AN ECONOMIC AND ENVIRONMENTAL ASSESSMENT OF EASTERN CANADIAN CRUDE OIL IMPORTS

Transcription

1 Study No. 167 CANADIAN ENERGY RESEARCH INSTITUTE AN ECONOMIC AND ENVIRONMENTAL ASSESSMENT OF EASTERN CANADIAN CRUDE OIL IMPORTS Canadian Energy Research Institute Relevant Independent Objective

2

3 AN ECONOMIC AND ENVIRONMENTAL ASSESSMENT OF EASTERN CANADIAN CRUDE OIL IMPORTS

4 ii Canadian Energy Research Institute An Economic and Environmental Assessment of Authors: Paul Kralovic Andrei Romaniuk Anna Vypovska Dinara Millington ISBN Copyright Canadian Energy Research Institute, 2018 Sections of this study may be reproduced in magazines and newspapers with acknowledgement to the Canadian Energy Research Institute Printed in Canada Front photo s courtesy of various Google searches Acknowledgements: The authors of this report would like to extend their thanks and sincere gratitude to the Canadian Fuels Association and ClipperData; as well as all CERI staff involved in the production and editing of the material, including but not limited to Allan Fogwill and Megan Murphy. ABOUT THE CANADIAN ENERGY RESEARCH INSTITUTE CANADA S VOICE ON ENERGY Founded in 1975, the Canadian Energy Research Institute (CERI) is an independent, registered charitable organization specializing in the analysis of energy economics and related environmental policy issues in the energy production, transportation, and consumption sectors. Our mission is to provide relevant, independent, and objective economic research of energy and environmental issues to benefit business, government, academia and the public. For more information about CERI, visit CANADIAN ENERGY RESEARCH INSTITUTE 150, Street NW Calgary, Alberta T2L 2A6 info@ceri.ca Phone:

5 An Economic and Environmental Assessment of Table of Contents LIST OF FIGURES... LIST OF TABLES... ACRONYMS AND ABBREVIATIONS... EXECUTIVE SUMMARY... CHAPTER 1 INTRODUCTION... 1 CHAPTER 2 BACKGROUND INFORMATION... 3 Crude Oil Supply and Disposition... 3 Refineries: Background and Trends... 6 The Refinery Process... 6 Eastern Canadian Refineries within a North American and Global Context Crude Oil Transportation Pipeline, Rail and Tanker/Barge Pipeline Crude-by-Rail Oil Tankers and Marine Terminals CHAPTER 3 METHODOLOGY AND ANALYSIS Overview of Methodology and Description of Scenarios Models and Their Approaches Crude Flows Model Costs of Feedstock Model Emissions/LCA Model CHAPTER 4 MODELLING RESULTS Made in Canada Scenario Crude Flows Cost of Feedstock Emissions Expanded Access Scenario Crude Flows Cost of Feedstock Emissions Current Reality Scenario Crude Flows Cost of Feedstock Emissions International Social Concerns Scenario Crude Flows Cost of Feedstock Emissions iii v vii ix xiii

6 iv Canadian Energy Research Institute CHAPTER 5 CONCLUSIONS AND IMPORTANT FUTURE DYNAMICS Conclusions by Scenario Made in Canada Expanded Access Current Reality International Social Concerns Conclusions by Province Inter-scenario Comparisons Expanded Access vs. Current Reality Expanded Access vs. Made in Canada Important Future Dynamics Crude Oil Price Dynamics Availability of Crude Oil Final Remarks BIBLIOGRAPHY APPENDIX A FACTUAL AND USED 2016 IMPORT VOLUMES APPENDIX B DETAILED CRUDE FLOWS AND TRANSPORTATION PATH OF THE BASE CASE APPENDIX C INVENTORY OF CANADIAN AND FOREIGN CRUDE OILS USED IN THE STUDY AND RESULTS OF UPSTREAM AND MIDSTREAM GHG EMISSIONS MODELLING APPENDIX D TANKER TRANSPORTATION COSTS BY ORIGIN AND DESTINATION APPENDIX E DETAILED SUBSTITUTION OF CRUDE OIL BY SCENARIO APPENDIX F DETAILED EMISSIONS BY SCENARIO APPENDIX G COST OF FEEDSTOCK VS. GHG EMISSIONS BY PROVINCE APPENDIX H SELECTED FOREIGN CRUDE OIL BRANDS IMPORT VOLUMES AND PRICE COMPARISON WITH SELECTED CANADIAN CRUDES

7 An Economic and Environmental Assessment of List of Figures v 2.1 Canadian Crude Oil Production by Province, Conventional + Oil Sands US Imports from Canada by Crude Type by Country, Breakdown of an Average Barrel of Canadian Refined Petroleum Products Distribution of Selected Canadian and Imported Crude Oils by API Gravity and Sulfur Content Refining Flow Scheme A Topping/Hydroskimming Refinery A Catalytic Cracking Refinery A Coking Refinery Canada s Refining Sector, Supply and Disposition of Refined Petroleum Products in Canada Canadian Refining Capacity and Number of Refineries, Crude Distillation Capacity Additions by Region, Eastern Canadian Crude Oil Transportation Infrastructure Enbridge s Mainline System Configuration Enbridge s Line 9 System Configuration Canadian Railway Network for Crude Oil Transportation Quarterly Volumes of Crude Oil Exported by Rail to the US The Four Scenarios Schematic of the Modelling Approaches Crude Intake by Type by Province Canadian Oil Supply Routes to Central and Eastern Refineries Base Case Availability and Usage of Crude and Infrastructure by Eastern Canadian Refineries Base Case Canadian Crude Oil Production Forecast, Available Western Canadian Crude for Central and Eastern Refineries and Demand by Central and Eastern Refineries by Crude Type Selected Foreign Crude Oil Brands Import Volumes and Prices Upstream and Midstream GHG Emissions for Selected Canadian and Foreign Crude Oils Used in the Study GHG Emissions Ranges for Crude Oils Used in the Study Upstream and Midstream GHG Emissions for Canadian Proxy and Modelled Crude Flows Total Upstream GHG Emissions for Canadian and Foreign Crude Oils Used for the Study Upstream GHG Emissions for Canadian and Foreign Crude Oils Used for the Study Split by Main Emissions Drivers Midstream GHG Emissions Results for Canadian and Foreign Crude Oils Used for the Study... 74

8 vi Canadian Energy Research Institute 3.15 Transportation GHG Emissions for Canadian and Foreign Crude Oils Used in the Base Case, Refineries Located in ON Transportation GHG Emissions for Canadian and Foreign Crude Oils Used in the Base Case, Refineries Located in QC, NB and NL Total Crude Intake for Central and Eastern Refineries Made in Canada Canadian Crude Supply to Central and Eastern Refineries Made in Canada Availability and Usage of Crude and Infrastructure Made in Canada Made in Canada and Base Case Emissions Change in Emissions Intensity Made in Canada Total Crude Intake for Central and Eastern Refineries Expanded Access Canadian Supply to Central and Eastern Refineries Expanded Access Availability and Usage of Crude and Infrastructure Expanded Access Expanded Access and Base Case Emissions Change in Emissions Intensity Expanded Access Total Crude Intake for Central and Eastern Refineries Current Reality Canadian Supply to Central and Eastern Refineries Current Reality Availability and Usage of Crude and Infrastructure Current Reality Current Reality and Base Case Emissions Change in Emissions Intensity Current Reality Total Crude Intake for Central and Eastern Refineries International Social Concerns Canadian Supply to Central and Eastern Refineries International Social Concerns Availability and Usage of Crude and Infrastructure International Social Concerns International Social Concerns and Base Case Emissions Change in Emissions Intensity International Social Concerns Future Availability of Canadian Crude G.1 Ontario Cost of Feedstock vs. GHG Emissions G.2 Quebec Cost of Feedstock vs. GHG Emissions G.3 Quebec Cost of Feedstock vs. GHG Emissions (zoomed in portion of G.2) G.4 New Brunswick Cost of Feedstock vs. GHG Emissions G.5 Newfoundland & Labrador Cost of Feedstock vs. GHG Emissions H.1 Selected Foreign Crude Oil Brands Import Volumes and Prices H.2 Selected Foreign Crude Oil Brands Import Volumes and Full-cycle Delivered Average Prices

9 An Economic and Environmental Assessment of List of Tables vii E.1 Key Conclusions under the Four Scenarios... xiv 2.1 Eastern Canadian Refinery Capacity and Crude Oil Intake Canada s Refining Sector Configurations, Crude-by-Rail Offloading Terminals in Eastern Canada Oil Tanker Size Oil Marine Terminals in Canada Crude Intake by Eastern Canadian Refineries, by Oil Type Crude Intake by Eastern Canadian Refineries, by Source Oil Type Shares as Input to Western Canadian Refineries Crude Oil Demand by Western Canadian Refineries Available Oil Stock for Central and Eastern Refineries, Crude Oil Brands Used in the Study and Their Prices Pipeline Tolls for Crude Oil Price of Crude Oil Brands at Refinery Gate of Eastern Canadian Refineries Transportation GHG Emissions for Canadian and Foreign Crude Oils Used in the Base Case and the Four Modelled Scenarios Cost of Feedstock Made in Canada Cost of Feedstock Expanded Access Cost of Feedstock Current Reality Cost of Feedstock International Social Concerns Key Conclusions under the Four Scenarios Merchandise Trade Levels between Canada and the Nations Canada Imports Foreign Oil Made in Canada Merchandise Trade Levels between Canada and the Nations Canada Imports Foreign Oil International Social Concerns Comparison of Availability of Light Oil with Demand by the Central and Eastern Refineries, by Scenario C.1 Inventory of Canadian and Foreign Crude Oils Used in the Study for Crude Flows and GHG Emissions Modelling C.2 Results of Upstream and Midstream GHG Emissions Modelling for Canadian and Foreign Crude Oils Used in the Study E.1 Substitution of Oil in the Made in Canada Scenario E.2 Substitution of Oil in the Expanded Access Scenario E.3 Substitution of Oil in the Current Reality Scenario E.4 Substitution of Oil in the International Social Concerns Scenario F.1 Emissions in the Made in Canada Scenario F.2 Emissions in the Expanded Access Scenario F.3 Emissions in the Current Reality Scenario F.4 Emissions in the International Social Concerns Scenario

10 viii Canadian Energy Research Institute

11 An Economic and Environmental Assessment of Acronyms and Abbreviations ix AB AFRA API API Gravity ATL bbl bpd BC CA CARB CAPP CBA CBOB CHOPS CEPA CERI CFA CN CNRL CO2 CO2 eq COLC CP dilbit DWT EIA EPA FCC FERC GDP GHGs GO-HC GREET HS ICE IEA KOH LA LCA LLS Alberta Average Freight Rate Assessment American Petroleum Institute Measure (in degrees) of an oil s gravity or weight Atlantic Canada Barrel Barrels per day British Columbia California California Air Resources Board Canadian Association of Petroleum Producers Cost-Benefit Analysis Conventional Blendstock for Oxygenate Blending (gasoline production) Cold Heavy Oil Production with Sand Canadian Energy Pipeline Association Canadian Energy Research Institute Canadian Fuels Association Canadian National (railway) Canadian Natural Resources Limited Carbon Dioxide Carbon Dioxide Equivalent (including all GHGs) Crude Oil Logistics Committee (Canada) Canadian Pacific (railway) Diluted Bitumen Deadweight Tonnage Energy Information Administration (US) Environment Protection Agency (US) Fluid Catalytic Cracking Federal Energy Regulatory Commission (US) Gross Domestic Product Greenhouse Gases Gas Oil-Hydrocracker The Greenhouse gases, Regulated Emissions, and Energy use in Transportation Model High Sulfur InterContinental Exchange International Energy Agency Potassium hydroxide Louisiana Life-Cycle Assessment Louisiana Light Sweet

12 x Canadian Energy Research Institute LPG LR LS m3 MB Mbpd MMbpd MCR MI MMBtu MJ MR Mt MTCO2eq NB ND n.d. NEB NL NRCan NS NYMEX ON OPEC OPEM OPGEE PADD ppmw PRELIM ptb QC RBOB RFG RPP SCO SHP SK TAN TCPL tonne TX UK ULCC US Liquefied Petroleum Gas Large Range (tanker) Low Sulfur Cubic meter Manitoba Thousand barrels per day Million barrels per day Micro-Carbon Residue Michigan million British thermal units Megajoule Medium Range (tanker) Million tonnes Million tonnes of carbon dioxide equivalent New Brunswick North Dakota Not dated National Energy Board Newfoundland & Labrador Natural Resources Canada Nova Scotia New York Mercantile Exchange Ontario Organization of the Petroleum Exporting Countries Oil Products Emissions Module Oil Production Greenhouse Gas Emissions Estimator United States Petroleum Administration for Defense District parts per million weight Petroleum Refinery Life-cycle Inventory Model Pounds per thousand barrel (salt content) Québec Reformulated Blendstock for Oxygenate Blending (gasoline production) Reformulated Gasoline Refined Petroleum Product Synthetic Crude Oil Selective Hydrogenation Process Saskatchewan Total Acid Number TransCanada Pipelines Limited Metric ton Texas United Kingdom Ultra Large Crude Carrier United States

13 An Economic and Environmental Assessment of xi VFF VLCC WCB WCS wt% WTI Venting, Flaring or Fugitive (emissions) Very Large Crude Carrier Western Canadian Blend Western Canadian Select Percentage on weight basis Western Texas Intermediate

14 xii Canadian Energy Research Institute

15 An Economic and Environmental Assessment of Executive Summary xiii Canada is the fifth largest oil producer in the world, accounting for 4.8 percent of world production in 2016, ranking behind the US (13.4 percent), Saudi Arabia (13.4 percent), Russia (12.2 percent) and Iran (5.0 percent) (BP 2017). Canada s proved reserves, totaling billion barrels or 10 percent of the world s share of proved reserves, are behind Venezuela (300.9 billion barrels) and Saudi Arabia (266.6 billion barrels) (BP 2017). Yet, despite this, Canada still imports oil. Eastern Canadian crude oil imports rose slightly in 2016, reaching 607 thousand barrels per day (Mbpd), with the majority of imports from the US (259 Mbpd), followed by Saudi Arabia (87 Mbpd) and Algeria (85 Mbpd) (Government of Canada 2017; NEB 2017f; Statistics Canada 2017a). The nature of the imports, however, is different between western and eastern Canada. Western Canadian crude oil imports are entirely from the US and are attributed to diluent or condensate, product used to dilute oil sands bitumen to facilitate transportation by pipeline. Eastern Canadian provinces, on the other hand, use domestic oil (either from western Canada or offshore Newfoundland & Labrador (NL)) and imported oil from various parts of the world for feedstock into eight refineries (four in Ontario (ON), 2 in Quebec (QC) and single refineries in New Brunswick (NB) and NL & Labrador) to process oil into gasoline, diesel, heating oil, propane, asphalt, and petrochemical feedstock. The scope of this study is to analyze the potential complete or partial substitution of eastern Canadian crude oil imports via domestically-sourced oil. The research provides a cost and emissions comparison based on four potential scenarios of substituting domestic vs. foreign crude oil in the central and eastern Canadian refinery market. The four scenarios include: Made in Canada in this scenario we consider complete substitution of Canadian crude oil for imported oil regardless of the cost. This scenario allows for expanded pipeline infrastructure in Eastern Canada. Expanded Access in this scenario we allow for the economic substitution of less expensive Canadian crude for imported oil. Again, a new oil pipeline is assumed in this scenario. Current Reality in this scenario we use the existing infrastructure but maximize the selection of Canadian crude from an economic perspective. The Base Case which now exists, has a portion of Canadian crude being used in central and eastern Canadian refineries but not the full amount that would be selected if only price was used to make the selection.

16 Emissions Costs Flows xiv Canadian Energy Research Institute International Social Concerns is a scenario that substitutes Canadian crude for imported oil from countries where international organizations have identified some social concerns. CERI uses the Economist s Democracy Index as a proxy for the range of social concerns (e.g., living standards, food security, human rights, environmental protection, health outcomes, etc.). Table E.1: Key Conclusions under the Four Scenarios Category Additional western Canadian supply (Mbpd) Additional eastern Canadian supply (Mbpd) Total additional Canadian crude (Mbpd) Expanded Infrastructure Made in Canada Expanded Access Existing Infrastructure Current Reality International Social Concerns Substituted foreign oil (%) 100% 57% 47% 50% Annual Cost of feedstock ($ million) Emissions (million tones CO2eq per year) Change of emissions (%) -6.2% -5.7% -5.7% -7.9% This study provides a comprehensive look at the substitution impact on refinery costs and emissions for refineries in ON, QC, NB and NL. Refineries in these provinces only process lighter crudes and are not able to process some of the heavier crude oils from Western Canada. As such, the substitution is on a like-for-like basis. The crude oil quality being imported is substituted with the same quality of crude from other parts of Canada. In all the scenarios, the substitution of Canadian crude oil for imported oil reduces overall global CO2 emissions compared to a business as usual (Base Case) option. Emissions are reduced by 2 MTCO2eq per year to 2.8 MTCO2eq per year. In some cases, Canadian emissions increase, but overall emissions which are linked to climate change decrease. Cost savings range from $23 million in the Made In Canada scenario to $317 million in the Expanded Access scenario. Both scenarios call for a new oil pipeline. In the case of Expanded Access, only those Canadian crude oil supplies cheaper than their imported counterparts are consumed in central and central and eastern refineries.

17 An Economic and Environmental Assessment of xv One scenario results in higher costs for QC and NB refineries. This is the substitution of oil from countries where international organizations have noted some social concerns. In this case, the cost of this policy would be approximately $79 million in additional crude oil costs.

18 xvi Canadian Energy Research Institute

19 An Economic and Environmental Assessment of 1 Chapter 1: Introduction Canada is the fifth largest oil producer in the world, accounting for 4.8 percent of world production in 2016, ranking behind the US (13.4 percent), Saudi Arabia (13.4 percent), Russia (12.2 percent) and Iran (5.0 percent) (BP 2017). Canada s proved reserves, totaling billion barrels or 10 percent of the world s share of proved reserves, are behind Venezuela (300.9 billion barrels) and Saudi Arabia (266.6 billion barrels) (BP 2017). Yet, despite this, Canada still imports oil. Canadian crude oil imports were 759 thousand barrels per day (Mbpd) in 2016 with the majority coming into the Eastern part of the country. ON, QC, NB, and NL imported 607 Mbpd combined, with most of imports from the US (259 Mbpd), followed by Saudi Arabia (87 Mbpd) and Algeria (85 Mbpd). Other imports in 2016 over 10 Mbpd include Nigeria (74 Mbpd), Norway (42 Mbpd), Kazakhstan (19 Mbpd), and the Ivory Coast (13 Mbpd) (Government of Canada 2017; National Energy Board [NEB] 2017f; Statistics Canada 2017a). The nature of the imports, however, is different between western and eastern Canada. Western Canadian crude oil imports are all from the US and are attributed to imports of diluent or condensate, which is used to dilute oil sands bitumen to facilitate transportation by pipeline. Eastern Canadian provinces, on the other hand, use domestic oil (either from western Canada or offshore NL) and imported oil from various parts of the world for feedstock in local refineries. The theoretical objective of an oil refiner is simple to minimize operating expenses and maximize margins. In practice, however, the economics of a refinery is dictated by a complex set of intertwined factors, at the heart of which are three variables: crude oil type used (crude slate), refinery size and configuration and final product range (product slate) (Natural Resources Canada [NRCan] 2016). Other important variables include a refinery s operational efficiency, or utilization rate, as well as regulatory requirements, often reflecting environmental considerations (NRCan 2016). These factors, as well as their proximity to crude oil producing areas and available transportation infrastructure, help explain why each province has very different crude oil supply dynamics. The role of transportation options (pipelines, rail, and tankers and/or barges) in determining a refiners crude sources cannot be overstated. In the west, it is not surprising that western Canadian refineries, on the other hand, use exclusively crude oil produced in the region. In the east, the dynamics are more complicated. ON accesses western Canadian production via the Enbridge Mainline pipeline, as does Suncor s refinery in Montréal via the Line 9b reversal. ON is, however, unique in that it uses feedstock from western Canada, but also imports crude oil, primarily from the US. In QC, while Montréal s Suncor refinery can be supplied with western Canadian crude oil by pipeline and rail, as well as receive eastern Canadian oil through the Portland-Montréal pipeline, QC s Valero refinery relies more on importing foreign crudes by

20 2 Canadian Energy Research Institute tankers. Located near Quebec City, Valero can access western Canadian crude oil from Montréal via smaller tankers and is also able to receive oil from the west by rail. The facility can also access eastern Canadian oil by tanker through its offloading facilities, which are primarily used for importing foreign-sourced crude oil. Like QC s Valero refinery, Atlantic Canada also depends on importing crudes due to a lack of pipeline infrastructure. They access a diverse slate of foreign crudes, including the US, via tankers. In 2016, QC received crude from nations such as Algeria, Azerbaijan, Kazakhstan, Norway and Nigeria while NB received crude from nine different nations including Saudi Arabia, Congo, Ivory Coast, Colombia, Nigeria and Norway (Government of Canada 2017; NEB 2017f; Statistics Canada 2017a). The crude slate for NL s single refinery is equally diverse. And while the facility uses some of its offshore oil as a feedstock, it does not use any crude from western Canada. NB s Irving Oil refinery, however, can access western Canadian crude oil via rail, as well as Offshore NL crude oil by tanker. The scope of this study is to analyze the potential complete or partial substitution of eastern Canadian crude oil imports via domestically-sourced oil. The research provides a cost and emissions comparison based on four potential scenarios of substituting domestic vs. foreign crude oil in the central and eastern Canadian refinery market. It is important to note that transportation capacity is not the constraint and consideration in foreign oil substitution. Other constraints in refinery decision-making include economics of supply and crude availability in western and eastern Canada. The former means that refineries modus operandi is to optimize cost of feedstock while gaining desired yields demanded by the refined petroleum markets. Thus, if Canadian oil is more expensive at the gate of a refinery compared to foreign oil, in a market-driven non-policy constrained decision-making world, it would prefer the cheaper feedstock (if it can get the same desired yields). Prices of particular brands of foreign and domestic oil as well as transportation costs play a role in the selection of the crude slate. This CERI study will show market- and socially-pushed scenario outcomes, via existing and expanded transportation infrastructure, which could serve as a sign for Canadians on how beneficial or costly such substitution could be. This research is a first of its kind comparison that has not been done by any other organization. Before discussing the methodology and assumptions undertaken in this study, it is prudent to provide a brief review of several important concepts central to this study. Prior to discussing any substitution, it is important to grasp the sheer complexity of the eastern refinery market, in terms of the refinery business as well as the complex transportation system connecting western and eastern Canadian producers, and their international counterparts, to the eastern Canadian refineries.

21 An Economic and Environmental Assessment of 3 Chapter 2: Background Information This chapter is divided into three parts: an overview of Canada s crude oil supply and disposition, an overview of the eastern Canadian refining sector within a global context, and a brief overview of the various transportation modes used in the study (pipeline, rail, and tanker and/or barge). In addition to providing a discussion of the refinery sector in eastern Canada (including refiner capacities, use rates, crude intakes, and mode of supply), this section will also include a brief discussion of trends and the Canadian refinery sector within a global context. Likewise, the overview of transportation modes will focus on pipelines, rail, and tankers important to transport crude oil to eastern Canadian refineries. Crude Oil Supply and Disposition Canada s total oil production, including bitumen and synthetic, has been continuously growing over the last seven years except for the period where production is flat, due to the Fort McMurray wild fires. In 2016, production was 3,872 Mbpd, up from 3,869 Mbpd in 2015 and up from 3,698 Mbpd in 2014 (2017a, 2017b). Canada s total conventional (including light, C5+/condensate and heavy) oil production in 2016 was 1,448 Mbpd, led by 666 Mbpd in AB, followed by 461 Mbpd in SK and 210 Mbpd in Atlantic Canada (NEB 2017a, 2017b). AB oil sands production, including bitumen and synthetic production, totals 2,418 Mbpd (NEB 2017a, 2017b). Figure 2.1 illustrates total oil production in Canada from 2009 to Canadian conventional production includes light, C5+/condensate and heavy crude oil production and is shown by producing province. AB s oil sands production includes bitumen and synthetic.

22 4 Canadian Energy Research Institute Figure 2.1: Canadian Crude Oil Production by Province, Conventional + Oil Sands (Mbpd) 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1, Atlantic ON MB SK AB BC NWT AB - Bitumen & Synthetic Data Sources: (NEB 2017a, 2017b), Canadian Energy Research Institute (CERI) calculations. Figure by CERI. Canada is a significant exporter of crude oil, with most oil exported to the US. Figure 2.2 illustrates the US imports from Canada by type. Total US imports from Canada in 2016 were 3,264 Mbpd, up from 3,169 Mbpd in 2015 and up from 1,939 Mbpd in 2009 (US Energy Information Administration [US EIA] 2016). In addition, it is important to mention that Canadian production, in western Canada, does supply other provinces.

23 An Economic and Environmental Assessment of 5 3,500 Figure 2.2: US Imports from Canada by Crude Type (Mbpd) 3,000 2,500 2,000 1,500 55% 57% 59% 59% 59% 60% 62% 60% 1, % 27% 27% 26% 25% 25% 21% 22% Light Sweet Light Sour Medium Heavy Sweet Heavy Sour Data Source: (US EIA 2016). Figure by CERI. It is interesting to note that the largest share of exports is heavy sour bitumen, accounting for 60 percent of type of crude exported from Canada to the US. Heavy sour exports to the US increased from 1,063 Mbpd in 2009 to 1,970 Mbpd in 2016 (US EIA 2016) and are destined to refineries capable of processing heavier crudes in the US Midwest (US Petroleum Administration for Defense District [PADD] II) and the US Gulf Coast (PADD III). Eastern Canadian crude oil imports rose slightly in 2016, reaching 607 Mbpd, with most of the imports from the US (259 Mbpd). Rounding out the top 5 in 2016 is Saudi Arabia (87 Mbpd), Algeria (85 Mbpd), Nigeria (74 Mbpd) and Norway (42 Mbpd) (Government of Canada 2017; NEB 2017f; Statistics Canada 2017a). This is illustrated in Figure 2.3.

24 6 Canadian Energy Research Institute Figure 2.3: by Country, 2016 (Mbpd) United States Algeria Saudi Arabia Nigeria Norway 41.8 Kazakhstan Ivory Coast United Kingdom Azerbaijan Colombia Congo Equatorial Guinea Denmark Data Sources: (Government of Canada 2017; NEB 2017f; NRCan 2017b; Statistics Canada 2017a), CERI calculations. Figure by CERI. Refineries: Background and Trends Refineries and their economics are dictated by a complex set of variables. In the short term, refineries must juggle their choices of inputs (crude slate) and refined outputs (product slate), both of which are typically price-takers. Refineries must optimize production against a backdrop of changing demand patterns and increased global competition among refiners to be profitable. In the longer term, refiners should decide whether to invest in changing their configuration or shutting down. While changing environmental regulations are typically not part of the day-to-day crude slate decision making, these changes affect the refinery capital program and long-term plan. The Refinery Process The refining sector is a critical component of the value chain of oil as a commodity, an intermediary between crude oil and valuable and usable refined products. Oil is the primary feedstock into the petroleum refining industry, processing oil into products such as gasoline, diesel, heating oil, propane, asphalt, and petrochemical feedstock. The latter transforms crude oil into a variety of consumer goods and products, ranging from plastics to pharmaceutical products.

25 An Economic and Environmental Assessment of 7 Figure 2.4 illustrates the breakdown of an average barrel of Canadian refined petroleum products, produced by Canadian refineries. The products are illustrated from heaviest to lightest, by percentage. In Canada, between 2011 and 2016, the largest component is motor gasoline (36 percent), followed by diesel and middle distillates (33 percent), petrochemical feedstocks and LPGs (7 percent), heavy fuel oil (6 percent), other heavy products (6 percent), aviation fuel (5 percent), other light products (4 percent) and asphalt (4 percent) (NEB 2017g). Figure 2.4: Breakdown of an Average Barrel of Canadian Refined Petroleum Products (%) Source: (NEB 2017g) The primary end-products are typically classified into three separate categories: light distillates, middle distillates, and heavy distillates. Light distillates include liquified petroleum gas (LPG), gasoline and heavy naphtha while middle distillates include kerosene, automotive and railroad diesel fuels, residential heating fuel and other light fuel oils. Heavy distillates include heavy fuel oils. However, before discussing the complex refining process, it is prudent to review the distinct types of crude oil. Globally, there are over 150 different types of crude and even more brands of crude (CFA 2013). Not only do they impact the price of feedstock to the refiner, they also produce different yields for the refiners (CFA 2013). On the simplest level, crude oils are differentiated by sulfur content (sweet or sour) and by American Petroleum Institute (API) gravity (light or heavy). Aside from density (kg/m3), API gravity (degrees), sulfur (percentage on weight basis [wt%]), grades of crude are also differentiated by several other physical characteristics including micro-

26 8 Canadian Energy Research Institute carbon residue (MCR) (wt%), sediment (parts per million weight [ppmw]), total acid number (TAN) (mg potassium hydroxide [KOH]/g), salt (pounds per thousand barrel [ptb]), nickel (mg/kg), vanadium (mg/kg) and olefins (wt%). Figure 2.5 illustrates the distribution of selected Canadian and imported crude oils used in this study by API gravity and sulfur content. Figure 2.5: Distribution of Selected Canadian and Imported Crude Oils by API Gravity and Sulfur Content Data Sources: (PRELIM v1.1 (J. A. Bergerson et al. 2016); CrudeMonitor.ca (Crude Quality Inc. 2017)), crude oil assays from EcoPetrol, ExxonMobil, Maersk Oil, TOTAL, Tullow Oil websites. Figure by CERI. Notes: Canadian crude oils, foreign crude oils. API gravity values represented in Figure 2.5 are the crude oil API gravity values at the refinery inlet gate. The higher the sulfur content, the sourer the crude oil while the lower the sulfur content, the sweeter the crude oil. Likewise, crude oil with a low API gravity is considered a heavy crude oil and typically has a higher sulfur content, resulting in a larger yield of lower-valued products (NRCan 2016). Therefore, the lower the API of a crude oil, the lower the value it has to a refiner, as crude will either require more processing or yield a higher percentage of lower-valued byproducts such as heavy fuel oil, which usually sells for less than gasoline (NRCan 2016). Refiners

27 An Economic and Environmental Assessment of 9 are willing to pay more for light, low sulfur crude oil. While lighter grades require less upgrading at the refinery, they are in decreasing supply (CFA 2013). Heavy crude oils, on the other hand, are cheaper, but more expensive to refine since they require significant investments (CFA 2013). It is important to note that at the core of this research study is the substitution of foreign crude oils by western Canadian crude, by type. Imported light oil is, hence, substituted by light oil of a similar crude type and physical characteristics by either western Canadian crude oil or from Offshore NL. This is discussed in greater detail later in this chapter. Figure 2.6 illustrates the complex refining process, as crude oil moves to its various derivative products, such as gasoline, kerosene and jet fuel, diesel, as well as heavier products such as lubricants (lubricating oil), fuel oil and residual, used for roads and building materials. The derivative products are listed on the right side of the figure. Figure 2.6: Refining Flow Scheme Source: (Honeywell UOP 2017) Note: FCC fluid catalytic cracking; LPG liquified petroleum gas; SHP selective hydrogenation process.

28 10 Canadian Energy Research Institute While there are common features and processes, no two refineries are identical. Refineries process different crude oils with different chemical characteristics and process them into a range of refined products. In doing so, refinery configurations, or the type of processing units at the refinery, are different. Different types of processing units can process certain types of crude oil. As such, some refineries are simply not able to process heavier crude oils. Canadian oil production has shifted from declining conventional reserves to heavier, higher sulfur crudes, such as bitumen. Eastern Canadian refineries that wish to take advantage of lower priced bitumen will need to change their refinery configurations and to invest into additional coking, hydrocracking and hydrotreating units to process heavier barrels into the same yield of refined products. Hydrotreating will be needed to meet the Canadian fuels regulations for lower sulphur products, especially if the crude slate includes higher sulphur crudes. While there are multiple processing units used, most refineries fall into three broad categories: a topping plant with a distillation unit, a cracking refinery, and a coking refinery. They represent increasing levels of complexity and increase in cost. A topping refinery is the simplest type of refinery. They have a distillation tower and convert light sweet crude oil into either light or middle distillates. Some topping refineries are also equipped with a catalytic reformer, adding a layer of complexity, as well as the capability to produce gasoline (Simmons and Phillip 2008). These refineries are referred to as hydroskimming refineries. The latter provides octane to gasoline for example. In addition, some topping refineries can be asphalt plants. The atmospheric distillation units produce hydrocarbon boiling point cuts from resid to LPG (Sep- Pro Systems 2009). The crude oil is heated with fire heaters to temperatures ranging between 650 and 700 degrees Fahrenheit (Inkpen and Moffett 2011). With most of it vapourizing, it enters the atmospheric distillation tower or column, getting cooler as it goes up in the tower (Inkpen and Moffett 2011). This process is illustrated by Figure 2.7.

29 An Economic and Environmental Assessment of 11 Figure 2.7: A Topping/Hydroskimming Refinery Source: (Simmons and Phillip 2008) Note: CARB gasoline gasoline that meets California Air Resources Board (CARB) regulations; CBOB conventional blendstock for oxygenate blending; HS high sulfur; LS low sulfur; RBOB reformulated blendstock for oxygenate blending. As the various hydrocarbons cool below their respective boiling points, they revert to a liquid state. Heavier hydrocarbons such as heavy gas oil or lubricating oil are collected on the lower trays of the distillation tower, while lighter hydrocarbons are collected near the top of the tower (Inkpen and Moffett 2011). Vacuum distillate units are often used to separate the gas oil from the lower quality heavy fuel or resid (Simmons and Phillip 2008). Lighter hydrocarbon products range from the heavier middle distillate, to kerosene and naphtha and low octane gasoline. Catalytic reformers increase the octane number of gasoline. Light sweet crude or condensate is used as the crude slate at a topping or hydroskimming refinery. Asphalt plants are topping refineries that run heavy crude oil because they are primarily interested in producing asphalt. The second type of refinery is a cracking refinery. This level of process is often referred to as medium conversion. This type of refinery takes the gas oil portion from the crude distillation unit and breaks it down further into gasoline and distillate components. Cracking breaks or cracks the heavier products into more valuable products such as gasoline and diesel (Inkpen and Moffett 2011). The fluid catalytic cracker (FCC) process is shown in red and illustrated in Figure 2.8.

30 12 Canadian Energy Research Institute Figure 2.8: A Catalytic Cracking Refinery Source: (Simmons and Phillip 2008) Note: CARB gasoline gasoline that meets California Air Resources Board (CARB) regulations; HS high sulfur; LS low sulfur; RFG reformulated gasoline. The FCC, for example, converts heavier hydrocarbon molecules into lighter hydrocarbon molecules. This is done using catalysts, high temperature and/or pressure, operating at approximately 986 degrees Fahrenheit (Vellenga 2017). It is important to note that cracking refineries tend to use more sour crude oils and upgrade them to higher value products. They have increased yields of higher value products than topping refineries. These types of refineries, using FCC, are common in the US, due to the higher demand in gasoline. FCC units are, on the other hand, less common in Europe and parts of Asia, due to the higher demand in diesel and kerosene, both of which can be satisfied by hydrocracking. The latter is similar to FCC but yields more diesel and represents a more complex type of refinery (Inkpen and Moffett 2011). The most complex type of refinery, or high conversion refinery, is the coking refinery. This type of refinery processes residual fuel, the heaviest material from the crude unit and thermally cracks it into lighter product. This is done in a coker or a hydrocracker. The addition of a FCC unit significantly increases the yield of higher-value products like gasoline and diesel oil from a barrel of crude. High conversion refinery processes are illustrated in Figure 2.9.

31 An Economic and Environmental Assessment of 13 Figure 2.9: A Coking Refinery Source: (Simmons and Phillip 2008) Note: CARB gasoline gasoline that meets California Air Resources Board (CARB) regulations; RFG reformulated gasoline. It also allows the refinery to process cheaper, heavier crude while producing an equivalent or greater volume of high-value products (Simmons and Phillip 2008). Other processes such as hydrotreating, which removes sulfur from finished products, give the refiners the ability to process crude oil with a higher sulfur content. Canada has primarily cracking refineries, using a mix of light and heavy crude oils to meet Canadian demand for gasoline, diesel, and other products. Historically, the abundance of domestically produced light sweet crude oils and a higher demand for distillate products, such as heating oil, reduced the need for upgrading capacity in Canada (NRCan 2016). However, in more recent years, the supply of light sweet crude has declined, and newer sources of crude oil tend to be heavier. Eastern Canadian Refineries within a North American and Global Context Canada s refinery market is considered as three separate entities: Western Canada; ON; and QC and Atlantic Canada. In some cases, QC and Atlantic Canada are considered separate entities as well. Western Canada is land-locked and refineries have primarily sourced their crude from local and regional production. ON can access western Canadian production via pipelines, QC and Atlantic Canada depend on importing crudes because of a lack of pipeline infrastructure, and the cost of transporting crude from western to eastern Canada is more expensive than importing foreign crudes. There are 15 refineries in Canada, seven of which are in western Canada and eight located in eastern Canada. In addition, Canada also has two refineries producing asphalt (Husky s

32 14 Canadian Energy Research Institute Lloydminster Refinery and Gibson Energy s Moose Jaw Refinery) and a lubricants plant (Hollyfrontier s Clarkson Refinery). Figure 2.10 illustrates the location of Canada s refineries as well as their capacity and product demand (by region). The two main clusters are the Edmonton area (AB) and the Sarnia area (ON). It is important to note that Canada s first new refinery since 1984 is the North West Redwater Partnership, sometimes referred to as the Sturgeon Refinery (AB). North West Refining and Canadian Natural Resources Limited (CNRL) have a 50/50 partnership in the Sturgeon Refinery. Figure 2.10: Canada s Refining Sector, 2016 Source: (CFA 2017b) Note: kb/d thousand barrels per day. Canada is a net exporter of refined products refinery capacity exceeds domestic demand, notably in QC and Atlantic Canada. Net exports reached nearly 17 billion litres in 2013, decreasing to 6 billion litres in 2016 (Statistics Canada 2017e). Figure 2.11 illustrates the net exports, as well as domestic sales of refined petroleum products by product (motor gasoline, diesel, and other refined petroleum products [RPP]) and total refinery production. Total refinery production reached 117 billion litres in 2010 and decreased to 105 billion litres in 2014, but increased slightly to 109 billion litres in 2016 (Statistics Canada 2017d).

33 An Economic and Environmental Assessment of 15 Figure 2.11: Supply and Disposition of Refined Petroleum Products in Canada (billion litres) Refinery Production Motor Gasoline Sales Diesel Fuel Oil Sales Other RPPs Sales Net Exports Source: (Statistics Canada 2017e) and CERI calculations. Figure by CERI. This section delves into the eight eastern Canadian refineries and various important metrics, including their location, capacity, use rates, crude intakes (or crude slate) and mode of supply (transportation). Consumption of feedstock (crude slate) and product slate of the eight refineries is reviewed, establishing the current state of refineries in eastern Canada. The crude slate is reviewed more closely, breaking it down by domestic-sourced crude and foreign-sourced crude. It is first important to note the sources of data used. The data was obtained from several sources including Statistics Canada, Natural Resources Canada (NRCan), Canadian Fuel Association (CFA), Canadian Association of Petroleum Producers (CAPP), various refinery websites, midstream companies websites (Enbridge and others), Clipper Data (transportation of oil by tankers), as well as calculations and estimates by CERI where needed. Statistics Canada CANSIM Table Supply and Disposition of Crude Oil and Equivalent and Table Refinery Supply of Crude Oil and Equivalent were used (Statistics Canada 2017c, 2017d). If the data was suppressed, missing, or conflicted with one another, CERI reconciled using multiple sources to infer quality and certainty in data. For instance, for QC, 10 months of 2016 were used to determine the shares of the different crude oil intake; likewise, for ON, five months of 2016 were used due to the availability of data (other months are either not available or suppressed by Statistics Canada s data). Information on type and volumes of crude that Irving Oil Refinery and North Atlantic Refining are receiving is not available from Statistics Canada. As such, numbers are estimated by CERI using data from other sources and various publications. Refining capacity in eastern Canada is represented by eight refineries with a total capacity of 1,228 Mbpd (CFA 2017b). Four refineries are in ON, two in QC, and one each in NB and NL. All eight refineries, their capacities, use rates and total 2016 crude intakes are illustrated in Table 2.1.

34 16 Canadian Energy Research Institute Table 2.1: Eastern Canadian Refinery Capacity and Crude Oil Intake Province Refinery Capacity (Mbpd) Utilization Rate (%) Total Intake (Mbpd) ON Imperial, Sarnia % Shell Canada, Corunna 75 88% 65.9 Suncor Energy, Sarnia 85 92% 77.8 Imperial, Nanticoke % 96.3 ON Subtotal 393 QC Valero, Lévis % Suncor Energy, Montréal % QC Subtotal 402 NB Irving Oil, Saint John % NL North Atlantic Refining, % 93.1 Come By Chance Atlantic Canada Subtotal 433 Total 1,228 1,073.0 Data Sources: (CFA 2017b; Statistics Canada 2017c, 2017d), various refinery websites & CERI (calculations and estimates). Table by CERI. All of ON s refineries are in the south of the province, with three refineries found in the Sarnia area (Suncor Energy and Imperial Oil operate refineries in Sarnia while Shell Canada operates the Corunna Refinery in nearby St. Clair). Imperial Oil operates a refinery in Nanticoke. As can be seen from Table 2.1, Imperial Oil s Nanticoke Refinery and Sarnia Refinery have capacities of 112,000 and 121,000 bpd, respectively. Suncor Energy s Sarnia refinery has a capacity of 85,000 bpd and Shell Canada s St. Clair refinery has a capacity of 75,000 bpd. QC has two refineries, Suncor Energy s Montréal Refinery and Valero s Jean-Gaulin Refinery, in Lévis, near Quebec City. The refineries have capacities of 137,000 bpd and 265,000 bpd, respectively. Atlantic Canada also has two refineries, Irving Oil Refinery in Saint John, NB, and the North Atlantic Refinery in Come By Chance, NL. The former has a capacity of 318,000 bpd and is Canada s largest refinery. The North Atlantic Refinery has a capacity of 115,000 bpd. Table 2.2 illustrates the refinery configurations for Canada s central and eastern refineries. These include coking (25.5 Mbpd in ON), visbreaking (5 Mbpd in ON and 40 Mbpd in Atlantic provinces), hydrocracker (around 60 Mbpd in ON, 22 Mbpd in QC, and 72 Mbpd in Atlantic provinces), catalytic cracker (116 Mbpd in ON, 100 Mbpd in QC, and 95 Mbpd in Atlantic provinces), catalytic reformer (107 Mbpd in ON, 84 Mbpd in QC, 69 Mbpd in Atlantic provinces), and hydrotreating (291 Mbpd in ON, 321 Mbpd in QC, and 148 Mbpd in Atlantic provinces) (Oil & Gas Journal 2016b).

35 An Economic and Environmental Assessment of 17 Table 2.2: Canada s Refining Sector Configurations, 2016 (Mbpd) Refinery Location Vacuum Coking Visbreaking Catalytic Cracker Catalytic Reformer Hydrocracker Hydrotreating Imperial Oil Sarnia Shell Canada Corunna Suncor Energy Sarnia Imperial Oil Nanticoke Valero Lévis Suncor Energy Montreal Irving Oil Saint John North Atlantic Come by Refining Chance Data Source: (Oil & Gas Journal 2016b). Table by CERI. The economics behind whether to invest in these additional coking units is determined by the current and expected spread between light and heavy oil prices. The gain in value for investing in additional units must yield a high enough return to offset the increased costs of building more units. Refining capacity to process heavy crude is limited because recently the light-heavy crude differential has not been consistently wide enough to cover the cost of installing additional upgrading or refining capacity. It is important to put the current state of Canadian refineries into a historical and global context. The former is illustrated in Figure 2.12.

36 18 Canadian Energy Research Institute Figure 2.12: Canadian Refining Capacity and Number of Refineries, Source: (Oil Sands Magazine 2016). Note: kbpcd thousand barrels per calendar day From a historical perspective, the number of refineries has been decreasing over the past several decades. With the closures of Shell s Montréal refinery in 2010 and the Imperial Oil refinery in Dartmouth, NS, there are now 15 refineries. Recall, Figure 2.12 also includes the Hollyfrontier Refinery, the Moose Jaw Refinery, and the Lloydminster Refinery, which are either asphalt or lubricant plants. This is down from 36 in 1980 and down from a high of 45 refineries in The same is true south of the border. The number of operating refineries decreased from 300 in 1982 to 137 in 2017 (US EIA 2017). Refinery capacity in Canada increased dramatically from 264 Mbpd in 1947 to a record-high of 2,232 Mbpd in The capacity subsequently decreased in the 1990s, hovering around the 1,800 Mbpd level. It has increased to 1,910 Mbpd in 2016 (Oil Sands Magazine 2016). Again, these totals include asphalt and lubricant plants. A similar pattern occurs in the US. Operating refinery capacity decreased from 18 million barrels per day (MMbbl per calendar day) to a low of 15 MMbbl per calendar day in 1993, where capacity increased back to 18 MMbbl per calendar day in 2017 (US EIA 2017).

37 An Economic and Environmental Assessment of 19 However, the number of refineries in Canada is likely to decrease further, due to several factors. First, like their global counterparts, including in the US, refineries are getting larger and more efficient, and increasingly satisfying local demand for gasoline and diesel. With cars and trucks increasing fuel efficiency, the future demand for refined petroleum products is flat in North America, if not decreasing. While increased vehicle efficiency and renewable fuels will likely impact the near-term, another question is the penetration of electric vehicles into the US transportation fleet. As a net exporter of refined petroleum products, this will leave several Canadian refineries vulnerable. It should, however, be emphasized that the rate of decline varies widely depending on the assumptions made by different studies; factors include the growth of electric vehicle sales, growth in biofuel production and consumption, and the effectiveness of current climate and transportation policies. Second, as previously mentioned, refineries could also be susceptible to changes in regulations, most notably, climate change initiatives. A recent study (March 2017), Cumulative Impacts on Climate Change and Other Policy Scenarios Facing the Canadian Petroleum Sector, suggested that eastern Canadian refineries are more vulnerable to potential closure by 2030 (Tamm and Milburn 2017). A potential 20 to 40 percent drop in demand could lead to shutting down three refineries in the case of the former and five refineries in the case of the latter. Some pundits expect a decrease of percent of capacity in the next few years (Tamm and Milburn 2017). It should be noted that in the context of current policies, decrease in demand of more than 20 percent (and associated refinery closures), appear to be an unlikely scenario at this point. Within a global context, North America s share of global refining capacity decreased slightly from 25 percent to 24 percent between 2005 and 2015 (Oil Sands Magazine 2016). Over the same time frame, Europe and Eurasia decreased from 28 percent to 25 percent. While the Middle East s share of the global refining capacity stayed at 9 percent, the largest growth occurred in the Asia Pacific, which increased from 28 percent to 34 percent between 2005 and The latter is led by rapid growth in China, increasing from a capacity of approximately 9,000 Mbpd in 2008 to over 14,000 Mbpd in 2015 (Oil Sands Magazine 2016). North American refineries, particularly Canadian, face competition from abroad as global refining capacity is capable of processing heavy crudes at lower costs than in North America. Aside from the economic challenges, locational and political issues affect access to this processing capacity. Refineries tend to be found close to their markets for refined products. 1 Some of the reasons behind this are that it is less expensive to transport crude oil than refined products, the specifications for gasoline vary across the world, and countries prefer energy security for refined products. For example, historically the US refinery configuration has been to maximize gasoline production, but in Asia, refinery configurations support industrial development especially for diesel and petrochemical production (Hackett et al. 2013). 1 Even though the scale of refining capacity and proximity to tidewater could mean that the refineries do not necessarily need to be in close proximity to their markets.

38 20 Canadian Energy Research Institute Refiners also face challenges in that the North American market has peaked in its product demand, meaning that an expansion of refining capacity to meet oil production would mean selling refined products to the export markets, specifically in developing nations of the Asia Pacific (IHS CERA 2013). Figure 2.13 depicts crude distillation capacity additions by region. Forty-six percent of the capacity additions are projected for the Asia-Pacific region between 2016 and 2040 (OPEC 2016). This is followed by the Middle East, Africa and Latin America at 17 percent, 16 percent and 11 percent, respectively (OPEC 2016). Figure 2.13: Crude Distillation Capacity Additions by Region, (MMbpd) Source: (OPEC 2016) On a global scale, China and the Asia-Pacific region are leading the growth of refined product demand. This is reflected in the region s refining capacity. The region is leading the worldwide trend of larger, more efficient refineries. The Asia-Pacific, including China, is expected to add around 3.3 MMbpd of new distillation capacity in the medium-term to 2021, while the Middle East should expand by 1.7 MMbpd (OPEC 2016). This is followed by growth in Latin America and Africa. The figure reflects that markets in Canada, US and Europe are mature and growing at a slower rate. It is not surprising that the 10 largest oil refineries in the world are primarily situated in the Asia- Pacific region: Reliance Jamnagar Refinery, India (1.24 MMbpd), Paraguana Refining Centre, Venezuela (955,000 bpd), Ulsan Refinery South Korea (840,000 bpd), Yeosu Refinery, South Korea,(775,000 bpd), Onsand Refinery, Ulsan, South Korea (669,000 bpd), Port Arthur Refinery, Texas, US (600,000 bpd), ExxonMobil s Singapore Refinery (592,000 bpd), ExxonMobil s Baytown Refinery, Texas, US (584,000 bpd), Ras Tanura Refinery, Saudi Arabia (550,000 bpd) and Garyville

39 An Economic and Environmental Assessment of 21 Refinery, Louisiana, US (522,000 bpd) (Dudu 2013). The largest Canadian refinery is NB s Irving Refinery at 318,000 bpd. Crude Oil Transportation Pipeline, Rail and Tanker/Barge Crude oil can be transported in four ways: pipeline, rail, tanker/barge, and truck. While discussed in greater detail in Chapter 3, transporting crude from western Canadian producers to eastern Canadian refineries use various modes of transportation, depending on the location of the refinery. The same is true for transporting foreign crude oil to eastern Canadian refineries. This section highlights the various elements of the three modes of transport primarily used in this study: pipeline, rail, and tankers. In the context of this study, the pipeline and rail are used to transport crude oil from western Canada and the US to refineries in eastern Canada. Tanker, and to lesser extent barges, on the other hand, are used to transport crude oil from the US and other parts of the world to refineries in eastern Canada. Figure 2.14 illustrates the complex web of liquids pipelines in eastern Canada and the US. While the pipeline illustrates Enbridge s Line 9 Reversal, the Portland-Montréal Pipeline, and the Energy East Pipeline (as proposed), the map does not show the Enbridge Mainline, transporting crude from AB, SK, and MB to eastern Canadian markets. The Energy East pipeline was cancelled on October 5, It is interesting to note that the Trans-Northern Pipeline, from Nanticoke, ON to Montréal, QC, and the St. Lawrence Pipeline, from Quebec City to Montréal, are transporting refined petroleum products. Also illustrated are the locations of Canada s central and eastern refineries, as well as the rail network in eastern Canada. Rail infrastructure is, however, discussed in greater detail later in this chapter.

40 22 Canadian Energy Research Institute Figure 2.14: Eastern Canadian Crude Oil Transportation Infrastructure Source: (NEB 2016a) Pipeline There are approximately 41,000 kilometers of crude oil pipelines in Canada, transporting 3.7 MMbpd (Canadian Energy Pipeline Association [CEPA] 2017). In addition, there are approximately 77,000 kilometers of natural gas pipelines in Canada, transporting 15.3 billion cubic feet per day (CEPA 2017). While it is not prudent to review over two dozen existing liquids pipelines that either originate in or operate within Canada, it is important to review several pipelines, within the context of this study, that transport liquids to eastern Canada that directly influence the movement of crude oil into eastern Canadian refineries, whether from western Canada or from the US. Western Canadian production is connected to domestic and US refining and export centers, through an extensive network of pipeline but also through rail. For example, most Canadian exports are destined for the US Midwest (PADD II) and the US Gulf Coast (PADD III) the latter being one the world s largest refining centers, with a refining capacity of 9.4 MMbpd. Nevertheless, 400 Mbpd of western Canadian crude oil was transported to refineries in ON, QC and NB, the bulk of which was transported by pipeline (CERI Crude Flow Model).

41 An Economic and Environmental Assessment of 23 The following section is divided into three parts: Enbridge s Canadian Mainline (with a focus on Line 5 & Line 78), Enbridge s Line 9b Reversal and the Portland-Montréal Pipeline. Enbridge Canadian Mainline Enbridge operates the world s longest crude oil and liquids transportation system, slightly over 30,000 kilometers in length (Enbridge 2017a). While moving 64 percent of Canadian exports to the US, it plays an important role in transporting crude oil across Canada, from western Canadian producing regions to eastern Canadian refineries (Enbridge 2017a). Enbridge s Canadian Mainline, or simply referred to as the Enbridge System, begins in Edmonton and runs to Montréal. It is 2,306 kilometers in length (Enbridge 2016a; Enbridge Inc. n.d.). The Canadian Mainline ends at Gretna, MB when the pipeline enters the United States and starts again in Sarnia, ON, where it extends through Toronto and on to Montréal (Enbridge 2016a; Enbridge Inc. n.d.). The Canadian Mainline transports crude oil and diluted bitumen, while the Enbridge Lakehead transports crude oil, condensate and NGLs (Enbridge Inc. n.d.). In combination with the Enbridge Lakehead System the capacity of the pipeline is 2,500,000 bpd (Enbridge Inc. n.d.). The US Mainline, or Enbridge s Lakehead System, connects the western portion of the Canadian Mainline from Edmonton, AB, to Gretna, MB and the eastern portion of the mainline that runs between Montréal and Sarnia, ON. The Lakehead System loops around Lake Michigan, as far north as Lewiston and runs southward through Bay City, Michigan and northwards to Sarnia, ON. The US Mainline is 3,057 kilometers in length and is owned by Enbridge Energy Partners LP (Enbridge Inc. n.d.). Figure 2.15 illustrates the Enbridge Mainline System configuration. The schematic includes 14 lines, or pipelines, as of Q The Canadian Mainline system includes Lines 1, 2, 3, 4, 7, 9, 10 and 11. The US Mainline, on the other hand, includes Lines 5, 6, 14/64, 61 and 62. It is important to note that Line 55 (Flanagan to Cushing) and Line 17 (Stockbridge to Toledo) are not part of the Enbridge Mainline system. The previously mentioned Line 9b, known as the Line 9b Reversal, runs between Sarnia, ON and Montréal, QC. It is discussed separately.

42 24 Canadian Energy Research Institute Figure 2.15: Enbridge s Mainline System Configuration Source: (Enbridge 2016a) Lines 4, 6, 7, 10, 11, 62, 14/64, 61 and 67 can transport heavy crude. Lines 4 and 67 are dedicated to heavy oil. Lines 2, 3, 5, 7, 10, 11 and 14/64 are used, in part, to transport condensates. All the lines, except for Lines 62 and 67 are used to transport various products (i.e., Line 2 is used to transport condensates, light synthetics, sweet crude, and light and high sour crude). The proposed multi-billion-dollar Line 3 Replacement Program expands capacity to satisfy Canadian crude oil production growth, as well as Canadian and US refinery demand (Enbridge 2017c). The Canadian-side and the US-side of the project are both 1,660 kilometers and the pipeline is expected to have an initial capacity of 760 Mbpd (Enbridge 2017d). In establishing crude flows in this study, both Line 5 and Line 78 play important roles. The Mainline enters Sarnia via two entry lines Line 5 and Line 78 with a total capacity of 1,040 Mbpd. The former transports western Canadian crude along Lines 1, 2, 3 and 67 to Superior, Minnesota, where it is transported to Sarnia. Line 78, on the other hand, expands Enbridge s capacity to transport crude produced in the Williston Basin region, North Dakota, and light and heavy crude production in western Canada and also transports product to Sarnia (Enbridge 2017b). Line 78 begins in Flanagan, Illinois and terminates in Sarnia, via Griffith/Hartsdale and Stockbridge (Enbridge 2017b). The pipeline runs parallel to Line 62 between Flanagan and Griffith/Hartsdale. As ON refineries consume 344 Mbpd (total refining capacity 393 Mbpd) (see Table 2.1), the incoming pipeline capacity is enough to carry more Canadian oil east of ON.

43 An Economic and Environmental Assessment of 25 Enbridge Line 9 Reversal While Line 9 is also operated by Enbridge, it operates outside of the Mainline, and is consequently discussed separately. In addition, Line 5 is being expanded and Line 9B reversed from Westover, ON to Montréal (Enbridge 2017e). The latter was approved by the NEB on March 6, 2014 (Enbridge 2017e). The mandate was to provide crude oil to refineries in Montréal from western Canada and the Bakken region in North Dakota. The pipeline originally shipped oil from the North Sea, West Africa, and the Middle East from Montréal to refineries in Sarnia. In 2012, Enbridge applied to reverse the flow direction and expand the capacity of Line 9, from 240,000 bpd to 300,000 bpd (NEB 2016b). Line 9, which now carries light crude from the Williston Basin to refineries in QC, allows refineries in QC to use North American light oil as a feedstock instead of importing crude oil from overseas, which trades at a premium to North American grades of crude. Figure 2.16 illustrates Enbridge s Line 9. Figure 2.16: Enbridge s Line 9 System Configuration Source: (Enbridge Geomatics Services 2011) In QC, the two refineries in Montréal and Lévis can be accessed by western and eastern Canadian oil producers. Montréal s refinery, with oil intake of Mbpd (total capacity 137 Mbpd), can be fully supplied from the west through the 300 Mbpd Enbridge Line 9, which is connected to Sarnia.

44 26 Canadian Energy Research Institute Portland Montréal Pipeline Today, the Portland-Montréal Pipeline consists of two pipes: one 18-inch diameter pipe, and one 24-inch diameter pipe. The two pipes that connect Montréal to Portland tidewater have a combined engineering capacity of 602,000 bpd of light crude (24 410,000 bpd, ,000 bpd) (Gillies 2009). The pipeline reported no throughputs from January to March 2016, likely in part due to the beginning of Enbridge s Line 9 Expansion and the Line 9b Reversal (NEB 2017d). QC refiners could access western Canadian oil through the latter. The reversal of Line 9 has also had an impact on oil tanker traffic in Portland. Not even a decade ago, with nearly 200 tankers each year, Portland was the largest oil port on the US east coast (Bell 2008). While refineries in QC are likely using a combination of light crude oil from Enbridge s Line 9, as well as overseas imports through the Portland-Montréal Pipeline, the pipeline s utilization rate was actually 22 percent in 2015 (61 Mbpd out of 280 Mbpd capacity in 2015, according to the NEB (NEB 2017d). There is speculation that the Portland-Montréal Pipeline could be reversed. By reversing the flow direction of one of the pipelines, western Canadian heavy oil could be transported from AB to Portland, where crude oil is later loaded onto tankers and shipped to refineries along the US Gulf Coast for refining. The Portland Pipe Line Corp. filed a lawsuit against the city of South Portland that has blocked the company from reversing the pipeline s flow (that the company would immediately undertake if not prohibited by legal barriers) (Bouchard 2017). Crude-by-Rail Non-intermodal traffic is primarily comprised of bulk cargo and tank cars, and includes commodities such as petroleum products, wheat, coal, or potash. Railcars are used to transport petroleum fuels (gasoline, diesel, aviation fuels, fuel oil and lubricants), chemical products (ethylene glycol, chlorine, ammonia, vinyl chloride and caustic soda), and LPG products (propane, butane, and pentanes). Crude-by-rail emerged several years ago, stemming from various regulatory issues that proposed pipelines faced. Rapid growth in western Canadian crude oil production (especially from oil sands operations) has outpaced pipeline capacity and pipeline companies expansion efforts. In more recent years, rail transport of crude oil has subsided as an alternative mode of transport to accommodate surplus volumes that exceed pipeline capacity, as the economic downturn reduced or eliminated the surplus. Figure 2.17 shows the current Canadian railway network for crude oil transport, and outlines the key shipping terminals and main refining centers receiving the product.

45 An Economic and Environmental Assessment of 27 Figure 2.17: Canadian Railway Network for Crude Oil Transportation Source: (CAPP 2017a) By global standards, Canadian National (CN) and Canadian Pacific (CP) are major players, both considered Class I rail carriers by US definition. As of June 8, 2017, CN had a market value of US$55.9 billion (ranked second in the world behind Union Pacific at US$86.9 billion) while CP had a market value of US$22 billion, ranked eighth in the world, between Central Japan Railway and China s Daqin Railway (Statista 2017). Of the two Canadian railways, CN has the longest railway network in Canada, reaching from the Pacific coast in BC to the Atlantic coast in NS, and extending to the US Gulf Coast. CN operates in 8 provinces and 16 US states (Meyer 2009). Oil from western Canada can be transported from as far north as Fort McMurray to marine terminals in Vancouver, to terminals in eastern Canada, as well as to refineries in the southern US and US Gulf Coast. CN-served ports include the Port of Halifax, Port Montréal, Prince Rupert Port Authority, Port Metro Vancouver (PMV) and Port of New Orleans (CAPP 2017a).

46 28 Canadian Energy Research Institute While CP was Canada s first transcontinental railway, it now does not reach the Atlantic coast. Its rail network stretches from Vancouver to Montréal, and as far north as Edmonton. The CP rail network serves several major US cities, such as Minneapolis, Detroit, Chicago, and New York. Various arrangements give CP access to terminals in Detroit, Buffalo, New York, Philadelphia, and Binghamton. Intermodal terminals located in Canada and the US include Vancouver, Calgary, Edmonton, Regina, Saskatoon, Winnipeg, Montréal, Minneapolis, Milwaukee, Detroit, and two terminals in Toronto and Chicago (CAPP 2017a). CP-served ports include Port Montréal, Port Metro Vancouver, and Port of Thunder Bay. CP s rail network is frequently divided into four primary corridors: Western, Southern, Central and Eastern (US Securities and Exchange Commission 2006). With the purchases of the Dakota, Minnesota and Eastern Railroad (DM&E) and the Iowa, Chicago and Eastern Railroad (IC&E) in September 2007, CP s rail network extends into the US Midwest as far as the natural resource-rich Powder River Basin in Wyoming (Dowd 2007). Additionally, CP connects with Kansas City Southern (KCS), Norfolk Southern (NS) and Union Pacific (UP) at Kansas City and connects with UP at Minneapolis and Minot, North Dakota (Canadian Pacific 2016). CAPP estimates current rail loading capacity originating in western Canada at 754,000 bpd (CAPP 2017b). Most of this capacity has come online in recent years. Table 2.3 illustrates the operator, location, and capacity of rail offloading terminals in eastern Canada. Table 2.3: Crude-by-Rail Offloading Terminals in Eastern Canada Terminal Operator Location Loading Capacity (Bpd) Irving Oil Rail Terminal Irving Oil Saint John, NB 145,000 Sorel-Tracey Rail Terminal Kildar Services ULC Sorel-Tracey, QC 33,000 Montréal Rail Terminal Suncor Energy Montréal, QC 30,000 Valero Rail Terminal Valero Lévis, QC 60,000 Nanticoke Rail Terminal Imperial Nanticoke, ON 20,000 Data Source: (Oil Sands Magazine 2017b). Table by CERI. Irving s oil rail terminal in Saint John is the largest in eastern Canada and the country s second largest. While Irving Oil in NB stands on the water and has direct access to Eastern offshore oil (47,000 barrels of eastern crude were used in 2016), the refinery also has a large rail offloading capacity of 145 Mbpd, the largest in Canada. Usage of rail however has decreased in recent years, as Irving Oil used to import around one-third of its intake by rail. This decrease also coincides with the fatal accident in Lac-Mégantic on July 6, 2013, when a train derailment resulted in 47 deaths. The unattended train, operated by Montréal, Maine & Atlantic Railway (MMA), was carrying crude oil from the Bakken formation to the Irving Oil refinery (Transportation Safety Board of Canada 2014).

47 An Economic and Environmental Assessment of 29 Recall, Montreal s Suncor Energy refinery can receive up to 35 Mbpd by rail. Valero can also receive oil from the west by rail (up to 60 Mbpd) as well as eastern oil directly by water to its offloading facilities. In western Canada, there are 14 rail terminals in AB and 13 terminals in SK (Oil Sands Magazine 2017b). The largest three terminals in western Canada are Kinder Morgan s Edmonton facility (210,000 bpd), Gibson/USDG s Hardisty terminal (140,000 bpd) and Cenovus Bruderheim terminal (100,000 bpd) (Oil Sands Magazine 2017b). Figure 2.18 illustrates the crude-by-rail exports by quarter to the US. The figure includes data from the NEB and the EIA. Monthly volumes as of July 2017 are 92.5 Mbpd, down considerably from the record high of 179 Mbpd in September 2014 (NEB 2017h). Figure 2.18: Quarterly Volumes of Crude Oil Exported by Rail to the US Source: (NRCan 2017c) Oil Tankers and Marine Terminals There are two types of oil tankers: crude tankers and product tankers. The former moves large quantities of unrefined crude oil to refineries while the latter transports petrochemicals from refineries to market for consumption. This section focuses on the crude tankers. Oil tankers are divided into subclasses: Product Tanker/Seawaymax, Panamax, Aframax, Suezmax, Very Large Crude Carriers (VLCC) and Ultra Large Crude Carriers (ULCC) (Rodrigo de Larrucea 2010). Table 2.4 illustrates the typical minimum and maximum deadweight tonnage (DWT).

48 30 Canadian Energy Research Institute Average Freight Rate Assessment (AFRA) Scale Table 2.4: Oil Tanker Size Flexible Market Scale Class Size in DWT Class Size in DWT General Purpose Tanker 10,000-24,999 Product Tanker 10,000-60,000 Medium Range Tanker 25,000-44,999 Panamax 60,000-80,000 LR1 (Large Range 1) 45,000-79,999 Aframax 80, ,000 LR2 (Large Range 2) 80, ,999 Suezmax 120, ,000 VLCC ULCC 160, , , ,999 Source: Evangelista 2002, as referred to in (Rodrigo de Larrucea 2010). Modified by CERI. VLCC 200, ,000 ULCC 320, ,000 Product tankers and Panamax vessels are known as product tankers or general purpose or Medium Range (MR) tankers, by the Average Freight Rate Assessment (AFRA) scale. These tankers often carry refined petroleum products. Panamax, or the Large Range 1 (LR1), as the name suggests, refers to ships that are able to travel through the Panama Canal with a crude capacity of up to 500,000 bbls (Marine Insight 2016). The Panama Canal, opened in 1914, has undergone an expansion so that ever-growing tankers can make use of the facility. New Panamax vessels with dimensions of 427 m in length, a 55 m beam and depth of 18 m will be able to use the new locks (Marine Insight 2016). The new Canal dimensions will not be able to fit vessels categorized as VLCCs and ULCCs, or supertankers. Aframax tankers, or Large Range 2 (LR2), have a DWT of less than 120,000 and the name is based on the AFRA tanker rate system (Marine Insight 2016). The latter was a rating system started by Shell in 1954 to categorize the size and purpose of vessels (AUUUU.com 2015). Aframax tankers are generally used in the North Sea, Black Sea, the Caribbean Sea, the China Sea and the Mediterranean (Marine Insight 2016). Aframax tankers have a capacity of approximately 650,000 bbls (National Research Council 1998). Suezmax, ranging between 120,000 and 200,000 DWT, refers to the category of vessels able to pass through Egypt s Suez Canal (Rodrigo de Larrucea 2010). Once regarded as supertankers, this distinction is now reserved for the much larger VLCC and ULCC vessels. Suezmax tankers have a capacity of approximately 1,000,000 bbls (Kinder Morgan Canada 2013). The VLCC can transport between 200,000 and 320,000 DWT and average 331 meters in length and 60 meters in width. The capacity of a VLCC is approximately 2,000,000 bbls (US EIA 2014; Maritime-Connector.com n.d.; AllOilTank.com n.d.). The latest generation of supertanker is the Ultra Large Crude Carrier (ULCC) and has a capacity of up to 550,000 DWT, or approximately 4,000,000 bbls (US EIA 2014; Maritime-Connector.com n.d.; AllOilTank.com n.d.). It is interesting to note that because of their sheer size, they are usually not permitted to enter a port fully loaded (Huber 2010).

49 An Economic and Environmental Assessment of 31 As of end-2016, there were 688 VLCCs, 509 Suezmaxes/LR3, 971 Aframaxes/LR2, 430 Panamaxes/LR1 and 1,986 MR-type (TankerOperator 2017). For comparison, as of January 2008, there were 492 VLCCs, 360 Suezmaxes, 783 Aframaxes, 329 Panamaxes and 981 MR-type (or Seawaymaxes) (TankerOperator 2008). While there are numerous marine terminals in Canada, there are seven marine terminals suitable for loading and offloading oil. Table 2.5 illustrates oil marine terminals in Canada. It is interesting to note that a single terminal is found on the west coast, the Westridge Marine Terminal in Burnaby. The terminal lies at the end of the TMX pipeline, transporting crude oil from Edmonton, AB to Burnaby, BC. Table 2.5: Oil Marine Terminals in Canada Name Operator Location Port of Come By Chance North Atlantic Refining Come By Chance, NL Canaport Crude Receiving Terminal Irving Oil Saint John, NB Port of Sainte Victoire Kildar Services ULC Sorel-Tracey, QC Port of Montréal Marine Terminal Suncor Energy Montréal, QC Port of Montréal East Suncor Energy Montréal, QC Port of QC Valero Lévis, QC Westridge Marine Terminal Kinder Morgan Canada Burnaby, BC Source: (Oil Sands Magazine 2017a) Montréal can also accept at least Panamax (60,000-80,000 DTW) size tankers directly from the eastern offshore projects.

50 32 Canadian Energy Research Institute

51 An Economic and Environmental Assessment of 33 Chapter 3: Methodology and Analysis This Chapter reviews the methodology of various models and calculations used in this study. It is divided into two sections: a) a discussion of the methodology and scenarios used and b) the models and their approaches. The latter discusses the various modelling approaches and their assumptions. It is subsequently divided into three parts: the crude flow model, the cost of feedstock model and the emissions/lca model. Each model is applied to the previously mentioned four scenarios, as well as the Base Case. Overview of Methodology and Description of Scenarios This section provides an overview of the scenarios and the various modelling approaches used in this study. Greater details of the models used are discussed separately later in this chapter. The scope of this study is to analyze the potential complete or partial substitution of imported foreign oil in the central and eastern Canadian refinery market with domestically-sourced oil supply. The research provides a cost and emissions comparison of four potential scenarios of feedstock for eastern Canadian refineries, substituting domestic versus foreign crude in central and eastern Canadian refinery markets. Modelling the different scenarios and their effects are in turn compared to the Base Case. The Base Case provides a snap shot of the crude flows in 2016, representing the reconstruction of current flows (Canadian and imported feedstock supply). It also explores the refineries themselves, their technologies, as well as existing and potential transportation routes for western and eastern Canadian oil to central and eastern refineries, as used in The data however is incomplete and CERI made assumptions where information is either missing or suppressed in the collected statistics. For instance, if crude brands of imported oil are not certain per refinery from an export country, the most abundant brands available in the export country were taken. Or, if there is no certainty of how crude oil is supplied per refinery from western Canada, an assumption was made based on the distances from the western crude sources, railway network and other available information. The objective was to reconstruct the existing flows accurately. While there are an infinite number of potential scenarios for displacing central and eastern refineries foreign oil imports, CERI created four scenarios defined in terms of two separate uncertainties: the level of infrastructure available to transport crude to central and eastern refineries transportation infrastructure (existing versus expanded) and the general approach of decision-making (a market- or policy-based approach). This is illustrated in Figure 3.1.

52 34 Canadian Energy Research Institute Figure 3.1: The Four Scenarios Expanded Infrastructure Social Push/ Policy-based Approach Made in Canada International Social Concerns Expanded Access Current Reality Economic Pull/ Market-based Approach Existing Infrastructure Mirroring the four quadrants and their scenarios in Figure 3.1, this research explores four key questions: 1. How much additional oil can be economically sourced through existing infrastructure to central and eastern refineries (from west and from east) substituting foreign oil, and how will this impact emissions? In this case, more expensive foreign oil volumes are substituted. 2. How would the cost of feedstock and emissions levels differ if oil from authoritarian states was substituted? 3. If there was additional transportation pipeline infrastructure (such as the cancelled Energy East project), how much more economic oil could be brought to central and eastern refineries and at what emissions levels? 4. What would the cost of feedstock and emissions levels be if all foreign crude was substituted with Canadian crude using expanded transport infrastructure? Before exploring the four scenarios, it is prudent to review several major assumptions. First, this research is not from the perspective of the producers and transportation sector, but from the refinery level, and more particularly, from central and eastern Canadian refineries. As such, this study explores the cost impacts in course of substitution for the refineries, but not for producers or transportation companies. Note that in this study, midstream is defined as the

53 An Economic and Environmental Assessment of 35 refining sector, while services of oil delivery (by any mode) is defined as the transportation sector. This deviates from the industry conventional value chain breakdown, but it is done so to be more consistent with LCA models used for the study (typically the transportation sector would be under midstream, and refining would fall under downstream activities). Second, this study assumes that Western Canadian crude oil used by eastern Canadian refineries in the various scenarios is assumed to be taken from the current export markets and redirected to eastern Canadian refineries. As such, this analysis does not assume incremental production to offset Canadian crude oil displacing foreign crude in eastern Canadian refineries. In addition, it is beyond the scope of this study to delve into downstream markets and implications due to substitution, if any. CERI does not explore where the refined petroleum products go, but focuses on displacing like-for-like foreign oil with domestically-sourced equivalents (i.e., light-for-light, etc.). The latter assumption enables CERI to assume that refineries yields in the Base Case and in the various scenarios remains the same, and allows a comparison of costs of feedstock and emissions between different hypothetical scenarios while keeping the sales variable constant. Recall, this a scenario-based approach and, as such, is not a forecast. This means there is limited impact on GDP associated with indirect and induced economic activities from changes in the trade balance (lower value of imports). The four scenarios and their main assumptions are summarized below. Current Reality adopts a market-based approach, optimizing for cost of feedstock and emissions. However, the difference is that it assumes crude oil can be transported from western and eastern Canada to eastern Canadian refineries via existing infrastructure pipelines, marine tankers, or rail. Crude oil substitutes feedstock in eastern Canadian refineries utilizing three modes of transport. Enbridge Mainline (Line 5 and Line 78) connects Hardisty, AB to Sarnia, ON, while Enbridge s Line 9 reversal transports oil from Sarnia to Montréal. Quebec City is connected via two tankers from Montréal. This scenario will identify the throughput capacity of existing pipelines and rail systems to move additional oil from west to central and eastern refineries. The remainder will be supplied by foreign crude oil from the sources which supplied central and eastern refineries in This is a market-based approach to substituting foreign crude oil in Canadian refineries meaning that more expensive foreign crude is displaced; cheaper foreign crude is kept in the slate of refineries. Proponents include the transportation sector and refineries. Of the four scenarios, Current Reality represents the closest to status quo, simply a more optimal solution regarding costs of feedstock and emissions utilizing the existing infrastructure available. Expanded Access is a market-based approach. Western and eastern Canadian crude oil is transported to central and eastern refineries via an expanded transportation infrastructure. The difference here is that Expanded Access allows for more domestic volumes to be substituted economically. Thus, even if more is available, if Canadian oil is more expensive at the refinery gate compared to foreign oil, in this market-driven decision making a refinery will prefer the cheaper imports (if it can get the same desired yields). Prices of particular brands of foreign and

54 36 Canadian Energy Research Institute domestic oil as well as transportation costs play a role in the selection of the crude intake slate. Tankers are cheaper transport options than pipelines and rail. As western production is far from the central and eastern refinery market, transportation costs by pipeline and rail will almost always be at a cost disadvantage to foreign tanker costs. In other words, in the Expanded Access scenario, more expensive foreign crude is displaced; cheaper foreign crude is kept in the slate of refineries. Made in Canada assumes that all Canadian crude, be it from western or eastern parts of the country, will substitute for all foreign imported crude oil via an expanded transportation infrastructure. The pipeline infrastructure with a new pipeline is used and transports oil from Hardisty to Saint John, NB, via Montréal and Lévis. This is in addition to existing infrastructure, mainly the Enbridge Mainline (Line 5 and Line 78), which connects Hardisty, AB to Sarnia, ON, while Enbridge s Line 9 reversal transports oil from Sarnia to Montréal. While western Canadian crude is used in refineries in ON and QC, as well as a portion of Irving refineries feedstock, the remaining refineries and feedstock demands are satisfied with eastern Canadian oil production. This scenario is often categorized by a nationalist or protectionist stance, and is either highlighted by security of supply or keeping the macroeconomic benefits within Canada, particularly for a country which is a major producer of oil. International Social Concerns assumes that crude oil will be transported to eastern markets via the existing pipeline infrastructure and rail. This scenario, however, assumes that domesticallysourced crude oil will substitute for foreign crude oil coming from countries that have generated concern by international organizations for their treatment of their citizens or their environment. CERI reviewed several indices by internationally recognized organizations including those produced by the World Bank (various), the United Nations (Human Development Index) and the Freedom House, as well as other organizations. The indices range from ranking the democratic process to human rights to quality-of-life. For this study, however, CERI utilized the widely-cited Economist Intelligence Unit s Democracy Index Often cited in peer-reviewed journals, the Democracy Index includes five diverse elements to determine a nation s classification: electoral process/pluralism, functioning of government, political participation, political culture, and civil liberties. While it is not perfect, it provides a proxy for reflecting social concerns. Nations fit into four categories: full democracy, flawed democracy, hybrid regime, and authoritarian (The Economist Intelligence Unit 2017). Canada, Norway, Denmark, the United Kingdom, Colombia, and the US are classified as either full or flawed democracies. Except for Nigeria (hybrid regime), the rest of the countries that supply Canada s imported oil are classified as authoritarian, according to the index. As such, this scenario substitutes all foreign oil that are not classified as either full or flawed democracies. This is a policy-based approach to substituting foreign crude oil in central and eastern Canadian refineries. It is important to note that more expensive oil from democratic states is kept in the crude slate. This approach allows CERI to assess the direct impact on costs and emissions tied to substitution of oil from states that have caused international concern for the treatment of their citizens or environment.

55 An Economic and Environmental Assessment of 37 The four scenarios are highlighted by two critical uncertainties, based on the level of infrastructure available to transport crude to central and eastern refineries (y-axis) and the general approach of decision-making (x-axis), whether a market- or policy-based approach (Figure 3.1). CERI identifies and quantifies the costs to refiners and effects on emissions in four very different scenarios, guided by two critical uncertainties. Each of which contain several unique assumptions. First, the transportation-related uncertainty determines whether the scenario is influenced by transporting crude through existing pipeline, railway and marine infrastructure or transporting crude through existing pipeline, railway, and marine infrastructure plus additional, or expanded, pipeline infrastructure. The available capacity of transporting crude is an important factor, determining in part the amount and how crude is transported to central and eastern refineries, displacing foreign crude oils. Existing transportation infrastructure allows more Canadian crude from the west and east to central and eastern Canadian refinery markets, however, it is not sufficient to substitute all foreign oil. The major infrastructural bottleneck for western crude is Line 9 s capacity to supply any refinery east of Montréal. This affects the Made in Canada and Expanded Access scenarios. In these two cases of an expanded infrastructure, CERI assumes a new pipeline extending from Hardisty, AB to Saint John, NB via Montréal and Lévis, QC. To simplify matters, CERI assumes that the route of the pipeline, its capacity and toll structure are the same as the now-cancelled Energy East pipeline, proposed by TransCanada Pipeline (TCPL). Energy East was cancelled on October 5, Second, CERI also examines whether partially or completely substituting foreign crude oil is fueled by a market-based approach (or economic pull) or policy-based approach (or social push or government policy). The general approach of decision-making and whether a market- or policy-based approach, the research question of partially or completely substituting foreign crude oil is fueled by various arguments. While each has their own motives, they are typically driven by economic (a market-based approach) or social rationale (a policy-based approach). The International Social Concerns scenario is reflective of a policy fueled by social push stemming from an increasing awareness in elements of society asking, where do the products we consume originate? or what are the labour and environmental standards for manufacturing in those countries we buy from? This trend has ramifications not only for the energy sector but other goods and services that Canada trades. CERI is not promoting policy or evaluating its relevance but exploring examples of policy initiatives that could possibly affect the substitution of foreign crude oil and thus the central and eastern Canadian refineries. As such, CERI does not discuss the type of policy instruments to implement such programs, or their effect on costs and emissions on the refinery-level. This CERI study will show market- and socially-pushed scenario outcomes, via existing and expanded transportation infrastructure, which could serve as a good indication for Canadians on how beneficial or costly such substitution could be. This research is a first of its kind comparison

56 38 Canadian Energy Research Institute that has not been done by any organization. However, in allowing CERI to make a comparison across the four scenarios, CERI uses its proprietary Crude Flow Model and its Costs of Feedstock Model, as well as utilizing a life-cycle assessment (LCA) of greenhouse gas (GHG) emissions of individual crude oils used in the study. Recall, CERI is substituting like for like crude oils, substituting crude oil from all over the world with the like Canadian equivalent. The objective of the emissions modelling is to establish an emission intensity profile for each crude oil used in the central and eastern Canadian refinery market. The latter is done through utilizing the publicly available GHG estimation models: OPGEE (upstream emissions) and PRELIM (midstream emissions). Figure 3.2 illustrates the four different modelling approaches and calculations and the relationship between them. There are several major components, or elements, to this study. Using the established crude flows for each of the scenarios, the following are, in a sense, the inputs into the costs of feedstock model and the emissions model. Costs of feedstock and emissions are influenced by the flow of crude in the four scenarios. Modelling the different scenarios, their effects are in turn compared to the Base Case a snap shot of crude flows in Figure 3.2: Schematic of the Modelling Approaches Crude Flows Model Costs of Feedstock Model Emissions/ LCA Model The crude flows from western and eastern Canada that displace foreign oil by the various scenarios is a critical component, impacting the other models. Costs of feedstock model and the emissions modelling determine the various costs and emissions of the four scenarios. Each of these models and their assumptions are discussed in greater detail in the following section. It is, however, important to mention that, despite the complexity of the various approaches, one of the most challenging elements of this study is the lack of data, particularly on the refinery level. One of the early core research questions was whether partial or entire substitution of imported crude in the central and eastern refineries will be of benefit or cost to Canada. This

57 An Economic and Environmental Assessment of 39 was initially answered by utilizing the cost-benefit analysis (CBA) from the unique perspective of the eastern Canadian refiner. Providing a robust method for evaluating the costs and benefits, including economic and non-economic impacts, this method is ideal for comparing a policy change, projects, or a mode of action in today s dollars to a society. And while the CBA is far from perfect, the CBA provides a framework that identifies, quantifies, and compares the costs and benefits of a proposed policy or mode of action. The CBA approach was, unfortunately, significantly limited by the data available, and was subsequently altered to the approach instead. Models and Their Approaches The following discusses the four models used in this study, their assumptions, and forecasts, where applicable. This section is divided into three parts: crude flows model, cost of feedstock model and emissions/lca model. The discussion is based on the Base Case. This section discusses the how and the why, pertinent to each model. The results of these models within the context of each of the four scenarios are discussed in Chapter 4. Crude Flows Model CERI s Crude Flow Model lies at the core of the modelling approach used in this study. The crude flows from western and eastern Canada that displace foreign oil are determined by the various scenarios as a critical component, impacting the other models. Costs of feedstock, emissions and calculating the economic effects are influenced by the flow of crude in the four scenarios. This section is divided into three parts: establishing a Base Case, calculating the stock of Canadian crude oil available for substitution purposes and discussing the assumptions used in modelling the scenario s crude flows. Establishing the Base Case This section delves into the eight eastern Canadian refineries and various important metrics, including their location, capacity, utilization rates, crude intakes (or crude slate) and mode of supply (transportation). The data represents Consumption of feedstock (crude slate) and product slate of the eight refineries is reviewed, establishing the current state of refineries in eastern Canada. The crude slate is reviewed more closely, breaking it down by domestic-sourced crude and foreign-sourced crude. The data for this study was obtained from several sources as mentioned in Chapter 2. If the data was suppressed, missing, or conflicted, CERI reconciled using multiple sources to infer quality and certainty in data. Refining capacity in eastern Canada is represented by eight refineries with a total capacity of 1,228 Mbpd. Four refineries reside in ON, two in QC, and one each in NB and NL & Labrador. All eight refineries, their capacities, utilization rates, crude intakes and mode of supply are illustrated in Table 3.1.

58 40 Canadian Energy Research Institute Table 3.1: Crude Intake by Eastern Canadian Refineries, by Oil Type (Mbpd) Province Refinery Capacity Capacity by Province ON QC NB Imperial, Sarnia Shell Canada, Corunna Suncor Energy, Sarnia Imperial, Nanticoke Suncor Energy, Montreal Utilization Rate (%) Total Intake Light SCO Heavy Bitumen Available Mode of Transport 86% Pipeline 75 88% Pipeline 85 92% Pipeline % Pipeline, Rail 92% Pipeline, Tanker, Rail Valero, Lévis % Irving Oil, Saint John % Tanker, Rail NL North Atlantic Refining, Come by Chance % Tanker Total 1,228 1,228 1, Percent 73% 15% 7% 5% Data Source: CERI Crude Flow Model. Table by CERI. Note that for this study, the intake for QC was altered from the factual data of In 2016, QC refinery utilization rates did not seem to be representative across the refining industry in Canada (an average 78 percent for two refineries based on CERI calculations of Statistics Canada data) (Statistics Canada 2017c, 2017d). For the purposes of modelling, an adjusted utilization rate of 88 percent was assumed for Valero, and 92 percent for the Suncor Energy refinery. Suncor Energy s estimate was taken from the company Annual Report (Suncor Energy Inc. 2017a). For reference, the factual 2016 import volume in QC was Mbpd. However, due to an increase in utilization rates, the modelled 2016 import volume for QC is 9.8 percent higher than factual 235 Mbpd (Government of Canada 2017). Hence, eastern offshore and western supply were also increased by the same rate. Modelled and factual imports are presented in Appendix A. Light oil prevails in the supply slate for the four provinces comprising 73 percent, followed by synthetic crude (SCO) with 15 percent in the crude slate; heavy and bitumen both comprise 12 percent of the crude intake. This is illustrated in Figure 3.3.

59 An Economic and Environmental Assessment of 41 Figure 3.3: Crude Intake by Type by Province (Mbpd) Data Source: CERI Crude Flow Model. Figure by CERI. Fifty-six percent of oil into the eastern region is imported, while 39 percent comes from western Canada and five percent is supplied from Canadian eastern offshore assets (offshore NL & Labrador as well as 0.6 Mbpd from ON). Among eastern Canadian refineries, ON receives the highest amount of western crude (almost 83 percent), followed by QC (33 percent) and NB (4 percent). NL & Labrador does not receive oil from western provinces. The largest exporters to Canada are the US, Saudi Arabia, Algeria, Nigeria and Norway (Government of Canada 2017; NEB 2017f; Statistics Canada 2017a). Detailed information on the sources of oil for eastern refiners is provided in Table 3.2 (imports per country in the last column are rounded; detailed information can be found in Appendix B). Based on the import data, up to 601 Mbpd could be substituted by Canadian crude oil.

60 42 Canadian Energy Research Institute Province ON Refinery Table 3.2: Crude Intake by Eastern Canadian Refineries, by Source Total Intake (Mbpd) Western Canada Supply (% / Mbpd) Eastern Canada Supply (% / Mbpd) Imported (% / Mbpd) Imperial / / 17.6 Shell Canada / / 11.1 Suncor Energy / / 13.1 Imperial / / / 16.3 ON / / / 58.1 QC Imports from Countries (Mbpd (Rounded)) US 58 Suncor Energy / / / US 100, Algeria 92, Kazakhstan 21, Valero / / Nigeria 12, Others 10 QC / / / NB Irving Oil / / / NL North Atlantic Refining / / 89.8 Atlantic Canada / / / Total 1, / / / Data Source: CERI Crude Flows Model. Table by CERI. Saudi Arabia 87, Nigeria 45, US 34, Norway 26, Others 27 US 49, Nigeria 18, Norway 14, Others 9 The following describes the flows of crude to the eastern refinery market in the Base Case. To obtain this, CERI used many sources as previously mentioned in Chapter 2. Statistics Canada information presented a detailed view of ON and QC intake of oil from western and eastern provinces, as well as intake by province by type of crude bitumen, heavy, light, and SCO. Statistics Canada trade information showed volumes and source of foreign oil per province (Statistics Canada 2017a). In ON, North Dakota was the foreign exporter. Its volume was distributed proportionally to all four refineries, as well as SCO. Bitumen was allocated to Imperial Oil s refinery as it is the refinery in ON equipped with the coker technology. In the case of QC, the volume of western, eastern and foreign oil was taken from Statistics Canada data on the provincial level. Clipper Data on tanker movements was used for apportioning specific volumes to a particular refinery as well as identifying transportation routes to refineries (ClipperData LLC 2017). The crude tanker shipments from Montreal to Valero data helped to allocate total volume of western oil between the two refineries. Likewise, shipment data to Portland, Maine (the origin of the Portland-Montreal pipeline), Quebec City, and Montreal ports and combining this information with Statistics Canada trade data, allowed CERI to allocate foreign oil volumes between the two refineries, as well as determine Canadian eastern offshore volumes. Utilizing Clipper Data, particular brands of imported oil were identified, which helped to assign light and heavy imported oil between refineries more accurately. This data was useful in

61 An Economic and Environmental Assessment of 43 reconciling the difference with Statistics Canada data as the latter showed a large volume of SCO which was not consistent with total volume of western crude (ClipperData LLC 2017; Statistics Canada 2017a). Thus, part of what Statistics Canada marked as SCO was marked as light crude oil in the study Base Case. Heavy and bitumen volumes were also adjusted slightly between these categories to reconcile refinery intake information, oil shipments data, and Statistics Canada data. NB s Irving oil feedstock was more easily established than QC s, as the refinery stands on water, and the bulk of their intake arrive in tankers. Statistics Canada trade data (Statistics Canada 2017a) as well as Clipper Data (ClipperData LLC 2017) were sufficient to arrive at volumes of foreign and eastern Canadian crude and type of oil. Western crude was derived as the difference between the total intake and the amount delivered by tankers. NL s North Atlantic Refining s intake was also more easily established at it also stands on water allowing for using both Statistics Canada trade data and Clipper Data to get a full picture of feedstock supply. Transportation routes were identified for all the refineries using multiple sources. As each refinery enjoys specific transportation options, combining a number of sources allowed CERI to establish a full picture of supply routes. Statistics Canada provided information that ON was fully supplied via pipelines (Line 5 and 78 enters ON). QC s crude oil supply was established via solving for volumes from tanker shipments for Valero and Portland-Montreal pipeline intake; the remaining portion was delivered via Line 9. Rail usage in the east was identified as follows. As 2016 data was limited, the year 2015 was studied in more detail to understand shipment volumes and routes. Data from 2015 showed domestic rail movements total volume of 53 Mbpd (NEB 2017e). Rail movements in 2015 for fuel oil and crude petroleum followed routes such as a) from AB to QC, ON and NB, b) from SK and MB to QC and ON, and c) from the US to QC (Canadian Pacific 2016; CAPP 2017a, 2017b; Statistics Canada 2017b). Domestic rail usage in the east was identified as limited to a relatively small number in Mbpd (CERI estimate). The number comes as a difference between movements of oil from west in 2016 in the amount of 100 mbpd (CAPP 2016), and export by rail data for 2016 (NEB 2017h). To simplify the modelling, all 11.9 Mbpd in 2016 were assigned to NB s refinery, rather than being distributed between the two refineries in QC and Irving Oil. For reference, railway offloading capacity is estimated at 255 Mbpd for all the central and eastern refineries (namely 20 Mbpd for Nanticoke, ON, 60 Mbpd for Lévis, QC, 30 Mbpd for Montreal, QC, and 145 Mbpd for Saint John, NB) (CAPP 2017b). Finally, to identify which western province supplied which province in the east, Statistics Canada data (Statistics Canada 2016) was used for 2015 as a proxy which showed that MB and SK s volumes of oil went to ON (as was assigned as such in the Base Case). Thus, ON s demand for feedstock was first satisfied predominantly from these provinces; the rest came from AB. AB oil, on the other hand, was assigned to all provinces except for NL. Bitumen, heavy oil and SCO was assigned exclusively to AB in the Base Case. In addition, as rail movements in 2015 (for export

62 44 Canadian Energy Research Institute and domestically) predominantly originate from AB and Atlantic Canada s supply in that year was also from AB, the above-mentioned volume 11.9 Mbpd was also allocated to this province. Below is a description of crude flows in the Base Case, based on assumptions and data sources. The single foreign exporter to ON s four refineries was the US (58 Mbpd). Light oil comes from North Dakota and is modelled in this study to be transported through Enbridge Line 81 to Clearbrook, MN and further through the Enbridge Mainline system (Line 5 entering Sarnia). Canadian light, heavy, and bitumen crude from three provinces AB, SK, and MB flow through various lines of the Enbridge Mainline and enter ON via Line 5 and 78. AB supplies 206 Mbpd, followed by SK s 72.6 Mbpd and MB s 6.4 Mbpd. Suncor s Montreal refinery imports light oil from four countries US (92.7 Mbpd), Azerbaijan (4.5 Mbpd), UK (4.1 Mbpd) and Norway (1.9 Mbpd). For crude that comes from the US, 62.1 Mbpd comes from North Dakota and Michigan through the Enbridge Mainline and Line 9 (going from Sarnia to Montreal). All other crude, including 30.6 Mbpd from Texas, is transported by tankers from loading ports to Portland, Maine and then transported via the Portland-Montreal pipeline to the refinery. Canadian supplies include 16.4 Mbpd of bitumen and 5.8 Mbpd of light crude that comes from eastern offshore assets. Bitumen comes through the Enbridge Mainline and Line 9, while the eastern crude follows the path of international oil via Portland. With Line 9 reversal and the purchase of two Panamax tankers by Valero, with the goal to move 50-60% (or Mbpd) from Montreal to Lévis (Van Praet 2014), the crude intake slate in the Valero refinery has changed dramatically. In 2016, it used 101 Mbpd of western Canadian crude, which includes synthetic, bitumen and heavy. The imported light oil came from four countries: Algeria (92.1 Mbpd), Kazakhstan (21.1 Mbpd), Nigeria (11.5 Mbpd), and US (Texas) (7.2 Mbpd). All foreign oil comes to Valero by tanker. Irving Oil s refinery in NB relies more heavily on imported oil and eastern offshore supply. Western Canada supplied 11.9 Mbpd of SCO via rail, 47.7 Mbpd of light crude came by tankers from Hibernia and other eastern offshore projects, and Mbpd came by tankers from predominantly four countries. Except for 15.3 Mbpd of heavy oil coming from Colombia and the Ivory Coast, the remaining imported oil was light. The largest supplier of imported feedstock is Saudi Arabia (86.7 Mbpd), followed by Nigeria (45 Mbpd), United States (34 Mbpd), Norway (24 Mbpd), Ivory Coast (12.6 Mbpd), Colombia (5.3 Mbpd), Azerbaijan (2.8 Mbpd) and Congo (2.7 Mbpd). Lastly, NL s North Atlantic Refinery was supplied almost entirely by foreign crude via tankers. Eastern offshore light oil accounted for 3.3 Mbpd out of a modelled 93.7 Mbpd intake for the refinery. Imported oil comes from the US (49 Mbpd), Nigeria (18.3 Mbpd), Norway (13.7 Mbpd), UK (6.2 Mbpd) and Denmark (1.7 Mbpd). Figure 3.4 shows the transportation routes of Canadian crude in the Base Case.

63 An Economic and Environmental Assessment of 45 Figure 3.4: Canadian Oil Supply Routes to Central and Eastern Refineries Base Case (Mbpd) Data Sources: (CAPP 2017b; Enbridge Inc. n.d.; NEB 2017c). Figure by CERI, based on cartography from (Enbridge Inc. n.d.). Note: L5 Line 5 pipeline; L78 Line 78 pipeline, L9 Line 9 pipeline, P-M - Portland Montréal Pipeline. Sarnia (3) illustrates the number of refineries located in the Sarnia region. There are two interesting key findings. First, in 2016 the reversed Line 9 was used at the level of 117 Mbpd out of its throughput capacity 285 Mbpd (assumed 95 percent of its total capacity). The lower than expected level may potentially be explained by the Fort McMurray wild fires in 2016 or other factors. Second, eastern offshore oil has limited use in central and eastern refineries due to some infrastructure restrictions. The largest user is Irving Oil with 47.7 Mbpd, followed by Suncor Energy in Montreal with 5.8 Mbpd, and NL & Labrador s refinery with 3.3 Mbpd. According to the National Energy Board (NEB 2017c), the rest of NL Offshore Light production in 2015 went to US PADD I (116 Mbpd), Asia and Europe (20 Mbpd) and the US Gulf Coast (4 Mbpd). The availability and usage of crude and transportation infrastructure by central and eastern refineries in the Base Case is illustrated in Figure 3.5. The top portion of the figure lists the various Canadian types of crude, ranging from AB Light and NL Offshore Light to AB Bitumen, in terms of availability and the amount used. The bottom portion of the figure, on the other hand, illustrates the various transportation modes potentially used in Canada, in terms of available infrastructure and their usage. In short, this figure provides a snap shot in time (2016) of the current crude flows from western and eastern Canada to the eastern Canadian refineries. Canada has enough light oil to be used in the eastern and western refinery markets.

64 Transportation Crude 46 Canadian Energy Research Institute Figure 3.5: Availability and Usage of Crude and Infrastructure by Eastern Canadian Refineries Base Case (Mbpd) AB Light SK Light AB SCO AB Heavy SK Heavy AB Bitumen NL Offshore Light New infrastructue PL Mainline (Line78, 5) PL Portland-Montreal PL Line 9 PL Valero's tankers Rail Available Infrastructure Available Used Data Source: CERI Crude Flows Model. Figure by CERI. Note: The volumes shown for Mainline (Line 78 and Line 5) represent Canadian crudes for Canadian refineries. The graph does not show movements in the Mainline of Canadian or US crudes destined for US refineries.

65 An Economic and Environmental Assessment of 47 It should be noted that the new pipeline infrastructure does not currently exist; it impacts the Made in Canada and Expanded Access scenarios, discussed in Chapter 4. The important input into the study s models is the availability of western Canadian crude supply for substitution. For this study, CERI considers oil produced in AB, SK, and MB, leaving BC production out of scope as it is assumed to be used in local refineries and not likely to be used in central and eastern refineries. Available Stock of Canadian Crude for Substitution To arrive at crude stocks available for central and eastern refineries, CERI uses the crude oil forecast from its Study 159, Canadian Crude Oil Production Forecast (CERI 2016). Based on average percentage of type of crude from each province, a forecast of available supply is derived from total crude oil production. This is illustrated in Figure 3.6. Note AB bitumen includes bitumen, not diluent. Figure 3.6: Canadian Crude Oil Production Forecast, (Mbpd) 6,000 5,500 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1, AB Light NL Offshore Light SK Light MB Light AB SCO SK Heavy AB Heavy AB Bitumen Data Source: (CERI 2016). Figure by CERI. The model assumes there are eight Canadian oil types modelled that represent the spectrum of the current crude intake (as of 2016) by refineries in eastern Canada. Four of them are based on existing crudes or blends AB Bitumen as a proxy for Western Canadian Select, AB Heavy as a proxy for Western Canadian Blend, AB SCO as a proxy for Synthetic Sweet Blend, and NL Offshore Light as a proxy for Hibernia. The other four Canadian crudes that include AB Light, SK Light, MB

66 48 Canadian Energy Research Institute Light and ON Light are assumed to be blends of either two or three existing crude oil blends selected from Mixed Sweet Blend, Mixed Sour Blend, Light Sour Blend or Midale in various ratios based on (COLC 2016) data. These crudes are highlighted in blue in Appendix C, please refer to Table C.1. The ratios are also detailed in the same table. To establish available stock for substitution purposes, the demand from western refineries is subtracted from the production forecast. The crude utilization in western Canadian refineries is shown in Table 3.3. Utilization rates are taken from the refineries annual reports or websites where available (the average of 2017/2016 or 2016/2015 was taken when information was available). For Chevron, Shell Canada, North West Refining, and Fed Co-op, 90 percent utilization was assumed due to lack of data. The demand for light, SCO, heavy and bitumen crude per refinery was determined based on (Statistics Canada 2017d) demand province-wide, and then adjusted using corporate websites and annual reports of refineries. Specifically, for AB and BC, the shares of each type of oil were calculated for each province for 2016 and are provided in Table 3.3. Table 3.3: Oil Type Shares as Input to Western Canadian Refineries Oil Type Oil Type Share in Refineries Intake (%) AB BC Conventional crude oil (light) 23.6% 32.5% Conventional crude oil (heavy) 6.7% 4.5% Synthetic crude oil (light) 59.9% 60.3% Crude bitumen charged 9.8% 0% Data Sources: (Statistics Canada 2017d), CERI calculations. Table by CERI. These ratios for the above table were applied to both BC refineries and in AB, unless more detailed information was available. For instance, in AB, the Suncor Refinery and North West Refining are assumed to receive bitumen. Oil demand by type for the Fed Co-op refinery in SK was available on the refinery website (Federated Co-operatives Limited 2015). The resulting Table 3.4 shows modelled crude oil demand by western Canadian refineries.

67 An Economic and Environmental Assessment of 49 Province Refinery Location Capacity (Mbpd) Table 3.4: Crude Oil Demand by Western Canadian Refineries Utilization Rate (%) Throughput (Mbpd) Light (Mbpd) SCO (Mbpd) Heavy (Mbpd) Bitumen (Mbpd) BC Husky Prince George BC Chevron Burnaby 57 90* AB Suncor Edmonton *** AB Imperial Strathcona AB Shell Scotford * Canada AB NW Sturgeon 50** 90* Refining Country SK Fed Co-op Regina * TOTAL Data Sources: (Statistics Canada 2017d), various refinery websites, CERI estimates. Figure by CERI. Notes: *Utilization assumed by CERI; **For NW refining, total bitumen capacity was used; ***Suncor Edmonton refinery was assumed to intake bitumen as its feedstock ((Suncor Energy Inc. 2017b) suggests that the refinery processes approximately 41 Mbpd of blended bitumen, 44 Mbpd of sour SCO, and 57 Mbpd of sweet SCO). As bitumen and SCO are not used for substitution of foreign crudes, this assumption does not impact the outcome of the study, but impacts the availability levels of SCO and bitumen provided below. The resulting available oil stock for central and eastern refineries after subtracting western refineries demand is depicted in Table 3.5. This is the total amount available for transporting to central and eastern refineries by type of crude and province. Table 3.5: Available Oil Stock for Central and Eastern Refineries, (Mbpd) Type of Crude and Province AB Heavy AB Bitumen 1,423 1,565 1,661 1,717 1,758 1,817 1,910 2,027 2,130 2,310 2,415 AB Light AB SCO ,008 1,047 1,058 1, SK Light SK Heavy MB Light NL Offshore Light Sub-total Light Sub-total Heavy Sub-total SCO ,008 1,047 1,058 1, Sub-total Bitumen 1,423 1,565 1,661 1,717 1,758 1,817 1,910 2,027 2,130 2,310 2,415 Total 3,614 3,823 3,952 4,018 4,078 4,113 4,168 4,222 4,294 4,452 4,589

68 50 Canadian Energy Research Institute Figure 3.7 illustrates the available crude oil for eastern Canadian refineries, as well as the demand by the refineries by crude type. The figure also shows that supply of light oil (conventional light + SCO) is enough to satisfy all the demand for light crude in the east. In fact, if all foreign oil was substituted in year 2027, there would still be 94 Mbpd of light oil left (852 Mbpd of available supply minus 758 Mbpd of refinery demand). Based on the availability of crude and demand by refineries, it is also clear that there is enough heavy, SCO and bitumen to supply central and eastern refineries needs. Figure 3.7: Available Western Canadian Crude for Central and Eastern Refineries and Demand by Central and Eastern Refineries by Crude Type (Mbpd) Light (includes Medium) Heavy 1, Available West and East East Refinery Demand Difference Available West East Refinery Demand Difference Synthetic Light Bitumen 1,000 2, , , , Available West East Refinery Demand Difference Available West East Refinery Demand Difference Source: CERI Note: Western Canadian crude supply available for central and eastern refineries was calculated as western Canadian crude production minus demand by western refineries.

69 An Economic and Environmental Assessment of 51 Modelling Scenario s Flows In all the scenarios, imported oil per refinery was substituted with Canadian crude using the following guiding principles: 1. Foreign oil was always displaced with the same oil type (light oil was substituted with light, heavy with heavy). This approach ensured that a) refineries can process crude of the same quality without a need for retrofit and/or expansion of their technologies, and b) that refinery yields will remain the same. The above-mentioned are important assumptions to modelling as CERI aims to account and measure effects of substitution of imported oil and isolate the modelling from other effects (e.g., modelling different yields leading to different downstream sales compared to the Base Case 2016 situation). 2. Availability of certain crude provided a stacking order in the substitution. If the required Canadian oil was not available in the volume needed from a source, the next available source of the required oil was used. 3. Western Canadian crude was used more in substitution in ON and QC, while eastern offshore Canadian crude was used in substitution in NL and NB. However, eastern oil was also modelled as far as Montreal s refinery, and western oil as far as NB. 4. Transportation infrastructure capacities also served as constraints: Line 5 and Line 78 pipelines (988 Mbpd), Line 9 pipeline (285 Mbpd), Portland-Montreal pipeline (270.8 Mbpd), rail offloading capacities at refineries (up to existing limits 255 Mbpd), tanker s capacity from Montreal to Valero s refinery up to 159 Mbpd. For all the pipelines, throughput was set as 95 percent of their total capacities. 5. For all the scenarios, pipeline delivery would take precedence over rail in the modelling. If pipeline capacity was full, but the scenario required further substitution, rail was used to compensate for lack of pipeline availability. At the same time, in the existing infrastructure scenarios, rail shipment of 11.9 Mbpd to NB s Irving Oil was kept intact. In the Expanded Infrastructure scenarios, this shipment was switched to pipeline. 6. For the market-based scenarios Current Reality and Expanded Access more expensive foreign crude oil was substituted. 7. For the policy-driven International Social Concerns scenario, cost of feedstock was not considered as a factor. All oil from undemocratic states was substituted even if it was cheaper. At the same time, in this scenario, oil from democratic countries was kept fully intact even if there was cheaper Canadian crude. 8. For the policy-driven Made in Canada scenario, all foreign crude was substituted guided by point #3, irrespective if Canadian crude was more expensive than foreign. The following section delves into some details of modelling crude flows for the four scenarios.

70 52 Canadian Energy Research Institute For the Made in Canada scenario, the crude flows modelling approach was as follows. All foreign oil was substituted with western and eastern Canadian crude (details of substitution are discussed further in Chapter 3). SK Light was used for ON refineries and in limited volume in QC (Suncor Energy refinery) until the availability of SK Light was used. AB Light and AB Heavy were used in QC and NB to substitute foreign crude. NL Offshore Light oil was used to displace all imported crude in North Atlantic Refining, the rest of available oil was modelled to be used in NB s Irving Oil until the availability of NL Offshore Light was fully used. Valero s and Irving Oil s refineries used the new pipeline capacity available in the scenario, while Montreal s Suncor Energy refinery proceeded to use the existing Mainline and Line 9. No rail deliveries were modelled. For the Current Reality scenario, existing supply from western and eastern Canada to central and eastern refineries remains intact or unchanged. ON s foreign oil was not substituted as foreign oil was found to be cheaper. In QC, AB Light was used until it was more expensive than foreign oil; then SK Light was used until it ran out of availability; then NL Offshore Light was used until it ran out of availability (because this oil was also used in Atlantic Canada). In all cases, more expensive oil was substituted. In NB, AB Light was used until it was more expensive than foreign crude. After that, NL Offshore was used for displacement, until it also was more expensive than foreign crude. NL Offshore was used to displace more expensive imports. For the Expanded Access scenario, the difference with Current Reality became the availability of the expanded transportation infrastructure (it was used by Irving Oil and Valero). ON s US foreign oil remained unsubstituted. QC s refineries were modelled to use AB Light for more expensive foreign crude; Irving Oil s was modelled to use AB Light via the new pipeline until it could not compete with cheaper foreign oil; then NL Offshore Light was used until it could not compete either. NL Offshore was used to displace more expensive imports the same volume levels as in Current Reality. Lastly, in the International Social Concerns scenario, SK Light was used in QC until it ran out of availability; NL Offshore Light was modelled to be used in NB, NL, and Valero s refinery until it also ran out of availability; the rest was substituted with AB Light. This approach ensured that while pursuing a policy scenario, the cheapest oil was used for substitution purposes (as it is seen in section below, NL Offshore Light is cheaper in all provinces (QC, NB, and NL) than AB Light, but it is more expensive than SK Light in QC). In this scenario, additional rail movements were modelled as tanker s capacity to Valero was not sufficient. Costs of Feedstock Model To understand the effect on cost of oil substitution, CERI had to establish costs of feedstock for each refinery for Canadian and foreign feedstock, as well as for Canadian crudes which are not currently used in the east refineries but could be used for substitution. This section describes the methodology that was used to generate costs for each refinery. It is subsequently divided into two parts: crude prices and transportation costs; and crude oil costs at refinery gate.

71 An Economic and Environmental Assessment of 53 Crude Oil Prices and Transportation Costs There are two elements of feedstock costs for a refinery: a) the cost of crude and b) the cost of delivery. The first element is represented by the price of a particular brand at a geographical price-point. The second element of feedstock costs for a refiner is the cost of transportation by a pipeline, rail, tanker, or a mix of the various modes. Table 3.6 shows the crude oils and their prices used in the study. The table includes the oil type (light, heavy, SCO and bitumen), the crude brand, proxy crude (or the formula used to determine the price, and the weighted price ($CAD) in Recall, there are over 150 different types of crude and even more brands of crude (CFA 2013). Listed are the crude oils used in this study. The presented prices are shown at location where they are sold, not at refinery gate. These were identified as crudes that are imported into Canada. Crude oil brands for foreign oil were taken from Clipper Data (ClipperData LLC 2017). In some cases, however, when a crude brand is not explicitly specified, CERI used a proxy crude from the location. Proxy crude oils are illustrated in the third column, where necessary. In four cases, crude prices were available, and the differential was not easily established. In these cases, CERI used Statistics Canada trade data to get an average price for a crude in 2016 (Government of Canada 2017; Statistics Canada 2017a). This was necessary with Denmark DUC, Equatorial Guinea Crude, and the two crude oils from Ivory Coast (Baobab and Espoir). It is also interesting to note that 21 out of 24 imported brands of crude oil are more expensive than AB Light. The exceptions are US North Dakota, Louisiana Thunderhorse, and US Eagle Ford. Also illustrated is that 10 out of 24 are more expensive than NL Offshore Light.

72 54 Canadian Energy Research Institute Oil Type Table 3.6: Crude Oil Brands Used in the Study and Their Prices Crude Category/ Brand Proxy Crude or Formula Used to Determine Price Name Used in the Study Weighted Price in 2016 ($CAD) Light Canada Hibernia Brent Canada Hibernia 57.9 Light Canada NL Offshore Light Brent NL Offshore Light 57.9 Light Canada AB Light Edmonton Light AB Light 51.9 SCO Canada AB SCO Edmonton Light multiplied by factor of (average factor between AB SCO and Edmonton AB SCO 56.3 Light for Jun-Aug 2017) Heavy Canada AB Heavy Western Canadian Select (WCS) AB Heavy 39.0 Bitumen Canada AB Dilbit Western Canadian Select (WCS) AB Bitumen 39.0 Light Canada SK Light, MB Light Edmonton Light SK Light, MB Light 51.9 Light Canada ON Light Edmonton Light ON Light 51.9 Light Algeria Saharan Blend No proxy needed Algeria Saharan Blend 58.6 Light Azerbaijan Azeri Light No proxy needed Azerbaijan Azeri Light 60.7 Heavy Colombia Castilla Colombia Castilla Brent minus $9.5 Blend Blend 45.3 Light Congo Djeno Blend Brent minus $2.5 Congo Djeno Blend 54.6 Light Denmark DUC Average price paid in 2016 by Canadian refineries (Statistics Canada trade data) Denmark DUC 56.8 Equatorial Guinea New Zafiro Light Blend, Equatorial Guinea Equatorial Guinea Average price paid in 2016 by Crude New Zafiro Blend Canadian refineries (Statistics 62.6 Canada trade data) Heavy Ivory Coast Baobab Average price paid in 2016 by Ivory Coast Baobab 52.0 Light Ivory Coast Espoir Canadian refineries (Statistics Canada trade data) Ivory Coast Espoir 52.0 Light Kazakhstan CPC Blend Brent minus $0.15 Kazakhstan CPC Blend 57.7 Light Michigan Sweet Michigan Sweet No proxy needed Crude Oil Crude Oil 57.0 Light Nigeria AKPO Blend No proxy needed Nigeria AKPO Blend 58.4 Light Nigeria Bonga No proxy needed Nigeria Bonga 58.8 Light Nigeria Bonny Light No proxy needed Nigeria Bonny Light 58.3 Light Nigeria Brass River No proxy needed Nigeria Brass River 59.4

73 An Economic and Environmental Assessment of 55 Oil Type Crude Category/ Brand Proxy Crude or Formula Used to Determine Price Name Used in the Study Weighted Price in 2016 ($CAD) Light Nigeria Qua Iboe Brent plus $0.5 Nigeria Qua Iboe 58.8 Light Nigeria Usan No proxy needed Nigeria Usan 55.5 Light Norway Ekofisk No proxy needed Norway Ekofisk 58.3 Light Norwegian Crude Origin: Norway Oseberg; Price: Norway Ekofisk Norwegian Crude 58.3 Light Saudi Crude Source: Saudi Ghawar; Price: Saudi Arabia Light Saudi Arabia Light 54.2 Light UK Brent No proxy needed UK Brent 57.9 Light UK North Sea UK Brent UK North Sea 57.9 Light US Louisiana LLS No proxy needed US Louisiana LLS 53.0 Light US Louisiana Source: Louisiana Thunderhorse, US Louisiana Thunderhorse Price: West Texas Sour Thunderhorse 47.8 Light US North Dakota No proxy needed US North Dakota 48.7 Light US Texas Eagle Ford Source: Texas Eagle Ford; Price: Gulf Coast Sweet US Texas Eagle Ford 46.2 Light US WTI No proxy needed US WTI 57.3 Data Sources: (Government of Canada 2017; IEA 2016; NRCan 2017a; Oil & Gas Journal 2016c; OPEC 2017; Statistics Canada 2017a), African countries websites. Table by CERI.

74 Crude Price, CAD Volume, mbpd 56 Canadian Energy Research Institute Figure 3.8 illustrates selected volume of imported crudes per refinery and prices in comparison with AB Light and NL Offshore Light. Volumes of more than 10 Mbpd are shown. Figure 3.8: Selected Foreign Crude Oil Brands Import Volumes and Prices US North Dakota Ontario refineries US North Dakota US WTI US Michigan Algeria Kazakhstan US WTI US Lousiana LLS Saudi Arabia Nigeria Bonga Nigeria Usan Norway US WTI US Lousiana LLS Montreal-Suncor Quebec-Valero NB-Irving Oil NL-North Atlantic Nigeria Bonny Light Volume Foregin Crude Price, CAD AB/SK Light Price NL Offshore Light Price Data Sources: (Government of Canada 2017; IEA 2016; NRCan 2017a; Oil & Gas Journal 2016c; OPEC 2017; Statistics Canada 2017a), African countries websites. Figure by CERI. At least three foreign crudes deserve attention based on their import volumes. First, the US North Dakota price beats the prices for both AB Light and NL Offshore Light. Second, Algerian oil is more expensive than both western and eastern Canadian crudes and thus has a good potential to be displaced cost-effectively. Third, Saudi Arabian oil is less expensive than NL Offshore Light but more expensive than AB Light; difference in cost for transportation will play a role in determining if substitution is cost-effective by both Canadian crudes. The following section describes the methodology used to determine transportation costs. For certain pipelines Enbridge Mainline, Line 9, Portland-Montreal, New Pipeline, Enbridge pipelines in North Dakota existing and projected tolls were used. In several cases where tolls were not available (i.e., transporting crude oil from Cushing, Oklahoma to the Gulf Coast or transporting Michigan oil to Sarnia, ON), a formula was used based on the cost per kilometer. The formula was originally established based on tolls of certain Canadian low, medium, and longrange pipelines. In all cases where committed tolls were available from the tariffs, those were used for the study.

75 An Economic and Environmental Assessment of 57 Table 3.7 illustrates tolls used for calculations in the study. Pipeline Oil Type To (Destination) Enbridge Mainline New Pipeline Table 3.7: Pipeline Tolls for Crude Oil ($CAD/barrel) From Clearbrook, MI Edmonton, AB Hardisty, AB Cromer, MB Heavy Sarnia, ON 6.38 Light/SCO Light 2.61* Sarnia, ON Light Nanticoke, 3.11* Light/SCO ON Heavy 6.99 Light/SCO Montreal, ** Heavy QC ** N/A Montreal, 6.71 (20 QC years) ** N/A QC, QC 7.02 (20 years) N/A Saint John, (20 NB (Combo years) toll, 20 years) Light Montreal, QC Light Clearbrook, MI Portland, ME North Dakota Cushing, OK Portland- Montreal ND 2.83 Pipeline Co. Seaway Light US Gulf, TX 4.46* Data Sources: (Enbridge Pipelines Inc. 2017a, 2017b, 2017c; FERC 2017; Montreal Pipe Line Limited 2015; TransCanada 2016), CERI estimates. Table by CERI. Notes: *CERI estimates based on tariffs for other pipelines of similar lengths; ** a toll from Hardisty to Montreal via new Pipeline is illustrated for comparison; it is not used in the study as Montreal keeps using Line 9 in all scenarios for western oil intake. Tanker costs were established using both Clipper Data and Oil & Gas Journal data (ClipperData LLC 2017; Oil & Gas Journal 2016a). Clipper Data oil shipments provided information on shipments volume, type of tanker and days in transition from a country to a particular port (Montreal, Portland, QC, etc.). The following tanker charter daily rates were used: Panamax (23 thousand US dollars per day), Aframax (40 thousand US dollars per day), Suezmax (48 thousand US dollars per day), VLCC (60 thousand US dollars per day) (Oil & Gas Journal 2016a). Daily rates are determined by the size of the tanker. Using this information, a cost per barrel per shipment was established, and then averaged for the country. A spill tax, tug fee and other fees of $0.36 USD was added for shipments. Other specific destinations needed for the study, for instance, Montreal to Saint John or Whiffen Head Terminal, NL & Labrador to QC, were estimated based on obtained costs for international and domestic tanker costs per barrel. Detailed transportation costs are provided in Appendix D. It is important to note that different costs for the same country and destination port are provided for different crude brands originating in that country. The difference in shipment costs is explained by different ports of origin within a state/country and/or different type of tanker.

76 58 Canadian Energy Research Institute To put tanker costs in comparison with pipeline costs for western crude, the lowest tanker cost is $0.91/bbl from Algeria to the North Atlantic refinery in NL & Labrador while the highest tanker cost is $2.12/bbl from Colombia to Irving Oil. Half of the shipments are under $1.45/bbl. At the same time, to move AB Light to Montreal, QC, and Saint John (the latter two by the New Pipeline) would cost $7.44/bbl, $7.02/bbl, and $8.60/bbl, respectively. NL Offshore Light oil can reach the same refineries by tanker for $0.80/bbl, $0.60/bbl and $0.70/bbl. For more details on tanker transport costs, refer to Appendix D. Rail transportation is used once in the modelled Base Case, transporting 12 Mbpd from AB to NB. This mode of transport is not projected to be used more intensively in scenarios with expanded pipeline infrastructure than in the Base Case, but may be needed to fulfill targets for existing infrastructure scenarios. Rail costs for light oil/sco ($/bbl) were calculated using the following metric $/barrel per km. The metric is estimated for each route using the formula derived from (Olateju and Kumar 2016): y = x + 0.2, where y is cost ($/bbl), and x is distance, kilometers 1 The most likely rail routes, which might be employed for crude flows modelling, are AB/SK to refineries in QC and NB. Heavy oil is not expected to be additionally moved by rail in the modelling. If a small amount of foreign heavy oil is to be displaced according to a scenario rationale, it will be transported by pipelines from the western provinces. Crude Oil Costs at Refinery Gate The combination of crude prices and transportation costs make up the total cost of crude feedstock at the refinery gate. To obtain cost of feedstock per refinery per route (from the specific origin by the specific mode), all the above-mentioned data was combined. Coupled with changes in crude flows per scenario, these costs of feedstock at a refinery gate will drive the cost results of oil substitution for each refinery. Table 3.8 illustrates the cost of feedstock per refinery per route. Along the top of the table there is the list of the eight refineries in eastern Canada, as well as the mode of transportation used to deliver the crude oil to each refinery gate. The grey cells denote rail shipments. Columns on the left of the table illustrate the name of the producing country as well as the crude brands that are used in this study. 1 The formula was estimated by CERI from a graph Comparative Transportation Cost: Rail vs Pipeline Impact on Distance in (Olateju and Kumar 2016), where costs were a linear function of distance.

77 An Economic and Environmental Assessment of 59 From a Canadian perspective, it is interesting to note that prices of crude oils increase at the refinery gate the further east that western Canadian oil is transported. This is particularly true when rail and tankers are used to transport western crude, adding other layers of cost. The comparison of prices of selected foreign and Canadian crudes, at place of initial sale and at the refinery gate (cost of feedstock with transportation costs), are provided in Appendix H.

78 An Economic and Environmental Assessment of 60 Country of Crude Oil Origin Canada US Crude Oil Brand Imperial Oil Shell Canada Table 3.8: Price of Crude Oil Brands at Refinery Gate of Eastern Canadian Refineries ON QC Atlantic Canada Suncor Energy Mainline Mainline Mainline Imperial Oil Mainline/ Rail Mainline/ Line 9 New Pipeline Suncor Energy Valero Irving Oil North Atlantic Refining Tanker Tanker/Portland Pipeline Mainline/ Line 9/Tanker New Pipeline Tanker/ Rail Mainline/ Line 9/Tanker New Pipeline Tanker/ Rail Mainline/ Line 9/Tanker AB Light AB SCO AB Heavy AB Bitumen SK/MB Light SK Heavy ON Light 55.2 New Pipeline/Tanker NL Offshore Light North Dakota WTI Eagle Ford Louisiana, LLS Louisiana, Thunderhorse 50.0 Michigan 62.0 Algeria Saharan Blend Azerbaijan Azeri Light Colombia Castilla Blend 48.1 Congo Djeno Blend 56.5 Denmark DUC 58.2 Equatorial Guinea New Zafiro Blend 64.9 Ivory Coast Espoir 54.8 Ivory Coast Baobab 54.4 Norway Ekofisk 60.7 Norwegian Crude 60.0 UK Brent Kazakhstan CPC Blend 59.8 Nigeria AKPO 60.9 Bonny Light 60.1 Qua Iboe 60.3 Brass River 61.0 Bonga 61.5 Usan 57.6 Saudi Arabia Saudi Arabia Light 56.5 Data Sources: CERI estimates based on: (ClipperData LLC 2017; Oil & Gas Journal 2016a); (Government of Canada 2017; IEA 2016; NRCan 2017a; Oil & Gas Journal 2016c; OPEC 2017; Statistics Canada 2017a); African countries websites; (Enbridge Pipelines Inc. 2017a, 2017b, 2017c; FERC 2017; Montreal Pipe Line Limited 2015; TransCanada 2016). Figure by CERI. Note: Grey cells denote rail shipments Tanker

79 An Economic and Environmental Assessment of 61 Emissions/LCA Model The objective of the emissions modelling is to establish an emissions profile for each crude oil in terms of CO2 equivalent per refinery. The specific methodology to obtain upstream, transportation and refinery emissions is stipulated below in greater detail. Life cycle assessment of GHG emissions of individual crude oils used in the study was used to inform environmental costs for each of the refineries. Analysis of available peer-reviewed literature, reports and documents from government, research institutions and international organizations shows that a number of approaches to model lifecycle GHG emissions of crude oils currently exists (Argonne National Laboratory 2017; J. Bergerson and MacLean n.d., 2016; Cai et al. 2015; El-Houjeiri and Brandt 2017; IHS Energy 2014; Jacobs Consultancy 2009, 2012; Lattanzio 2014; Moorhouse, Droitsch, and Woynillowicz 2011; Nimana, Canter, and Kumar 2015; Nimana et al. 2017; Pacheco et al. 2016; Tarnoczi 2013; TIAX LLC and MathPro Inc. 2009; Vafi and Brandt 2014). After detailed review of the above documents, the Oil Production Greenhouse Gas Emissions Estimator (OPGEE) developed at Stanford University was selected as a tool to assess GHG emissions from exploration, production, surface processing, and transport of crude oil to the refinery inlet (El-Houjeiri and Brandt 2017), and the Petroleum Refinery Life-Cycle Inventory Model (PRELIM) developed at the University of Calgary was selected to estimate GHG emissions related to crude oil processing in different types of refineries that combine different processing units (J. A. Bergerson et al. 2016). The approach used by CERI was a partial life cycle assessment, since downstream GHG emissions after the refinery outlet (resulting from the transport and end use of refined petroleum products) were out of the scope of this research. The reasons behind selecting these specific two models for GHG emissions calculations for the purposes of our study included the following: Both OPGEE and PRELIM models have been run on a wide range of individual crude oils with extensive geography (approximately 300 global crudes for OPGEE model, and more than 110 global crudes for PRELIM model) (J. A. Bergerson et al. 2016; CARB 2015b; Vafi and Brandt 2013; Wang et al. 2016). That represents more upstream and midstream crude runs than other modelling efforts, in particular, those presented in (IHS Energy 2014; Jacobs Consultancy 2009, 2012; TIAX LLC and MathPro Inc. 2009); OPGEE and PRELIM are built as engineering-based models that use public datasets where possible, open academic sources, technical reports, and industry references, with documenting sources for all equations, parameters, and assumptions. This approach allows more flexible and accurate estimations of upstream and midstream GHG emissions from various emissions sources in comparison with some other publicly available models. It also provides more transparency and accessibility in methods and data compared to private models that use proprietary datasets (J. Bergerson and

80 62 Canadian Energy Research Institute MacLean 2016; J. A. Bergerson et al. 2016; El-Houjeiri et al. 2017; Vafi and Brandt 2013, 2014); It was important to select tools where input data were available to model both upstream and midstream GHG emissions for many individual crude oils (Canadian and international) used in the study, with outputs that could be easily converted to the same functional units (kg of CO2 equivalent per barrel of crude oil). OPGEE model requires up to 50 data inputs, however, it can function sufficiently with limited key parameters entered and built-in defaults that can fill in data gaps. According to the model, the most important characteristics depend on extraction method specifications and activity per unit of production, and include steam-to-oil and water-to-oil ratios, flaring and venting rates, and crude density (measured as API gravity). Other important inputs also include gas-to-oil ratios, oil production rates, location (onshore, offshore) and depth of the oil field (El-Houjeiri et al. 2017; El-Houjeiri and Brandt 2017). The model uses the following classification of crude oils by their API gravity: extra-heavy ( 12 API), heavy (12 20 API), medium (20 30 API), light (30 40 API), ultralight ( 40 API) and condensate (El-Houjeiri et al. 2017). By sulfur content, oil is classified as sweet (<0.5% sulfur) or sour (>0.5% sulfur). It should be noted that for synthetic crude oils, the sulfur content is reported after the bitumen is upgraded. OPGEE generates final upstream GHG emissions outputs in kg CO2 eq/megajoule (MJ) which can be converted into kg CO2 eq/barrel (bbl) of crude oil by multiplying them by the lower heating value of each oil (in MJ/bbl; this factor depends on API gravity of oil) (El-Houjeiri and Brandt 2017; El-Houjeiri et al. 2017). The most important upstream GHG emission drivers identified by OPGEE include flaring, land use (that impacts the land s carbon storage capacity), steam, venting and fugitive emissions, pumping, and upgrading (Brandt et al. 2015; El-Houjeiri and Brandt 2017; El-Houjeiri et al. 2017). OPGEE also includes GHG emissions from the transportation of oil to the refinery inlet that depend on the mode of transportation (can include ocean tankers, pipeline, railway, or tanker trucks), transport fuel used, the distance traveled and mass of crude oil. While publicly available input datasets for OPGEE obtained from California ARB and Stanford University assume that all oil is transported by default to the refinery hub in Houston, TX, or Los Angeles, CA, as default destinations, CERI used the actual average distances traveled by crude oils from their countries/places of origin to a refinery in eastern Canada. Emissions of imported oil, including upstream activities and transportation of the oil, were accounted for, even if they are produced outside of Canadian territory. Volume of crude oil transported was converted to mass (from barrels to tonnes) using reported product densities, and the vehicle and fuel emissions factors as suggested in the OPGEE model (with references made in this model to other relevant sources) were applied (El-Houjeiri and Brandt 2017; Argonne National Laboratory 2017; US EPA 2015). Based on the above sources, transporting GHG emissions (kg of CO2eq) per tonne of fuel shipped for 1 km (kg CO2eq/tonne-km) are as follows:

81 An Economic and Environmental Assessment of 63 Tanker trucks 0.09 Railway 0.02 Pipeline 0.01 Barge For marine tankers (ocean crude carriers), transportation emissions factors also depend on an ocean tanker size (tonnes). GHG emissions for various ocean tanker sizes referred to in the study were provided in the OPGEE model in g/mmbtu-mi fuel transported (El-Houjeiri and Brandt 2017). They were further converted to kg CO2eq/tonne-km and are as follows: Ocean tanker (250,000 tonnes) Ocean tanker (120,000 tonnes) Ocean tanker (80,000 tonnes) Ocean tanker (60,000 tonnes) The PRELIM model for refinery GHG emissions calculations requires comprehensive crude oil assays as inputs; each parameter at the assay must be specified at the specific temperature cut, with consistent temperature cut ranges for each assay. According to the model, the most important parameters for oil assays include API gravity, sulfur, hydrogen and nitrogen content, density, volume/mass flow, micro-carbon residue or Conradson carbon residue, and viscosity for vacuum residuum (Abella et al. 2016a; J. A. Bergerson et al. 2016; J. Bergerson and MacLean 2016). Ideally, they should be reported for nine temperature cuts with corresponding products; however, the above parameters for at least four temperature cuts could also be entered into the model, but in this case, it will create uncertainty and could possibly affect GHG emissions outputs (Abella et al. 2016a). For the purposes of midstream GHG emissions estimates, in addition to the crude oil assays presented in the PRELIM assay inventory, CERI used other open-source crude assays consistent in a format with the PRELIM requirements, as discussed below. PRELIM uses ten different combinations of processing units within three different types of refinery (hydroskimming, medium conversion, or deep conversion). A default refinery configuration is assigned in the model as the most suitable for each crude oil, based on the oil assay parameters and using API gravity and sulfur content as the criteria. The default configurations are as follows (Abella et al. 2016a; J. Bergerson and MacLean 2016): hydroskimming for light sweet crudes; medium conversion (with fluid catalytic cracking and gas-oil hydrocracking units [FCC & GO-HC]) for medium sweet and sour crudes, and light sour crudes; deep conversion (with fluid catalytic cracking and gas-oil hydrocracking units [FCC & GO- HC]) for heavy sweet and sour crudes. Deep conversion can further include either a coking or hydrocracking refinery configuration.

82 64 Canadian Energy Research Institute However, the user can override these default selections if necessary. It is worth noting that for the upgraded extra-heavy oils, the model uses API gravity for their synthetic crude oil (instead of their initial gravity) to determine the default refinery configuration; for the bitumen that was simply diluted without being upgraded, the API gravity of the oil sand is used to determine the default refinery configuration (Abella et al. 2016b; J. A. Bergerson et al. 2016). The most important midstream GHG emission drivers identified by PRELIM include refinery heat and refinery hydrogen (which is produced for adding to carbon-heavy oils) (Abella et al. 2016a). While both models have numerous advantages discussed above which make them particularly useful for the current study, it is critical to remember their limitations. For both OPGEE and PRELIM, quality of input data is paramount. The accuracy of modelling results is fundamentally related to data inputs available; it is especially true for the OPGEE model. Key parameters that determine data quality are their consistency, completeness, representativeness, accuracy, and comparability. With proper assumptions and good quality detailed data, more accurate emissions estimates can be generated and the individual models can be tuned to match real world results. The main obstacles for obtaining quality input data for these models include lack of publicly available and transparent crude oil information, as well as data inconsistency and discrepancy. Input data for a crude oil are difficult to find and validate, and alternate interpretations may result in varying GHG estimates. OPGEE GHG calculations are the primary source of variation in GHG emissions estimates, whereas PRELIM provides little additional uncertainty (Abella et al. 2016a; El-Houjeiri et al. 2017; Vafi and Brandt 2013, 2014). The objective of the work on GHG emissions/lca modelling was to collect information on upstream, transportation and midstream emissions for all the crude oils (Canadian and international) that are being imported and that will be used for substitution in eight eastern Canadian refineries. One of the potential limitations of the study was lack of information on what type of crude oil or crude oil brand is imported from specific countries, as well as which particular brands of light, heavy, SCO and bitumen are supplied to refineries. Therefore, selection of crude oils for the purposes of GHG emissions modelling was based on: information on crude intake (country/province, crude type/category, etc.) for eight eastern Canadian refineries in 2016 obtained from several sources (Clipper Data 2017, COLC 2016, Crude Monitor 2017, Statistics Canada 2017, various refineries websites, CERI s assumptions); data availability for both upstream and midstream GHG modelling (operations specifications, publicly available assays, input data in the models); oil blend composition and characteristics (API gravity, sulfur content, etc.) that allowed to assume which individual crude brand or blend can represent each crude flow taken by the refineries;

83 An Economic and Environmental Assessment of 65 if the detailed information on an imported crude oil brand was absent, CERI used the most abundant crude oil brand from the import country per refinery when modelling baseline flows. More information on individual crude oils and blends selected for the purposes of crude flows and/or GHG emissions modelling, as well as assumptions on proxy crudes selection for GHG emissions modelling, is presented in Table C.1 (Appendix C). In Table C.1, there are 32 individual crude brands or blends selected for the purposes of GHG emissions modelling, including eight Canadian and 24 foreign crude oils (25 percent and 75 percent of total amount, respectively). The spectrum of crude types includes bitumen (one Canadian blend), three heavy oils (one of which is a Canadian blend), five medium crude oils (one of which is a Canadian blend), SCO (one Canadian blend), 18 light oils (four of which are Canadian blends) and four ultra-light oils (all foreign blends). For the foreign crude flows, 22 were actual crude oil brands or blends imported to Canada in 2016 (Statistics Canada, NRCan, CFA, Clipper Data), and for 20 of the 22 crudes, sufficient data for upstream and midstream GHG emissions modelling were available. In the absence of publicly available data for the remaining two crude flows, Denmark DUC and Kazakhstan CPC Blend, proxy crude oil blends (Denmark Dansk Blend and Kazakhstan Tengiz) were used for the purposes of GHG emissions modelling for these flows (see the Assumptions column). Data available for the other four crude flows (Norwegian Crude, Saudi Arabia Light, UK North Sea, and US Michigan Light) were not specific in terms of actual crude oil brands imported to Canada, so for each of them an individual proxy brand or blend (Norway Oseberg, Saudi Arabia Ghawar, UK Brent or US Bakken) was selected based on the criteria described above. Please note that UK Brent and US Bakken were used as proxies twice (see Table C.1 for details), this explains the difference between the number of foreign crude flows (26) and the number of individual foreign crude brands or blends (24) used for GHG emissions modelling. Table C.2, also in Appendix C, presents the results of upstream and midstream GHG emissions modelling for eight Canadian and 24 foreign crude oil brands and blends (referred to earlier in Table C.1 as individual brands and blends selected as proxies for crude flows modelling). Table C.2 also provides detailed information on models applied, sources of input data and assumptions behind the GHG emissions modelling. All upstream GHG emissions for the crude oils used in the study were calculated using the OPGEE v2.0a model which was the most current version of OPGEE as of September To run the model, CERI has collected input data for the 32 crude oils of interest from a variety of sources, mostly California ARB (CARB 2015a, 2015b) and Stanford University (Wang et al. 2016), but also from (Crude Quality Inc. 2017; Kable Intelligence Ltd. 2017; Wiki.dot 2015) and crude oil producing companies websites. All upstream GHG emissions calculated using the OPGEE model exclude transportation emissions. GHG emissions resulting from crude oil transport to the refinery inlet in Los Angeles, California, or Houston, Texas, as default destinations according to (CARB 2015b; Wang et al. 2016) were 2 The latest version of the model, OPGEE v2.0b, was released in July 2017, however, it was not available for download until November 2017.

84 66 Canadian Energy Research Institute deducted from the total upstream GHG emissions outputs for each modelled oil (see Notes to Table C.2). For several crude oils, a few sources of input data for upstream GHG emissions calculations were available, with comparable output results. However, for some crudes the resulting emissions based on input data from different sources varied substantially. For example, GHG emissions for Canada Synthetic Sweet Blend based on (CARB 2015b) data were estimated to be kg CO2eq/bbl crude. The same emissions based on (Wang et al. 2016) were calculated as a weighted average of upstream GHG emissions for Suncor A and Syncrude Sweet Premium that constitute Synthetic Sweet Blend (see the Assumptions column in Table C.2) and were estimated to be kg CO2eq/bbl crude. The result based on California ARB (2015) data was selected, since it aligns with upstream GHG emissions results for Canadian SCOs provided in other sources (Jacobs Consultancy 2009; CERI 2017). For Nigeria Brass River, upstream GHG emissions calculations based on input data provided by six producing companies (SPDC, Chevron, NAOC Phillips, Addax, AENR/AGIP, others) disclosed an almost ten-fold difference between the lowest and the highest emissions values (CARB 2015b). Upstream GHG emissions outputs based on Chevron data were selected for the purposes of this study to reflect the fact that Chevron has one of the highest oil production volumes and the biggest number of producing wells within this field (see also Notes to Table C.2). In the case of Nigeria Bonny Light, where input data from four producing companies (SPDC, Chevron, Total E&P, and others) resulted in a 2.5 times difference between the lowest and the highest upstream GHG emissions values obtained in OPGEE model (CARB 2015b), the outputs based on Chevron data were selected for the same reason. As it can be seen from Table C.2, all midstream GHG emissions for the crude oils used in the study were calculated using the PRELIM v1.1 model (J. A. Bergerson et al. 2016). For 21 crude oil brands or blends, input data based on numerous crude assays (CrudeMonitor, BP, ExxonMobil, Chevron, Statoil, Stratiev, Platt, etc.) were available from the assay inventory spreadsheet (PRELIM v1.1). For the other 11 crude oil blends, publicly available crude assays obtained by CERI from CrudeMonitor.ca, EcoPetrol, ExxonMobil, Maersk Oil, TOTAL and Tullow Oil websites had to be converted to the PRELIM format first and then run through the PRELIM v1.1 model. 3 4 It should be noted that different results for midstream GHG emissions could be obtained depending on different crude assay data available from various companies for some crude oils. A few examples include Azerbaijan Azeri Light where different PRELIM outputs were obtained based on Chevron, ExxonMobil, and Statoil assays; Norway Ekofisk (BP, Statoil, and Chevron assays), UK Brent (BP, Chevron, ExxonMobil); Canada Hibernia (ExxonMobil, Statoil, and Chevron), etc. To decide which crude assay (if there are multiple assays available) should be selected to run in the PRELIM model, CERI checked the availability of oil producing companies 3 The authors would like to personally thank John Guo, MSc., and Dr. Joule Bergerson at the University of Calgary (PRELIM project team) for their help with converting publicly available assays for the selected crude oils into the PRELIM format.

85 An Economic and Environmental Assessment of 67 input data for upstream GHG modelling. If a company s data were available for both upstream and midstream GHG modelling, this dataset (including crude oil assay) was used for the purposes of the study, for the sake of consistency. The results of GHG emissions modelling presented in Table C.2 demonstrate that upstream GHG emissions vary significantly among the individual crude brand/blends used for the study, ranging from to 7.7 kg CO2eq/bbl crude. Midstream GHG emissions for the same crude brand/blends vary less significantly (ranging from 92.7 to 12.5 kg CO2 eq/bbl crude), based on the assumption that heavy crudes undergo the deep conversion refining process with the hydrocracking unit. Table C.2 also provides numbers for the midstream GHG emissions from heavy crudes that undergo the deep conversion refining process with the coking unit. In this case, there will be approximately six times difference in midstream GHG emissions for the studied crude oils, ranging from 77.6 to 12.5 kg CO2eq/bbl crude. Upstream and midstream GHG emissions for selected Canadian and foreign crude oils used in the study are also shown on Figure 3.9. Crude oil brands/blends previously discussed in Table C.2, are split by crude categories heavy, medium, light (including SCO) and ultra-light crude oils. This provides visualization for the purposes of total, upstream and midstream GHG emissions comparison both between the categories, and between Canadian and foreign crude oils within each crude category.

86 68 Canadian Energy Research Institute Figure 3.9: Upstream and Midstream GHG Emissions for Selected Canadian and Foreign Crude Oils Used in the Study (by Crude Categories) Data Sources: (PRELIM v1.1 (J. A. Bergerson et al. 2016); OPGEE v2.0a (El-Houjeiri and Brandt 2017); CARB 2015b; Wang et al. 2016), CERI calculations. Figure by CERI. Notes: Stacked columns with borders represent Canadian crude oils, stacked columns without borders represent foreign crude oils. Upstream and midstream GHG emissions for Canada WCS dilbit are presented per barrel of diluted bitumen, not per barrel of crude oil produced. For midstream GHG emissions from heavy crude oils, the deep conversion refinery configuration with hydrocracking is assumed.

87 An Economic and Environmental Assessment of 69 Please note that midstream GHG emissions from heavy crude oils are shown on Figure 3.9 for the deep conversion refinery with the hydrocracking unit only. Total GHG emissions differ approximately 7.6 times between the highest numbers (196.3 kg CO2eq/bbl, Canadian WCS Dilbit) and the lowest ones (25.7 kg CO2eq/bbl, US Bakken No Flare) for the studied crudes. If comparing total GHG emissions within each crude category for the studied crudes as presented on Figure 2.10, there was approximately 1.6 times difference for four heavy crudes (from to kg CO2eq/bbl), approximately 3.3 times difference for five medium crudes (from to 45.3 kg CO2eq/bbl), approximately 5.6 times difference for 19 light crudes including one SCO (from to 25.7 kg CO2eq/bbl) and approximately 2.4 times difference for four ultra-light crudes (from 94.4 to 39.9 kg CO2eq/bbl). While the highest total GHG emissions were observed for Canadian WCS Dilbit, it should be noted that the Canadian oil sands industry is working diligently to reduce its bitumen production footprint, in terms of reducing natural gas use, fresh water use, steam management, well pad design, and so on. The technologies that are being tested, commercialized, and implemented are at various stages of adoption by industry players. The viability of some of these technologies on how they can help to reduce the environmental footprint was discussed in CERI Study 164, Economic Potential and Efficiencies of Oil Sands Operations: Processes and Technologies (CERI 2017). Figure 3.10 provides a graphic summary of the results of emissions modelling presented earlier in Table C.2 and on Figure 3.9, showing ranges of total GHG emissions separately for Canadian and foreign crude oils split by crude categories.

88 70 Canadian Energy Research Institute Figure 3.10: GHG Emissions Ranges for Crude Oils Used in the Study (by Crude Categories) Data Sources: (PRELIM v1.1 (J. A. Bergerson et al. 2016); OPGEE v2.0a (El-Houjeiri and Brandt 2017); CARB 2015b; Wang et al. 2016), CERI calculations. Figure by CERI. Notes: Average GHG emissions for crude oils in each category are represented by X. GHG emissions for Canadian dilbit are presented per barrel of diluted bitumen, not per barrel of crude oil produced. As it can be seen from Figure 3.10, Canadian Dilbit (i.e., Canada WCS Dilbit) shows the highest total GHG emissions, however, they are presented per barrel of diluted bitumen, not per barrel of crude oil produced. 4 5 Canadian Heavy (i.e., Canada WCB Heavy) is higher in total GHG emissions (150.9 kg CO2eq/bbl) than Foreign Heavy (average kg CO2eq/bbl), but Canadian Medium (i.e., Midale) has lower GHG emissions (83.6 kg CO2eq/bbl) than average total GHG emissions (100.3 kg CO2eq/bbl) for Foreign Medium crude. This is also the case for Canadian Light: the average total GHG emissions (58.8 kg CO2eq/bbl) for this crude category are lower than average total emissions for Foreign Light (68.7 kg CO2eq/bbl) and Foreign Ultra-Light (67.8 kg CO2eq/bbl) crude categories. In contrast to conventional Canadian light oils, Canadian SCO (i.e., Canada Synthetic Sweet Blend) shows high total GHG emissions (144.1 kg CO2eq/bbl). However, it is still comparable in emissions with some conventional, not upgraded light foreign oils, for example, US Louisiana Light Sweet (143.2 kg CO2eq/bbl). 4 For Canadian WCS it is assumed that a barrel of oil produced from bitumen is 70% bitumen blended with 30% diluent (CARB 2015b). Upstream GHG emissions from producing the diluent for dilbit are also considered in OPGEE (El-Houjeiri et al. 2015).

89 An Economic and Environmental Assessment of 71 Figure 3.11 graphically represents information provided in Tables C.1 and C.2 for upstream and midstream GHG emissions from individual Canadian crude oil blends used as proxies, as well as the resulting GHG emissions for Canadian crude flows modelled by CERI based on these proxy blends. Figure 3.11: Upstream and Midstream GHG Emissions for Canadian Proxy and Modelled Crude Flows Data Sources: (PRELIM v1.1 (J. A. Bergerson et al. 2016); OPGEE v2.0a (El-Houjeiri and Brandt 2017); CARB 2015b; Wang et al. 2016), CERI calculations. Figure by CERI. Stacked columns representing GHG emissions in Figure 3.11 have either solid fill (for individual proxy crudes), or diagonal stripes pattern fill (for crudes modelled by CERI). Columns representing each modelled crude flow along with columns representing individual crude brand(s) or blend(s) used to create the particular crude flow are surrounded with rectangles to visually separate flows from each other. As it can be seen from Figure 3.11, if the modelled crude flow is deemed to consist of one proxy crude brand/blend, GHG emissions for the flow are understandingly the same as the emissions for its proxy crude (e.g., Western Canadian Blend and AB Heavy, or Hibernia and NL Light Offshore, etc.). However, if the modelled crude flow consists of two or three individual proxy blends, the resulting upstream and midstream GHG emissions for the flow were calculated as a weighted average of OPGEE and PRELIM outputs for individual proxy blends (see, for example, AB Light, or SK Light, etc.). Recall, since AB Light and ON Light (as well as SK Light and MB Light) are assumed, for the purposes of modelling, to be blends of the same proxy crudes with the same proportions (see Table C.1 for assumptions), their resulting GHG emissions per barrel of modelled crude oil will be the same. Again, upstream, and

90 72 Canadian Energy Research Institute midstream GHG emissions for Western Canadian Select and AB Bitumen are presented per barrel of diluted bitumen, not per barrel of crude oil produced. Figures 3.12 and 3.13 provide detailed information on OPGEE GHG emissions results for Canadian and foreign crude oils used for the study, either showing the total for upstream emissions (Figure 3.12) or splitting them by the main upstream emissions drivers (Figure 3.13). Both figures are at the same scale, and are organized in the same order, with crude oils ranked by the total upstream emissions (highest to lowest) for the convenience of comparing. As it can be seen, Figure 3.13 looks different from Figure 3.12, even if total upstream emissions for each individual crude depicted on both figures are the same. This is related to the fact that the OPGEE model considers GHG emissions from a lifecycle perspective, therefore, upstream GHG emissions associated with electricity generated on-site or natural gas produced that is gathered, sold, and not flared (i.e., off-site emissions) are credited back or deducted from total upstream GHG emissions (Brandt et al. 2015; Argonne National Laboratory 2017; El-Houjeiri et al. 2017; El- Houjeiri and Brandt 2017). For some crudes (e.g., UK Brent, Nigeria Usan or Nigeria Akpo Blend), the off-site emissions credited towards total GHG emissions compensate their otherwise high upstream emissions. For many crude oils presented on Figure 3.13, the main driver for upstream emissions is venting, flaring or fugitive emissions (VFF). For some crude oils with a high gas-to-oil ratio (especially light and ultra-light oils), gas can be vented during processing even if the flaring rate is low, whereas for other crude oils their relatively high GHG emissions can be attributed to flaring of associated gas if it is not very well managed (El-Houjeiri and Brandt 2017; El-Houjeiri et al. 2017). However, some crude oils with a high gas-to-oil ratio where their associated natural gas is responsibly managed and can be gathered and exported (e.g., US Bakken (No Flare), US Eagle Ford, US WTI or Norway Ekofisk,) generate negative off-site GHG emissions that can lower total upstream emissions (El-Houjeiri and Brandt 2017; Wang et al. 2016). For some crude oils, the main emission driver will be production and extraction, specifically for conventional oils from depleted fields (e.g., US Louisiana Light Sweet) as they use energy-intensive pumping techniques (Masnadi and Brandt 2017), or for bitumen (e.g., Canada WCS Dilbit), which requires substantial amounts of energy to be produced. For SCOs (e.g., Canada Synthetic Sweet Blend), upgrading operations are the main driver behind upstream GHG emissions.

91 An Economic and Environmental Assessment of 73 Figure 3.12: Total Upstream GHG Emissions for Canadian and Foreign Crude Oils Used for the Study Figure 3.13: Upstream GHG Emissions for Canadian and Foreign Crude Oils Used for the Study (Split by Main Emissions Drivers) Data Sources: (CARB 2015b; OPGEE v2.0a (El-Houjeiri and Brandt 2017); Wang et al. 2016), CERI calculations. Figure by CERI. Notes: Stacked columns with borders represent Canadian crude oils, stacked columns without borders represent foreign crude oils. Upstream GHG emissions for Canada WCS Dilbit are presented per barrel of diluted bitumen, not per barrel of produced crude oil. Off-site emissions can be either added to (positive) or deducted from (negative) total upstream GHG emissions.

92 74 Canadian Energy Research Institute Figure 3.14 provides detailed information on PRELIM GHG emissions results for Canadian and foreign crude oils used for the study, with crude oils ranked by the total midstream emissions (highest to lowest). Please note that the studied crude oils on Figure 3.14 are ranked differently from those on Figure 3.12 (though in the order of decreasing emissions on both figures), since there is no direct correlation between upstream and midstream GHG emissions for these crudes. Figure 3.14: Midstream GHG Emissions Results for Canadian and Foreign Crude Oils Used for the Study Data Sources: (PRELIM v1.1 (J. A. Bergerson et al. 2016); Abella et al. 2016b), CERI calculations. Figure by CERI. Notes: Stacked columns with borders represent Canadian crude oils, stacked columns without borders represent foreign crude oils. Midstream GHG emissions for Canada WCS Dilbit are presented per barrel of diluted bitumen, not per barrel of produced crude oil. Medium and deep conversion refineries use fluid catalytic cracking (FCC) and gas-oil hydrocracking (GO-HC). It can be seen from the figure that the main difference in midstream GHG emissions from the selected crudes is related to the default refinery configuration (deep conversion, medium conversion or hydroskimming). The studied crude oils undergoing the refining process in a hydroskimming refinery have the lowest GHG emissions ranging from 26.4 to 12.5 kg CO2eq/bbl crude. Crude oils depicted on Figure 3.14 that undergo the refining process in a medium conversion refinery have their emissions in a range from 47.1 to 21.0 kg CO2eq/bbl crude. Midstream GHG emissions from heavy crude oils on Figure 3.14 are assumed to be from the refining process in a deep conversion refinery with the hydrocracking unit, and they range from 92.7 to 78.5 kg CO2eq/bbl crude. Canadian crudes are present on both ends of the emissions spectrum, both for upstream GHG emissions (Figure 3.12) and midstream GHG emissions (Figure 3.14). As discussed above, all upstream GHG emissions calculated using the OPGEE model exclude transportation GHG emissions. They were calculated separately in accordance with the

93 An Economic and Environmental Assessment of 75 methodology described at the beginning of this section. The results for transportation GHG emissions for Canadian and foreign crude oils used in the Base Case are graphically presented on Figures 3.15 and As it can be seen from Figure 3.15, the main transportation mode for all crudes shipped to ON refineries in the Base Case is by pipeline (gathering and main pipelines), except for ON Light that is shipped to Imperial refinery in Nanticoke by gathering pipeline and rail. Transportation GHG emissions in this case range from 5.59 kg CO2eq/bbl crude (AB Heavy to Imperial refinery in Nanticoke) to 0.26 kg CO2eq/bbl crude (ON Light, Imperial refinery in Nanticoke). All refineries shown on Figure 3.16 use transportation by tanker (either combined with transportation by pipeline, or without pipeline, if tanker was offloaded offshore). Despite the large distances associated with marine transportation, GHG emissions for this mode are low related to the lower transportation emissions factors for tankers to kg CO2eq/tonne-km depending on tanker size (tonnes). Recall, the factor for pipeline transportation is 0.01 kg CO2eq/tonne-km, and for railway it is 0.02 kg CO2eq/tonne-km. This means an ocean tanker moving a tonne of crude oil one kilometer will emit 3.3 times less GHG emissions than a pipeline moving a tonne of crude oil one kilometer, or 6.6 times less than moving a tonne of crude the same distance by rail. Transportation GHG emissions in the case of a tanker range from 6.01 kg CO2eq/bbl crude (Saudi Arabia Light to Irving Oil refinery, NB) to 0.32 kg CO2eq/bbl crude (NL Light Offshore to North Atlantic Refining, NL). The highest transportation emissions (13.20 kg CO2eq/bbl crude) for all studied crude oils were obtained for AB SCO shipped by rail to the NB refinery. Table 3.9 provides a detailed summary of transportation GHG emissions for Canadian and foreign crude oils used in the Base Case and the four modelled scenarios introduced in Chapter 2, split by eight refineries in four provinces (ON, QC, NB, and NL & Labrador). Unlike upstream and midstream GHG emissions, transportation emissions in Table 3.9 (as well as on Figures 3.15 and 3.16) are attributed to the modelled crude flows (see Table C.1, Appendix C) based on proxy crude oil brands/blends. Transportation emissions are low; for the crude oils used in the study, they do not exceed 30 percent of total upstream GHG emissions (except for Azerbaijan Azeri Light, US Bakken, Kazakhstan CPC Blend, and Saudi Arabia Light), and are lower for many of the selected crudes. The lowest ratio for transportation emissions was 0.6 percent of upstream GHG emissions for ON Light transported to the Imperial refinery in Nanticoke due to a short transportation distance (100 km by railway). For Azerbaijan Azeri Light, this ratio is almost 52 percent and can be explained with very low upstream GHG emissions for the crude oil and the distance (approximately 1,770 km by pipeline and 9,600 km by tanker) Azerbaijan Azeri Light travels to the Suncor Energy refinery in QC. The same (low upstream GHG emissions combined with long distances the crude travels by pipeline and tanker) is true for Kazakhstan CPC Blend whose transportation to upstream GHG emissions ratio is 34.1 percent (shipping to the Valero refinery in QC) and Saudi Arabia Light (31.6 percent ratio, shipping to the Irving Oil refinery in NB). The high transportation to upstream GHG emissions ratio for US Bakken is also related to its very low upstream GHG emissions (in terms of kg CO2eq/bbl, the lowest ones among all crude

94 76 Canadian Energy Research Institute oils used for the study) and long traveling distances (approximately 2,800 km by pipeline to Suncor Energy refinery in QC).

95 An Economic and Environmental Assessment of 77 Figure 3.15: Transportation GHG Emissions (kg CO2 eq/bbl crude) for Canadian and Foreign Crude Oils Used in the Base Case Refineries Located in ON Source: CERI calculations Notes: Stacked columns with borders represent Canadian crude oils, stacked columns without borders represent foreign crude oils. Transportation GHG emissions for AB Bitumen are presented per barrel of diluted bitumen, not per barrel of produced crude oil. Transportation by pipeline includes gathering pipeline and/or main pipeline. ON Light is transported by gathering pipeline and railway.

96 78 Canadian Energy Research Institute Figure 3.16: Transportation GHG Emissions (kg CO2 eq/bbl crude) for Canadian and Foreign Crude Oils Used in the Base Case Refineries Located in QC, NB and NL Source: CERI calculations Notes: Stacked columns with borders represent Canadian crude oils, stacked columns without borders represent foreign crude oils. Transportation GHG emissions for AB Bitumen are presented per barrel of diluted bitumen, not per barrel of produced crude oil. Transportation by pipeline includes gathering pipeline and/or main pipeline. Transportation by tanker assumes tanker was offloaded offshore. For US LLS (LA), transportation by tanker also includes transportation by barge for a portion of the route.

97 Foreign Crude Oils Canadian Crude Oils Imperial, Sarnia, ON Shell Canada, ON Suncor Energy, ON Imperial, Nanticoke, ON Suncor Energy, QC Valero, QC Irving Oil, NB North Atlantic Refining, NL Imperial, Sarnia, ON Shell Canada, ON Suncor Energy, ON Imperial, Nanticoke, ON Suncor Energy, QC Valero, QC Irving Oil, NB North Atlantic Refining, NL Imperial, Sarnia, ON Shell Canada, ON Suncor Energy, ON Imperial, Nanticoke, ON Suncor Energy, QC Valero, QC Irving Oil, NB North Atlantic Refining, NL Imperial, Sarnia, ON Shell Canada, ON Suncor Energy, ON Imperial, Nanticoke, ON Suncor Energy, QC Valero, QC Irving Oil, NB North Atlantic Refining, NL Imperial, Sarnia, ON Shell Canada, ON Suncor Energy, ON Imperial, Nanticoke, ON Suncor Energy, QC Valero, QC Irving Oil, NB North Atlantic Refining, NL An Economic and Environmental Assessment of 79 Table 3.9: Transportation GHG Emissions (kg CO2 eq/bbl crude) for Canadian and Foreign Crude Oils Used in the Base Case and the Four Modelled Scenarios (Split by Refinery) Base Case 'Made in Canada' Scenario 'Expanded Access' Scenario 'Current Reality' Scenario 'International Social Concerns' Scenario Crude Oil Flow Name AB Light / AB SCO AB Heavy AB Bitumen MB Light SK Light ON Light NL Hibernia NL Light Offshore Algeria Saharan Blend Azerbaijan Azeri Light Colombia Castilla Blend Congo Djeno Denmark DUC Equatorial Guinea Zafiro Ivory Coast Baobab Ivory Coast Espoir Kazakhstan CPC Blend Nigeria Akpo Blend Nigeria Bonga Nigeria Bonny Light Nigeria Brass River Nigeria Qua Iboe Nigeria Usan Norway Ekofisk Norway Norwegian Crude Saudi Arabia Light UK Brent UK North Sea US Bakken (ND) US WTI US Eagle Ford (TX) US LLS (LA) US Thunder Horse (LA) US Michigan Light (MI) Source: CERI calculations Note: Color coding in Table 3.9 specifies modes of crude oil transportation to the refinery inlet and are defined as follows: transportation by gathering pipeline and main pipeline; by gathering and/or main pipeline and tanker; by gathering pipeline and railway; various modes of transportation, with a portion of crude flow transported by gathering pipeline, main pipeline and tanker (first number), and another portion transported by gathering pipeline and railway (second number); by tanker (offloaded offshore). For US LLS (LA), transportation by tanker also includes transportation by barge for a portion of the route.

98 An Economic and Environmental Assessment of 80 A review of transportation GHG emissions for the Base Case and the four modelled scenarios resulted in the following conclusions and observations: transportation GHG emissions intensity in terms of kg CO2eq/bbl crude for all Canadian crude oils shipped to the four refineries in ON has not changed in any of the four scenarios, in comparison with the Base Case; transportation GHG emissions for Canadian crude oils shipped to the refineries in ON range from 5.59 kg CO2eq/bbl (AB Heavy, by pipeline, to Imperial refinery in Nanticoke) to 0.26 kg CO2eq/bbl (ON Light, by gathering pipeline and rail, to Imperial refinery in Nanticoke) in the Base Case and all four scenarios; transportation emissions (kg CO2eq/bbl crude) for the foreign crude, US Bakken, imported to the ON refineries, have not changed in three scenarios in comparison with the Base Case (US Bakken has been fully substituted in the Made in Canada scenario); two US crude oils (Bakken and Michigan Light) are transported to Suncor Energy refinery in QC by pipeline, whereas all other foreign crudes imported to this refinery in the Base Case and the International Social Concerns scenario come by tankers (with gathering/international pipelines involved for the portion or the route); all foreign crude oils imported to the remaining three refineries (Valero in QC, Irving Oil in NB and North Atlantic Refining in NL & Labrador) are transported by tankers, or by international pipelines for the portion of route and tankers, in the Base Case and three scenarios (excluding the Made in Canada scenario, where all foreign crudes have been substituted); seven foreign crudes (Azerbaijan Azeri Light, Equatorial Guinea Zafiro, Kazakhstan CPC Blend, Nigeria Akpo Blend, Nigeria Bonga, Nigeria Brass River and Nigeria Qua Iboe) have been substituted in all four scenarios and have not been used except in the Base Case; the largest variety of crude flows is taken by the Irving Oil refinery in NB (the Base Case). There are three Canadian crudes (transportation GHG emissions range from to 0.98 kg CO2eq/bbl crude) and 14 foreign crudes (transportation emissions range from 7.30 to 2.82 kg CO2eq/bbl); transportation emissions for Canadian crudes in the Base Case (all refineries) range from CO2eq/bbl crude (AB SCO, by gathering pipeline and rail, to Irving Oil refinery, NB) to 0.26 CO2eq/bbl crude (ON Light, by gathering pipeline and rail, to Imperial refinery in Nanticoke); transportation GHG emissions for foreign crudes in the Base Case (all refineries) range from 7.70 CO2eq/bbl crude (Azerbaijan Azeri Light, by tanker and pipeline, to Suncor Energy refinery in QC) to 1.51 CO2eq/bbl crude (US Michigan Light, by pipeline, to Suncor Energy refinery in QC);

99 An Economic and Environmental Assessment of 81 while GHG emissions for the same mode of crude oil transportation will expectedly increase with distance, this increase was the most noticeable for NL & Labrador Light Offshore (0.32 CO2eq/bbl when transported by tanker to North Atlantic refinery vs CO2eq/bbl when transported by tanker to Suncor Energy refinery, QC). Similarly, for AB Light supplied to seven out of eight refineries in all four scenarios, transportation emissions have increased from 4.11 CO2eq/bbl (by pipeline to Imperial refinery, Sarnia) to 6.25 CO2eq/bbl (by pipeline to Irving Oil refinery, NB). Based on the results of GHG emissions modelling for each individual crude oil and crude flow, calculation of upstream, transportation and midstream emissions per crude stream, per refinery and per scenario was undertaken. Refer to Chapter 4 for additional details. However, CERI does not discuss emission costs in terms of carbon pricing. It should be clear that for impact of carbon pricing, the focus should be on emissions reported by refineries and associated with their respective production process, since this is the focus of current carbon pricing policies (either through provincial cap-and-trade/carbon tax) or federal output-based allocation systems for large final emitters. It is interesting to note that under QC and ON capand-trade systems, most of the emissions from oil refineries are covered by free allowances, as they are considered trade-exposed emitters.

100 82 Canadian Energy Research Institute

101 An Economic and Environmental Assessment of 83 Chapter 4: Modelling Results This section discusses the results of the modelling and is divided into four parts. Within each section, the results are reviewed and highlighted using the same four elements: crude flows (including transportation capacities used and remaining), costs of feedstock, emissions, and economic effect. It is important to note that the results of each of the scenarios is compared to the Base Case (2016). A set of drivers are presented to understand the difference between the Base Case and each scenario. The order of the scenario description and results is not indicative of the importance of any scenario. Made in Canada Scenario Made in Canada assumes that all Canadian crude will substitute all foreign imported crude oil, occurring via an expanded transportation infrastructure. This is a policy-based approach to substituting foreign crude oil in eastern Canadian refineries. Crude Flows Figure 4.1 illustrates the crude intakes for all the refineries in the Made in Canada scenario. As previously mentioned, the existing transportation infrastructure is augmented by the new pipeline, allowing the substitution of all foreign crude, 601 Mbpd. This includes an additional Mbpd of light and heavy crude from western Canada (408.4 Mbpd of light and 15.3 Mbpd of heavy), as well as Mbpd of light oil from eastern Canadian offshore assets. The scenario assumes the usage of almost all available oil from the eastern offshore (234.2 Mbpd) and all of SK (149.7 Mbpd). It is important to note that for AB, Mbpd of light oil remains available in the west for domestic refining after supplying Mbpd to the central and eastern refineries. In total, scenario modelling results in Mbpd of crude supply from Canadian western provinces, Mbpd from Canadian eastern offshore assets, and no imports.

102 Imperial Oil Shell Canada ON QC NB NL Suncor Energy Imperial Oil Suncor Energy Valero Irving Oil North Atlantic Refining 84 Canadian Energy Research Institute Figure 4.1: Total Crude Intake for Central and Eastern Refineries Made in Canada (Mbpd) Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario 77.8 Base Scenario 65.9 Base Scenario Base Western Eastern Imported A detailed review of crude substitution per refinery is illustrated in Table E.1 in Appendix E. While it is not realistic to review all the transactions, it is prudent to discuss the highlights. In ON, all of North Dakota s Bakken oil (58.1 Mbpd) is displaced with light oil from SK via Enbridge s Mainline. This occurs in all four of ON s refineries. In QC, the dynamics are more complex. Suncor Energy s refinery in Montreal is modelled to proceed using Enbridge Mainline and Line 9 to satisfy its crude intake. Feedstock is predominantly from AB, as well as SK. Although for reference, the cost of feedstock at the refinery gate for AB light oil is $1.38 per bbl cheaper using a new pipeline than using the Mainline and Line 9, based on projected tolls and current Enbridge tolls. Total replaced volume for the Suncor refinery is Mbpd. Valero s refinery in Lévis, on the other hand, is modelled to use a new pipeline, substituting all its foreign crude in the volume of Mbpd. In this scenario, Valero stops using Line 9 as well as its two tankers currently transporting crude oil between Montreal and Quebec City. Recall, supplying crude oil to Valero fully on new infrastructure pipeline versus utilizing the existing Line 9 and two tankers is an assumption of the study. However, during the modelling process, this assumption is underpinned by economics, as the costs of AB oil through Line 9 plus the usage of tankers is $1.66 per bbl more expensive for AB Light, $1.13 per bbl more expensive for AB SCO, and $2.52 per bbl more expensive for AB Heavy and bitumen.

103 An Economic and Environmental Assessment of 85 Irving Oil s refinery in NB is modelled to use both western oil as well as eastern oil to substitute its vast foreign crude slate. Western crude oil comes through new infrastructure pipelines, while eastern crude is delivered by tankers from offshore NL. NL & Labrador s North Atlantic Refining s foreign crude slate was fully substituted with nearby Canadian offshore production. The crude flows in the Made in Canada scenario are illustrated in Figure 4.2. Figure 4.2: Canadian Crude Supply to Central and Eastern Refineries Made in Canada (Mbpd) Source: Figure by CERI based on cartography from (Enbridge Inc. n.d.). The capacity of the expanded infrastructure scenario and the availability of light oil in western Canada would allow a full western Canadian crude substitution in NB. However, eastern Canadian crude costs $1.8 per bbl less then western crude at the gate of this refinery, hence, a significant portion of eastern crude supply is modelled the amount of additional eastern oil which was used for NB Mbpd makes up the remaining available eastern oil after satisfying demand of North Atlantic Refining of 93.1 Mbpd and delivering 5.8 Mbpd to Montreal s Suncor Energy. The availability and usage of crude and transportation infrastructure by central and eastern refineries is illustrated in Figure 4.3. Canada has enough light oil to be used in the eastern and western refinery markets. As mentioned above, after all foreign oil has been substituted, there is still Mbpd of light oil available in the western part of the country. The amount of SCO and Bitumen used by the eastern refinery market is small.

104 Transportation Crude 86 Canadian Energy Research Institute Figure 4.3: Availability and Usage of Crude and Infrastructure Made in Canada (Mbpd) Mbpd AB Light SK Light AB SCO AB Heavy SK Heavy AB Bitumen NL Offshore Light New infrastructue PL Mainline (Line78, 5) PL Portland-Montreal PL Line 9 PL Valero's tankers Rail Available Infrastructure Available Used Note: The volumes shown for Mainline (Line 78 and Line 5) represent Canadian crudes for Canadian refineries. The figure does not show movements in the Mainline of Canadian or US crudes destined for US refineries. In terms of infrastructure, several conclusions stand out. First, with new pipeline infrastructure being in place and used by Valero, Line 9 becomes underused. Simultaneously, there is zero usage of rail from west to east, and the Portland-Montreal pipeline becomes obsolete in its current flow direction. If Valero s refinery is modelled to use a mix of Line 9 (to its capacity), tankers (to their planned capacity) as well as the new infrastructure pipeline, the new infrastructure pipeline usage would drop from Mbpd to Mbpd. Rail and the Portland-Montreal Pipeline would not be needed. This is due to rail being more expensive than pipeline delivery. In addition, eastern offshore oil production in the Montreal refinery via the Portland-Montreal Pipeline is $0.48 per bbl more expensive than AB Light via the Mainline and Line 9, and $1.87 per bbl more expensive via the new infrastructure pipeline.

105 An Economic and Environmental Assessment of 87 Cost of Feedstock Feedstock costs varied across all examined refineries. This is illustrated in Table 4.1. Province Refinery Table 4.1: Cost of Feedstock Made in Canada Base Case ($bbl) Scenario ($bbl) Difference ($bbl) Difference Cost of Feedstock (annual, million $) 10-year Aggregated Difference Cost of Feedstock (million $) Annual Feedstock Costs (%) Imperial Oil ON Shell Canada Suncor Energy Imperial Oil QC Suncor Energy Valero NB Irving Oil North NL Atlantic Refining Total In ON, under the Made in Canada scenario, if all foreign oil (US Bakken) is substituted, all four refineries end up having more expensive feedstock. The results are driven by the difference in costs of US Bakken oil versus SK light oil $1.59 per bbl at the gates of the ON refineries (difference versus AB Light would be $3.69). The increase constitutes a 0.5 percent increase in the refineries annual costs and is due to the fact that oil from the Bakken is more cost competitive than its Canadian counterpart. On the other hand, if all foreign crude was substituted, the two QC refineries could save on feedstock costs approximately $103 million per year. For Suncor Energy, the difference in costs is driven by using AB Light ($59.97 per bbl) and SK Light ($57.75 per bbl) instead of oil from US WTI ($63.87 per bbl), Michigan ($61.96 per bbl), and Azerbaijan ($63.59 per bbl). In fact, AB Light and SK Light are cheaper than any other foreign light oil supplied to Suncor Energy s refinery in 2016, except for oil from the US Bakken. The difference between AB Light and North Dakota is $3.31 per bbl. This illustrates the competitiveness of Enbridge Mainline tariffs coupled with western crude prices. It also supports the economics of the Line 9 reversal project. Recall, transportation costs are included in these estimated costs. The Valero refinery in Lévis is the largest benefactor in this scenario, with potential savings of $80.8 million per year. The largest driver for such a decrease is the cost difference of oil from AB ($59.03 per bbl), via the new infrastructure pipeline, and from Algeria ($60.17 per bbl) and

106 88 Canadian Energy Research Institute Kazakhstan ($59.77 per bbl). Recall that these two countries represent 86 percent of foreign intake for this refinery. Overall, AB Light is cheaper than any foreign light oil at Valero s gate, except for US Texas Eagle Ford ($47.98 per bbl). North Atlantic refinery s cost of feedstock would decrease by almost $7 million annually if foreign crude was substituted entirely with eastern offshore crude, as domestic offshore crude ($58.73 per bbl) is cheaper than any 2016 foreign crude in the Base Case except for US Louisiana LLS ($55.46 per bbl) and Denmark DUC ($58.19 per bbl). The Irving Oil Refinery, on the other hand, will face higher feedstock costs in the amount of $51.1 million per year. This represents an increase in cost of around 1 percent. The reason for this increase is the refineries low-priced crude oil slate in the Base Case. For heavy oil, Irving Oil could benefit as AB Heavy costs $47.57 per bbl (transported via the new infrastructure pipeline route) and has lower tolls than heavy oils from the Ivory Coast ($54.41 per bbl) and Colombia ($48.13 per bbl). For light oil, AB Light ($60.62 per bbl) and Eastern Offshore Light ($58.83 per bbl) are more expensive than 7 of 13 imported crude brands, especially versus Saudi Arabia oil ($56.46 per bbl). The latter comprises nearly 40 percent of total imports of this refinery. Canadian crude is still less expensive than many imported crudes used for this refinery. The feedstock cost would be less if Irving Oil could use 100 percent of domestic NL offshore oil for substitution purposes. However, per the crude flows modelling, there is not enough light oil in eastern Canada to meet the two Atlantic Canada refineries full capacity. Emissions Based on the results of the emissions/lca modelling, total upstream, transportation and refining emissions decrease if all imports are substituted with Canadian crude supply, 6.2 percent less than the Base Case or 2.2 million tones CO2eq (MTCO2eq). In absolute terms, it is a decrease from MTCO2eq to MTCO2eq. This difference represents how much emissions would increase/decrease globally if substitution was to happen. It is important to note that this decrease represents a global emissions decrease, as these emissions reductions are not all realized within Canadian borders. The major driver for the emissions decrease is overall lower upstream emissions of Canadian light oil versus the imported slate (-2.24 MTCO2eq). Midstream emissions increase by 0.14 MTCO2eq, while transportation emissions decrease in the Made in Canada scenario by 0.12 MTCO2eq. The detailed information can be found in Appendix F in Table F.1. Figure 4.4 illustrates the difference between the Base Case and Made in Canada scenarios, by refinery. Emissions are also divided by upstream, refining (midstream) and transport.

107 Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario An Economic and Environmental Assessment of 89 Figure 4.4: Made in Canada and Base Case Emissions (tonnes CO2eq per year) 10,000,000 9,000,000 8,000,000 7,000,000 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 - Global-wide Total Emissions: -2,222,442 tones CO2eq Annually Imperial Oil Shell Canada Suncor Energy Imperial Oil Suncor Energy Valero Irving Oil North Atlantic Refining ON QC NB NL Upstream Midstream Transport In all ON refineries, as well as Suncor Energy in Montreal, total emissions increase modestly because of full substitution. ON s result is driven by the fact that North Dakota s light oil emission intensity of kg CO2eq/bbl is lower than SK Light which is used for substitution (emission intensity of kg CO2eq/bbl). For reference, AB Light average at ON refineries is kg CO2eq/bbl. For Suncor Energy s refinery in Montreal, the increase in emissions is linked to upstream activities, and less so, to midstream and transportation. The average of foreign oil total emissions is kg CO2eq/bbl in the Base Case, while the emissions intensities from AB Light is kg CO2eq/bbl and SK Light is kg CO2eq/bbl at the gate of this refinery. For Valero s refinery, full substitution of foreign oil decreases total emissions by 9.9 percent, on the upstream side. The average foreign crude emissions are kg CO2eq/bbl, while AB Light is kg CO2eq/bbl. Regarding the latter, the difference between the emissions intensity in AB Light for Valero and the refineries in ON is due to transportation. For reference, Algeria s Saharan Blend, which at 91.1 Mbpd, is the largest foreign intake for Valero, has total emissions of 92.9 kg CO2eq/bbl at refinery gate. Irving Oil s emissions profile decreases significantly, by 14.9 percent. Most of the decrease occurs in upstream and midstream activities, while transportation related emissions also decrease. Average foreign crude emissions are kg CO2eq/bbl, while AB Light via the new pipeline is kg CO2eq/bbl and the eastern offshore average is 37.7 kg CO2eq/bbl. Western Canadian

108 90 Canadian Energy Research Institute crude emissions profile was lower than 7 of 14 foreign crudes which were imported in the Base Case. Eastern offshore crude had lower emissions from any other crude oil in the Base Case, except for Azeri Light. The largest decrease of emissions stemmed from the North Atlantic Refinery, with a significant 64 percent reduction in total emissions, the majority of which is derived from upstream (75 percent drop) and transportation activities. Midstream activities, on the other hand, experience an increase in emissions. Total transportation emissions for the refinery decrease by 92 percent compared to the Base Case. This is due in part to the proximity of the upstream offshore activities to the refinery. Average foreign crude total emissions are kg CO2eq/bbl, while Eastern offshore crude at this refinery gate is kg CO2eq/bbl. The difference in the weighted average kg CO2eq per barrel per refinery is illustrated in Figure 4.5. As different refineries increase or decrease their total intensity (upstream, transportation and midstream), the average decrease at the national level is 5.7 kg CO2/bbl. To obtain each refinery s weighted average increase or decrease of emissions (kg/bbl), each refinery s absolute increase or decrease of emissions (kg/bbl) was multiplied by share of intake of the refinery in total central and eastern refineries crude intake. Figure 4.5: Change in Emissions Intensity Made in Canada (kg/bbl) Global-wide Total Emissions: -5.7 kg CO2eq / bbl

109 An Economic and Environmental Assessment of 91 Expanded Access Scenario Expanded Access assumes that Canadian crude will substitute foreign imported crude oil, occurring via an expanded transportation infrastructure (new pipeline). This is a market-based approach to substituting foreign crude oil in eastern Canadian refineries, meaning that more expensive foreign crude is displaced; cheaper foreign crude is kept in the slate of refineries. Crude Flows Figure 4.6 illustrates the crude intakes for all the refineries in the Expanded Access scenario. As previously mentioned, the existing transportation infrastructure is augmented by a new pipeline, allowing additional substitution of foreign crude, Mbpd. This includes an additional 248 Mbpd of light and heavy crude from western Canada (232.7 and 15.3 Mbpd, respectively), as well 96.1 Mbpd of light oil from eastern Canadian offshore assets. The modelling results in the usage of 65 percent of available light oil from the eastern offshore (153 Mbpd), 87 percent of SK light (130.7 Mbpd) and 56 percent of AB light (292.5 Mbpd). The remaining 42.8 percent is foreign crude that is cheaper than Canadian supply options. In total, scenario modelling results in Mbpd of crude supply from Canadian western provinces, Mbpd from Canadian eastern offshore assets, and Mbpd from foreign sources. It is important to note that Mbpd of light oil remains available in western Canada after supplying the above-mentioned volumes to central and eastern refineries; 81.5 Mbpd of light oil remains available in eastern Canada.

110 Imperial Oil Shell Canada ON QC NB NL Suncor Energy Imperial Oil Suncor Energy Valero Irving Oil North Atlantic Refining 92 Canadian Energy Research Institute Figure 4.6: Total Crude Intake for Central and Eastern Refineries Expanded Access (Mbpd) Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Western Eastern Imported A detailed review of crude substitution per refinery is illustrated in Table E.2 in Appendix E. While it is not realistic to review all the transactions, it is prudent to discuss the highlights. If foreign oil is mentioned as not substituted, then it is less expensive than any of the available Canadian supply under consideration. In ON, North Dakota s Bakken oil (58.1 Mbpd) is not displaced as it is cheaper than AB Light or SK Light by $1.59 and 3.69 per bbl, respectively. In QC, the dynamics are a little more complex. Suncor Energy s refinery in Montreal is modelled to proceed using Enbridge Mainline and Line 9 to satisfy its crude intake. North Dakota s Bakken (48 Mbpd) remains unchanged, while a mix of crude in the total volume of 55.1 Mbpd from US Michigan, US WTI, Azerbaijan, Norway, and UK are displaced with AB Light brought via the Enbridge Mainline and Line 9. For reference, the cost of feedstock at the refinery gate for AB light oil is $1.38 per bbl cheaper using the new pipeline, based on projected tolls and current Enbridge tolls.

111 An Economic and Environmental Assessment of 93 The Valero refinery in Lévis, on the other hand, is modelled to use the new pipeline. The competitive price of AB Light at this refinery gate allows to substitute almost all its foreign crude in the volume of Mbpd, except for 6.3 Mbpd Texas Eagle Ford imported from US Texas in the Base Case. In this scenario, Valero stops using Line 9 as well as its two tankers currently transporting crude oil between Montreal and Quebec City. Recall, supplying crude oil to Valero fully on the new infrastructure pipeline versus utilizing the existing Line 9 and two tankers is an assumption of the study. However, during the modelling process, this assumption appeared to be underpinned by economics, as the cost of AB oil through Line 9 plus usage of tankers is $1.66 per bbl more expensive for AB Light, $1.13 per bbl more expensive for AB SCO, and $2.52 per bbl more expensive for AB Heavy and Bitumen. For the Irving Oil refinery in NB, 50 percent of foreign feedstock is substituted with cheaper western or eastern Canadian oil. The remaining 50 percent of foreign intake is from the US (16.7 Mbpd), Congo (2.7 Mbpd), Ivory Cost (2.6 Mbpd) and Saudi Arabia (86.7 Mbpd), and is less expensive for the refinery than any Canadian option. The difference between AB Light via new pipeline and Saudi Arabia light is $4.16 per bbl. NL s North Atlantic Refining s foreign crude slate was substituted by 60 percent with nearby Canadian offshore production. US Louisiana LLS and oil from Denmark are not substituted as the former is $3.27 per bbl cheaper than Canadian offshore and the latter is cheaper by $0.55 per bbl. The crude flows in the Expanded Access scenario are illustrated in Figure 4.7. Figure 4.7: Canadian Supply to Central and Eastern Refineries Expanded Access (Mbpd) Source: Figure by CERI based on cartography from (Enbridge Inc. n.d.)

112 Transportation Crude 94 Canadian Energy Research Institute The availability and usage of crude and transportation infrastructure by central and eastern refineries is illustrated in Figure 4.8. Canada has enough light oil to be used in the eastern and western refinery markets. As mentioned above, after all foreign oil being substituted, there is still 330 Mbpd of light oil available in the western and eastern parts of the country. The amount of SCO and Bitumen used by the eastern refinery market is small. Figure 4.8: Availability and Usage of Crude and Infrastructure Expanded Access (Mbpd) Mbpd AB Light SK Light AB SCO AB Heavy SK Heavy AB Bitumen NL Offshore Light New infrastructue PL Mainline (Line78, 5) PL Portland-Montreal PL Line 9 PL Valero's tankers Rail Available Infrastructure Available Used Note: The volumes shown for Mainline (Line 78 and Line 5) represent Canadian crudes for Canadian refineries. The figure does not show movements in the Mainline of Canadian or US crudes destined for US refineries. In terms of infrastructure, several conclusions stand out. First, with new pipeline infrastructure being in place and used by Valero, Line 9 becomes underused. Simultaneously, there is zero usage of rail from west to east and the Portland-Montreal pipeline. If Valero s refinery is modelled to use a mix of Line 9 (to its capacity), tankers (to their planned capacity) as well as the new infrastructure pipeline, the new infrastructure pipeline usage would drop from Mbpd to Mbpd. Rail and the Portland-Montreal Pipeline would not be needed. This is due to rail being more expensive than pipeline delivery. In addition, eastern offshore oil production in the Montreal refinery via the Portland-Montreal Pipeline is $0.48 per bbl more expensive than AB

113 An Economic and Environmental Assessment of 95 Light via the Mainline and Line 9, and $1.87 per bbl more expensive via the new infrastructure pipeline. Cost of Feedstock The feedstock costs varied across all examined refineries. This is illustrated in Table 4.2. Province Refinery Table 4.2: Cost of Feedstock Expanded Access Base Case ($bbl) Scenario ($bbl) Difference ($bbl) Difference Cost of Feedstock (annual, million $) 10-year Aggregated Difference Cost of Feedstock, (million $) Annual Feedstock Costs (%) Imperial Oil ON Shell Canada Suncor Energy Imperial Oil QC Suncor Energy Valero , NB Irving Oil , NL North Atlantic Refining Total , In ON, under the Expanded Access scenario, as no oil is displaced, there is no difference to the Base Case. For the other three provinces, the partial substitution of more expensive foreign oil would save $317.4 million on feedstock annually. For Suncor Energy, the difference in costs is driven by using AB Light ($59.97 per bbl) and SK Light ($57.75 per bbl) instead of oil from US WTI ($63.87 per bbl), Michigan ($61.96 per bbl), and Azerbaijan ($63.59 per bbl). In fact, AB Light and SK Light are cheaper than any other foreign light oil supplied to Suncor Energy s refinery in the Base Case, except oil from the Bakken. The difference between AB Light and North Dakota is $3.31 per bbl. The Valero refinery in Lévis is the largest benefactor in this scenario, with potential savings of $106 million per year (or 2.1 percent). The largest driver for the decrease is the cost difference of oil from AB ($59.03 per bbl), via new infrastructure pipeline, and crude from Algeria ($60.17 per bbl) and Kazakhstan ($59.77 per bbl). Recall that these two countries represent 86 percent

114 96 Canadian Energy Research Institute of foreign intake for this refinery. Overall, AB Light is cheaper than any foreign light oil at Valero s gate, except for US Texas Eagle Ford ($47.98 per bbl). NB and the Irving Oil Refinery is the second-largest benefactor as its feedstock cost decreases by $100.4 million per year. This represents a decrease in cost of 1.7 percent. On the heavy oil side, Irving Oil could benefit as AB Heavy costs $47.57 per bbl (transported via the new infrastructure pipeline route) versus heavy oils from the Ivory Coast ($54.41 per bbl) and Colombia ($48.13 per bbl). On the light oil side, AB Light ($60.62 per bbl) and Eastern Offshore Light ($58.83 per bbl) are less expensive than 6 of the 13 imported crude brands US WTI ($63.54 per bbl), Nigeria Bonga ($61.53), and Azerbaijan ($62.77), to name a few. Saudi Arabia oil, which accounts for 40 percent in imports, is very competitive in price at the refinery gate to be substituted ($56.46 per bbl). North Atlantic refinery s cost of feedstock would decrease by almost $48.15 million annually if foreign crude was substituted entirely with eastern offshore crude, as domestic offshore crude ($58.73 per bbl) is cheaper than any 2016 foreign crude in the Base Case except for US Louisiana LLS ($55.46 per bbl) and Denmark DUC ($58.19 per bbl). Emissions Based on the results of the modelling, total upstream, transportation and refining emissions decrease in this scenario and will equal 2.1 million tones CO2eq (MTCO2eq) or 5.1 percent from the Base Case. In absolute terms, it is a decrease from MTCO2eq to MTCO2eq. This difference represents how much emissions would increase/decrease globally if substitution was to happen. It is important to note that this decrease represents a global emission decrease, as these emissions reductions are not all realized within Canadian borders. The major driver for the emissions decrease is overall lower upstream emissions of Canadian light oil versus the imported slate (by 1.96 MTCO2eq). However, emissions in both refining and transportation are also lower in this scenario compared to the Base Case, by a total of MTCO2eq (0.068 MTCO2eq for midstream, and MTCO2eq for transportation). The detailed information can be found in Appendix F in Table F.2. Figure 4.9 illustrates the difference between the Base Case and Expanded Access scenario, by refinery. Emissions are also divided by upstream, refining (midstream) and transport.

115 Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario An Economic and Environmental Assessment of 97 Figure 4.9: Expanded Access and Base Case Emissions (tones CO2eq per year) 10,000,000 9,000,000 8,000,000 7,000,000 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 - Global-wide Total Emissions: -2,048,275 tones CO2eq Annually Imperial Oil Shell Canada Suncor Energy Imperial Oil Suncor Energy Valero Irving Oil North Atlantic Refining ON QC NB NL Upstream Midstream Transport In all ON refineries, emissions stay the same as no substitution of oil is occurring. For Suncor Energy s refinery in Montreal, the increase in emissions is linked to predominantly upstream activities, and less so, to midstream and transportation. The average of foreign oil total emissions intensity is kg CO2eq/bbl in the Base Case, while AB Light is kg CO2eq/bbl at the gate of this refinery. For Valero s refinery, full substitution of foreign oil decreases total emissions by 10.4 percent, driven by lower emissions on the upstream side. The average foreign crude emissions intensity is kg CO2eq/bbl, while AB Light is kg CO2eq/bbl. For reference, Algeria s Saharan Blend, which at 91.1 Mbpd is the largest foreign intake for Valero, has total emissions intensity of 92.9 kg CO2eq/bbl at refinery outlet gate. Irving Oil s emissions profile decreases by 7.6 percent. The decrease occurs in all the three components upstream, midstream and transportation. Recall, that in this scenario all substituted oil comes from the west by new infrastructure pipeline. Average foreign crude emissions intensity average is kg CO2eq/bbl, while AB Light via the new pipeline is kg CO2eq/bbl and the Eastern offshore average is 37.7 kg CO2eq/bbl. The western Canadian crude emissions profile was lower than 7 of 14 foreign crudes which were imported in the Base Case. Eastern offshore crude had lower emissions from any other crude oil in the Base Case, except for Azeri Light.

116 98 Canadian Energy Research Institute The largest decrease of emissions stemmed from the North Atlantic Refinery, with a significant 24 percent reduction in total emissions, all of which is derived from upstream (31% drop) and transportation activities; midstream activities, on the other hand, experience an increase in emissions. Total transportation emissions for the refinery decrease by 56 percent compared to the Base Case. This is due in part to the proximity of the upstream offshore activities to the refinery. Average foreign crude total emissions are kg CO2eq/bbl, while Eastern offshore crude at this refinery gate is kg CO2eq/bbl. The difference in the weighted average kg CO2eq per barrel per refinery is illustrated in Figure As different refineries increase or decrease their total intensity (upstream, transportation and midstream), the total decrease on the Canadian level is 5.2 kg CO2/bbl. To obtain each refinery s weighted average increase or decrease of emissions (kg/bbl), each refinery s absolute increase or decrease of emissions (kg/bbl) was multiplied by share of intake of the refinery in total central and eastern refineries crude intake. Figure 4.10: Change in Emissions Intensity Expanded Access (kg/bbl) Global-wide Total Emissions: -5.2 kg CO2eq / bbl

117 An Economic and Environmental Assessment of 99 Current Reality Scenario Current Reality assumes that Canadian crude will substitute foreign imported crude oil, occurring via existing transportation infrastructure pipelines, marine tankers, or rail. This is a marketbased approach to substituting foreign crude oil in eastern Canadian refineries meaning that more expensive foreign crude is displaced; cheaper foreign crude is kept in the slate of refineries. Crude Flows Figure 4.11 illustrates the crude intakes for all the refineries in the Current Reality scenario. As previously mentioned, the existing transportation infrastructure allows substitution of foreign crude in the volume of Mbpd (47 percent). This includes an additional Mbpd of light crude from western Canada, as well as 160 Mbpd of light oil from eastern Canadian offshore assets. The modelling results in the usage of 93 percent of available light oil from the eastern offshore (216.9 Mbpd), 100 percent of SK Light (149.7 Mbpd) and 31 percent of AB Light (160.5 Mbpd). The remaining 53 percent of foreign crude is cheaper than Canadian supply options. In total, scenario modelling results in Mbpd of crude supply from Canadian western provinces, Mbpd from Canadian eastern offshore assets, and Mbpd from foreign sources. It is important to note that Mbpd of light oil remains available in western Canada after supplying the above-mentioned volumes to central and eastern refineries and 17.5 Mbpd of light oil remains available in eastern Canada.

118 Imperial Oil Shell Canada ON QC NB NL Suncor Energy Imperial Oil Suncor Energy Valero Irving Oil North Atlantic Refining 100 Canadian Energy Research Institute Figure 4.11: Total Crude Intake for Central and Eastern Refineries Current Reality (Mbpd) Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Western Eastern Imported A detailed review of crude substitution per refinery is illustrated in Table E.3 in Appendix E. While it is not realistic to review all the transactions, it is prudent to discuss the highlights. If foreign oil is not substituted, the reason is it was cheaper than any of the available Canadian supply under consideration (western and eastern) unless stated otherwise. In ON, none of North Dakota s Bakken oil (58.1 Mbpd) is displaced as it is cheaper than AB Light or SK Light by $1.59-$3.69 per bbl. In QC, the dynamics are a little more complex. Suncor Energy s refinery in Montreal is modelled to proceed using Enbridge Mainline and Line 9 to satisfy its crude intake. North Dakota s Bakken (48 Mbpd) is also not substituted, while the mix of crude in the total volume of 55.1 Mbpd from US Michigan, US WTI, Azerbaijan, Norway, and UK are displaced with AB Light brought via Mainline and Line 9. For reference, the cost of feedstock at the refinery gate for AB Light oil is $1.38 per bbl cheaper using a new pipeline, based on projected tolls and current Enbridge tolls.

119 An Economic and Environmental Assessment of 101 The Valero refinery in Lévis is modelled to use the Enbridge Mainline, Line 9 and tankers from Montreal. The competitive price of AB Light and NL Offshore Light at this refinery gate allows, from a price standpoint, to substitute all its foreign crude, except US Texas Eagle Ford. However, as offshore oil is also used in the Maritimes, the modelling showed not enough available oil from eastern assets. Thus, in addition to US Texas Eagle Ford, not all Algerian Saharan Blend oil was substituted (64 percent out of 92.1 Mbpd). In total, 66.6 Mbpd is displaced using AB Light, SK Light, and NL Offshore Light. For the Irving Oil refinery in NB, 36 percent of foreign feedstock could be substituted with cheaper western or eastern oil. The remaining part of the foreign intake, which comes in the Base Case from several countries, including the US (all by WTI), Saudi Arabia, Ivory Coast, Congo and others, is kept in the slate as more cost-effective for the refinery. Of the 78.3 Mbpd of Canadian crude used for substitution, more than half, 45.1 Mbpd, is supplied from eastern offshore production. The remaining western crude comes by rail and via Mainline, Line 9 and tanker from Montreal to Saint John. Colombia heavy crude, US (except for WTI) light, Congo light, Ivory Coast light, Saudi Arabian light, and Nigeria Usan brands are not displaced as they are cheaper than Canadian alternatives. The Ivory Coast s heavy oil volume of 10 Mbpd, though more expensive than AB Heavy, is not substituted as Line 9 capacity is reached. NL & Labrador s North Atlantic Refining s foreign crude slate was substituted by 60 percent with nearby Canadian offshore production. US Louisiana LLS and oil from Denmark could not be substituted as the former is $3.27 per bbl cheaper than Canadian offshore, and the latter is cheaper by $0.55 per bbl. The crude flows in the Current Reality scenario are illustrated in Figure 4.12.

120 102 Canadian Energy Research Institute Figure 4.12: Canadian Supply to Central and Eastern Refineries Current Reality (Mbpd) Source: Figure by CERI based on cartography from (Enbridge Inc. n.d.) The availability and usage of crude and transportation infrastructure by central and eastern refineries is illustrated in Figure Canada has enough light oil to be used in the eastern and western refinery market. As mentioned above, after all foreign oil being substituted, there is still Mbpd of light oil available in the western and eastern parts of the country. The amount of SCO and bitumen used by the eastern refinery market is small.

121 Transportation Crude An Economic and Environmental Assessment of 103 Figure 4.13: Availability and Usage of Crude and Infrastructure Current Reality (Mbpd) Mbpd AB Light SK Light AB SCO AB Heavy SK Heavy AB Bitumen NL Offshore Light New infrastructue PL Mainline (Line78, 5) PL Portland-Montreal PL Line 9 PL Valero's tankers Rail Available Infrastructure Available Used Note: The volumes shown for Mainline (Line 78 and Line 5) represent Canadian crudes for Canadian refineries. The figure does not show movements in the Mainline of Canadian or US crudes destined for US refineries. In terms of infrastructure, several conclusions stand out. First, Valero s tanker fleet and Line 9 becomes used to full capacity. Second, there is zero usage of the Portland-Montreal pipeline. Third, Line 9 capacity becomes a bottleneck on the path of western crude to NB; however, this is not a significant bottleneck as 10 Mbpd of heavy oil could not be delivered.

122 104 Canadian Energy Research Institute Cost of Feedstock Feedstock costs varied across all examined refineries. This is illustrated in Table 4.3. Province ON QC Refinery Table 4.3: Cost of Feedstock Current Reality Base Case ($bbl) Scenario ($bbl) Difference ($bbl) Difference Cost of Feedstock (annual, million $) 10-year Aggregated Difference Cost of Feedstock (million $) Annual Feedstock Costs (%) Imperial Oil Shell Canada Suncor Energy Imperial Oil Suncor Energy Valero NB Irving Oil NL North Atlantic Refining Total , In ON, under the Current Reality scenario, as no oil is displaced, there is no difference to the Base Case. For all other provinces, the partial substitution of more expensive foreign oil drives a total $ million of savings on feedstock. For Suncor Energy, cost savings is the same as the Expanded Access scenario (for additional detail and drivers please see previous scenario). Valero refinery s savings on feedstock reach $44.5 million per year or 0.9 percent. The savings come from using AB Light ($60.68 per bbl) versus Nigeria AKPO ($60.93 per bbl) and Nigeria Brass River ($61.02 per bbl); SK Light ($58.46 per bbl) versus Kazakhstan ($59.77 per bbl); NL Offshore Light ($58.68 per bbl) versus Algeria ($60.17 per bbl). The Irving Oil Refinery is the second-largest benefactor as its feedstock cost decreases by $54.4 million per year. The decrease of costs of feedstock come from using AB Light ($60.94 per bbl) and Eastern Offshore Light ($58.83 per bbl) instead of US WTI ($63.54 per bbl), Nigeria Bonga ($61.53), Azerbaijan ($62.77), and Norway ($60.02) to name a few. Saudi Arabia oil, which

123 An Economic and Environmental Assessment of 105 accounts for 40 percent of imports, is very competitive in price at the refinery gate to be substituted ($56.46 per bbl). The North Atlantic refinery s cost of feedstock would decrease by $48.2 million annually if foreign crude was substituted entirely with eastern offshore crude. Cost savings are the same as for the Expanded Access scenario (for additional detail and drivers please see previous scenario). Emissions Based on the results of the modelling, total upstream, transportation and refining emissions decrease in this scenario by 2.0 million tones CO2eq (MTCO2eq) or 5.7 percent from the Base Case. In absolute terms, it is a decrease from MTCO2eq to MTCO2eq. This difference represents how much emissions would decrease globally if substitution was to happen. It is important to note that this decrease represents a global emission decrease, as these emissions reductions are not all realized within Canadian borders. The major driver for the emissions decrease is overall lower upstream emissions of Canadian light oil versus the imported slate (by 2.23 MTCO2eq). Transportation emissions are also lower in the Current Reality scenario by a total of 0.14 MTCO2eq, however, midstream emissions grew by 0.33 MTCO2eq. Detailed information can be found in Appendix F in Table F.3. Figure 4.14 illustrates the difference between the Base Case and Current Reality scenarios, by refinery. Emissions are also divided by upstream, refining (midstream) and transport.

124 Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario Base Scenario 106 Canadian Energy Research Institute 10,000,000 9,000,000 Figure 4.14: Current Reality and Base Case Emissions (tones CO2eq per year) 8,000,000 7,000,000 6,000,000 Global-wide Total Emissions: -2,035,632 tones CO2eq Annually 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 - Imperial Oil Shell Canada Suncor Energy Imperial Oil Suncor Energy Valero Irving Oil North Atlantic Refining ON QC NB NL Upstream Midstream Transport In all ON refineries, emissions stay the same as no substitution of oil is happening. For Suncor Energy s refinery in Montreal, the increase in emissions is identical to the Expanded Access scenario (as the same substitution happens). For Valero s refinery, substitution of foreign oil deceases total emissions by 13 percent, on the upstream side and on the transportation side. The average foreign crude emissions intensity is kg CO2eq/bbl, while AB Light is kg CO2eq/bbl. For reference, Algeria s Saharan Blend, which at 91.1 Mbpd is the largest foreign intake for Valero, has total emissions intensity of 92.9 kg CO2eq/bbl at refinery gate. Irving Oil s emissions profile decreases by 4.5 percent. The decreases occur in the upstream and transportation emissions, with increases in midstream. The average foreign crude emissions intensity average is kg CO2eq/bbl, while AB Light via Mainline, Line 9 and tanker is kg CO2eq/bbl and the Eastern offshore average is 37.7 kg CO2eq/bbl. The western Canadian crude emissions profile was lower than 7 of 14 foreign crudes which were imported in the Base Case. Eastern offshore crude had lower emissions from any other crude oil in the Base Case, except for Azeri Light.

125 An Economic and Environmental Assessment of 107 The largest decrease of emissions stemmed from the North Atlantic Refinery, with a significant 24 percent reduction in total emissions. The emissions decreases are identical to the Expanded Access scenario. The difference in the weighted average kg CO2eq per barrel per refinery is illustrated in Figure As different refineries increase or decrease their total intensity (upstream, transportation and midstream), the total decrease on the national level is 5.2 kg CO2/bbl. To obtain each refinery s weighted average increase or decrease of emissions (kg/bbl), each refinery s absolute increase or decrease of emissions (kg/bbl) was multiplied by share of intake of the refinery in total central and eastern refineries crude intake. Figure 4.15: Change in Emissions Intensity Current Reality (kg CO2eq/bbl) Global-wide Total Emissions: -5.2 kg CO2eq / bbl International Social Concerns Scenario International Social Concerns assumes that Canadian crude will substitute foreign imported crude oil from our proxy index for those concerns (undemocratic states), occurring via existing transportation infrastructure pipelines, marine tankers, or rail. In 2016, Canada imported crude oil from a range of nations (in alphabetical order) including Algeria, Azerbaijan, Colombia, Congo, Denmark, Equatorial Guinea, Ivory Coast, Kazakhstan, Nigeria, Norway, Saudi Arabia, United Kingdom and the US (Government of Canada 2017; NEB 2017f; Statistics Canada 2017a). Recall, CERI used the Economist Intelligence Unit s Democracy Index 2016, which fits into four categories

126 108 Canadian Energy Research Institute including full democracy, flawed democracy, hybrid regime, and authoritarian (The Economist Intelligence Unit 2017). This is a policy-based approach to substituting foreign crude oil in eastern Canadian refineries meaning that crudes from those states is displaced; it is important to note that more expensive oil from democratic states are kept in the crude slate of eastern Canadian refineries. This approach allows CERI to assess the direct impact on costs and emissions tied to substitution of oil from states that cause international social concerns regarding the treatment of their citizens or the environment. Crude Flows Figure 4.16 illustrates the crude intakes for all the refineries in the International Social Concerns scenario. As previously mentioned, the existing transportation infrastructure is used to substitute for foreign crude, Mbpd (50 percent). This includes an additional 123 Mbpd of light crude from western Canada, as well as Mbpd of light oil from eastern Canadian offshore assets. The modelling results in the usage of 100 percent of available light oil from the eastern offshore (234.5 Mbpd), 100 percent of SK Light (149.7 Mbpd) and 22 percent of AB Light (115.9 Mbpd). The remaining crude comes from other foreign states. In total, scenario modelling results in Mbpd of crude supply from Canadian western provinces, Mbpd from Canadian eastern offshore assets, and Mbpd from foreign sources. It is important to note that Mbpd of light oil remains available in western Canada after supplying the above-mentioned volumes to central and eastern refineries.

127 An Economic and Environmental Assessment of 109 Figure 4.16: Total Crude Intake for Central and Eastern Refineries International Social Concerns (Mbpd) A detailed review of crude substitution per refinery is illustrated in Table E.4 in Appendix E. While it is not realistic to review all the transactions, it is prudent to discuss the highlights. If foreign oil is mentioned as not substituted, then it is less expensive than any of the available Canadian supply under consideration (western and eastern), unless stated otherwise. In ON, North Dakota s Bakken oil (58.1 Mbpd) is not displaced, as it comes from the US. In QC, 4.5 Mbpd from Azerbaijan is displaced with AB Light in Suncor Energy s refinery in Montreal. The remaining 98.7 Mbpd originate in democratic countries including the US, the UK and Norway. The substitution share in Valero is higher than in Montreal; out of 132 Mbpd, 7.2 Mbpd is kept intact as it is supplied from the US. The bulk of imported crude is from Algeria, Kazakhstan, and Nigeria. This crude slate is substituted with a mix of AB Light, SK Light, and NL Offshore Light. Western Canadian crude oil is delivered via Enbridge Mainline, Line 9 and tankers, as well as rail. The rail shipment comprises 50 percent of Valero s rail offloading capacity