas Emission Consequences of a Transition to Low-quality and Synthetic Petroleum Resources," Climatic Change, Vol. 84, 2007, pp. 241-263; and Burnham, A. et al., "Life-Cycle Greenhouse Gas Emissions of Shale Gas, Natural Gas, Coal, and Petroleum," Environmental Science and Technology, Vol. 46, 2012, pp. 619–627. Notes: NETL values converted from kgCO2e/MMBtu using conversion factors of 1,055 MJ/MMBtu and 1,000 g/kg; Brandt values converted from gCe/MJ using conversion factor of 3.667 Ce/CO2e.

Further Considerations

Life-cycle assessment has emerged as an influential methodology for collecting, analyzing, and comparing the GHG emissions and climate change implications of various hydrocarbon resources. However, because of the complex life cycle of hydrocarbon fuels and the large number of analytical design features that are needed to model their emissions, LCAs retain many uncertainties. These uncertainties often make comparing results across resources or production methods problematic. Hence, the usefulness of LCA as an analytical tool for policy makers may lie less in its capacity to generate comparative rankings, or "scores," between one source and another, and more in its ability to highlight "areas of concern," or "hot spots," in the production of a given hydrocarbon fuel. In this way, LCA can serve to direct policy makers' attention to those areas in resource development that present the greatest challenges to GHG emissions control, and hence, the biggest potential benefits if adequately managed.

Table 3 summarizes the GHG emissions impacts of the various stages of Canadian oil sands production and presents examples of mitigation strategies that have been offered by industry, academia, and other stakeholders.

Acknowledgments

Thanks to Amber Wilhelm of CRS for her help with graphics, and to Bryan Sinquefield of CRS for his help with editing.

Footnotes