Well-to-Wheels Greenhouse Gas Emissions of Canadian Oil Sands

Jun 9, 2015 - (1, 2) Canadian oil sands production is projected to grow 2.5-fold, from ..... requirements from the literature and was compared with co...
0 downloads 0 Views 814KB Size
Article pubs.acs.org/est

Well-to-Wheels Greenhouse Gas Emissions of Canadian Oil Sands Products: Implications for U.S. Petroleum Fuels Hao Cai,*,† Adam R. Brandt,‡ Sonia Yeh,§ Jacob G. Englander,‡ Jeongwoo Han,† Amgad Elgowainy,† and Michael Q. Wang† †

Systems Assessment Group, Energy Systems Division, Argonne National Laboratory, 9700 South Cass Avenue, Lemont, Illinois 60439, United States ‡ Department of Energy Resources Engineering, Stanford University, 367 Panama Street, Green Earth Sciences Building, 065, Stanford, California 94305, United States § Institute of Transportation Studies, University of California, 1605 Tilia Street, Davis, California 95616, United States S Supporting Information *

ABSTRACT: Greenhouse gas (GHG) regulations affecting U.S. transportation fuels require holistic examination of the life-cycle emissions of U.S. petroleum feedstocks. With an expanded system boundary that included land disturbanceinduced GHG emissions, we estimated well-to-wheels (WTW) GHG emissions of U.S. production of gasoline and diesel sourced from Canadian oil sands. Our analysis was based on detailed characterization of the energy intensities of 27 oil sands projects, representing industrial practices and technological advances since 2008. Four major oil sands production pathways were examined, including bitumen and synthetic crude oil (SCO) from both surface mining and in situ projects. Pathway-average GHG emissions from oil sands extraction, separation, and upgrading ranged from ∼6.1 to ∼27.3 g CO2 equivalents per megajoule (in lower heating value, CO2e/MJ). This range can be compared to ∼4.4 g CO2e/MJ for U.S. conventional crude oil recovery. Depending on the extraction technology and product type output of oil sands projects, the WTW GHG emissions for gasoline and diesel produced from bitumen and SCO in U.S. refineries were in the range of 100−115 and 99−117 g CO2e/MJ, respectively, representing, on average, about 18% and 21% higher emissions than those derived from U.S. conventional crudes. WTW GHG emissions of gasoline and diesel derived from diluted bitumen ranged from 97 to 103 and 96 to 104 g CO2e/MJ, respectively, showing the effect of diluent use on fuel emissions.



INTRODUCTION

in U.S. refineries in 2013 and forecast to reach 13.6% in 2020 (see Supporting Information, Table S3), Canadian oil sands are an increasingly important factor affecting the carbon intensities of U.S. petroleum fuels. Oil sands plants and upgraders generated about 7.8% of Canada’s total GHG emissions in 2011 and are projected to increase their GHG emissions from 55 million tonnes in 2011 to 101 million tonnes in 2020.4 Compliance with low-carbon fuel regulations, such as the California Low-Carbon Fuel Standard (LCFS)5 and the European Union Fuel Quality Directive,6 requires assessment of life-cycle GHG emissions, including the crude recovery stage for both conventional crude

Crude oil resources around the world vary significantly in regard to quality and production methods, resulting in significant variation in greenhouse gas (GHG) emission intensities associated with crude recovery.1,2 Canadian oil sands production is projected to grow 2.5-fold, from 1.95 million barrels per day (bpd) in 2013 to 4.81 million bpd by 2030, much of which could end up in U.S. refineries.3 In 2013, 0.85 million bpd were from surface mining, which removes the oil sands material with shovels and trucks and separates the bitumen (the highly viscous oil) from the sand using hot water. Another 1.1 million bpd were recovered from in situ projects, which employed Steam-Assisted Gravity Drainage (SAGD) and Cyclic Steam Stimulation (CSS) as the main thermal recovery technologies.3 Looking ahead to 2030, mining and in situ production are forecast to increase to 1.6 and 3.2 million bpd, respectively.3 Accounting for 9.4% of the total crudes processed © 2015 American Chemical Society

Received: Revised: Accepted: Published: 8219

March 11, 2015 June 8, 2015 June 9, 2015 June 9, 2015 DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology

bitumen (IS+B), surface mining bitumen (M+B), and in situ SCO (IS+SCO), were considered separately to evaluate the impact of differences in oil sands production technologies and types of products on energy and emission intensities. We expanded and employed the GREET model to estimate the WTW GHG (including CO2, CH4, N2O, black carbon [BC], and primary organic carbon [POC]) emissions of these pathways and analyzed their impacts on WTW GHG emissions of petroleum-derived fuels in the U.S. We used the Global Warming Potentials of 30, 265, 900, and 69 for CH4, N2O, BC, and POC, respectively, according to the Fifth Assessment Report of the United Nations Intergovernmental Panel on Climate Change. We adopted the methodology and data for previous BC and POC emissions analysis.21 See Supporting Information, Section 2, for WTW GHG emission calculations. The life-cycle system boundary includes the production of bitumen and SCO from recovery, separation, and upgrading of oil sands; production and transportation of diluent for bitumen dilution; transportation of dilbit and SCO to U.S. refineries; refining of bitumen, SCO, and dilbit in U.S. refineries; transportation and distribution of fuels; and combustion of fuels (see Supporting Information, Figure S1). Oil Sands Production Pathways. Facility-level energy consumption and emissions data were collected on a monthly basis for seven M+SCO projects and three M+B projects over the period from 2008 to 2012 and 15 IS+B projects and two IS +SCO projects for the period from 2009 to 2012.20 The cogeneration and noncogeneration projects were aggregated for each pathway on a production-weighted average basis. In 2012, the M+SCO and IS+B pathways accounted for 41.7% and 47.1%, respectively, of the total production of oil sands products by energy, while the M+B and IS+SCO pathways contributed only 5.3% and 6.0%, respectively, of the total oil sands production.20 For the two in situ pathways, SAGD and CSS technologies accounted for 51% and 49% of production, respectively. Of these thermal in situ methods, SAGD has a lower energy intensity than CSS, as SAGD typically has lower steam requirements per unit of bitumen produced.20 Energy Efficiencies and Process Fuel Consumption. Each oil sands project is distinct because of the differing characteristics of oil sands reservoirs, recovery technologies, and operational choices. Despite improved energy efficiencies of oil sands production over the period 1970−2010,15 a continued trend was unclear for the major oil sands production pathways over the most recent period 2008−2012 (see Supporting Information, Figure S2). Therefore, we focused on the multiyear production-weighted average energy efficiency and process fuel consumption for each pathway, primarily based on monthly records of energy consumption by fuel type and oil sands production from 2008 to 201220 (see Supporting Information, Figure S3). Characterization of month-to-month variability in the operational energy requirements of oil sands extraction and upgrading among projects (see error bars in Figure S3) allowed us to analyze the impacts of variability on the performance of each oil sands pathway on GHG emissions. See Supporting Information, Section 6, for details on the variability/uncertainty analysis and the way we interpreted the variability/uncertainty of the results. Cogeneration. While some projects purchase grid electricity, many cogenerate steam and electricity. When all oil sands projects were combined on a production-weighted average basis, there was a net surplus of cogenerated electricity, which was sold back to the grid.20 To include the cogeneration

and oil sands production. A variety of models exist to assess these emissions. General well-to-wheels (WTW) models such as Greenhouse gases, Regulated Emissions, and Energy use in Transportation (GREET)7 and GHGenius8 were developed to analyze WTW GHG emissions of various transportation fuel pathways. Other models focus more narrowly on hydrocarbon production. The Oil Production Greenhouse gas Emissions Estimator was developed as an engineering-based life-cycle assessment tool for estimating GHG emissions from the recovery, processing, and transport of crude petroleum.9 Using proprietary oil sands project operating data, the GreenHouse gas emissions of Oil Sands Technologies tool was developed to quantify well-to-refinery entrance gate (WTR) GHG emissions associated with production of diluted bitumen (dilbit) and synthetic crude oil (SCO).10 Previous studies reported generally higher GHG emissions from production of oil sands-derived fuels than conventional crude-oil-derived fuels.1,2,11−15 The range of the reported higher emissions, however, varies among studies, mainly owing to the use of different data and analysis boundaries; different assumptions about the source, type, share, and fate of diluents; different electricity sources displaced by the surplus electricity from cogeneration facilities; and different refining efficiencies and emissions for fuel products. Assessments based on proprietary data from a set of operating projects10,14 provided new estimates of life-cycle GHG emissions of oil sands pathways, but the limited transparency compromised the ability of other researchers to replicate the results. Because refining is an important component of WTW GHG emissions of oil sandsderived fuels, GHG emissions from refining of Canadian oil sands have been simulated by engineering-based process models, considering the impacts of qualities such as specific gravity and sulfur content.1,2,16,17 Canadian oil sands producers have recently improved the energy efficiency of their processes and deployed GHG mitigation technologies.18,19 The impacts of these efforts on the carbon intensities of Canadian oil sands products require evaluation. In this study, we (1) evaluated the WTR GHG emissions of four Canadian oil sands production options that employ surface mining and in situ technologies and produce SCO or bitumen, using public facility-level data that reflect recent trends of oil sands project operations,20 (2) estimated the WTW emissions of gasoline and diesel fuels derived wholly from Canadian oil sands products, i.e., SCO and bitumen, and (3) explored the WTW GHG emissions of gasoline and diesel derived from dilbit. Compared to previous work, this study improves the representativeness, transparency, and completeness of the carbon footprint estimates for the Canadian oil sands industry. This study advances our understanding of oil sands emissions by applying the GREET model with projectlevel intensities of process fuel consumption and flaring activities for 27 operating oil sands projects since 2008. Moreover, this study leverages newly compiled, entirely public data sources that represent the industrial practices and recent technological advances in oil sands production. Lastly, this analysis incorporates much more detailed estimates of land use disturbance-induced GHG emissions obtained from in-depth analysis of field operations involved in surface mining and in situ projects.



METHODOLOGY AND DATA System Boundary of WTW Analysis. Four oil sands pathways, namely, surface mining SCO (M+SCO), in situ 8220

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology

Figure 1. WTR GHG emissions for bitumen and SCO from four major Canadian oil sands production pathways. Error bars show the standard deviations of 1,000 random runs of stochastic simulation of WTR GHG emissions, by pathway, due to variation in energy consumption and fugitive emissions associated with operational activities of a variety of oil sands projects and uncertainties in other GREET parameters. Parametric assumptions about the U.S. conventional crudes pathway were summarized elsewhere.34,35 M+B, M+SCO, IS+B, and IS+SCO represent surface mined bitumen, surface mined and upgraded SCO, bitumen from in situ production, and SCO from in situ production, respectively.

and thus produce CH4 and CO2 emissions.27,28 Estimates of these CH4 and CO2 emissions vary widely.28,29 We estimated that 0.9 and 1.1 g CO2 equivalents per megajoule (in lower heating value, CO2e/MJ) of bitumen and SCO, respectively, were produced from tailing ponds (see Supporting Information, Section 5). Crude bitumen batteries from in situ production can also produce fugitive CH4 emissions. We calculated an emission rate of 4.5 g CO2e per MJ of bitumen for the IS+B pathway from monthly venting statistics for crude bitumen batteries, whereas the IS+SCO pathway produces no fugitive CH4 emissions from crude bitumen batteries.20 Land Disturbance GHG Emissions. Oil sands land disturbance is significant, particularly for surface mining practices.30 Land disturbance GHG emissions were calculated on the basis of pre- and post oil sands land disturbances from 1985 to 2009 that were identified using satellite remote sensing observations and the associated soil carbon dynamics and were amortized over the production lifetime of 50 years for surface mining and 30 years for in situ projects.31 Lifetime productionweighted land disturbance GHG emissions of 1.87−1.90 g CO2e/MJ for surface mining and 0.56−0.89 g CO2e/MJ for in situ projects were estimated and incorporated into the GREET model31 and subsequently used in this analysis. Transportation of Oil Sands Products to U.S. Refineries. Transportation modes and distances determine the carbon footprint of transporting Canadian oil sands products to U.S. markets. Currently, dilbit and SCO are mostly transported from Alberta to PADD2, mainly by pipeline,32 for redistribution through the crude oil distribution infrastructure.33 We estimated a weighted-average distance of 1,708 miles by pipeline, according to the shares of oil sands received by each PADD region3 and the distances between oil sands fields and each of the PADD regions. Despite its higher costs relative to pipeline transportation, raw bitumen delivery by rail may be favorable in the future, as major oil transfer hubs are being built for rail transportation of oil sands.32 Transport of raw bitumen

credits for GHG emissions, we assumed that the surplus electricity displaced the GHG-intensive local Alberta electricity (see Supporting Information, Table S7). Diluents. Diluent is required to reduce the viscosity of bitumen for pipeline transportation to refineries in the form of dilbit. The diluent pool in Western Canada (WC) comes from four main sources. See Supporting Information, Section 8, for details. Other diluents include SCO. SCO can be used to produce a mixture of bitumen and SCO called synbit, but this is a more costly diluent and is less common so we considered only dilbit. A blend of the various sources of diluents, which has an average American Petroleum Institute (API) gravity of 66,22 is commonly used to blend bitumen to pipeline specification.23 The volume of diluent blended with the raw bitumen varies, depending on the project and the desired quality grade of the output. A 70:30 mixture of bitumen and diluent is generally required3,24 to produce dilbit, which has an API gravity of about 21.5. See Supporting Information, Section 8, for parametric assumptions for diluent production and transportation in WC and the U.S. Some refineries in Petroleum Administration for Defense District (PADD) 2 recover diluent in dilbit and return it to WC by pipeline for reuse.25 Such refineries process bitumen while others can process dilbit.26 In the bitumen pathway, we allocated all of the emissions associated with transportation of the diluent in dilbit to the bitumen, whereas the emissions associated with production of the diluent that was recovered for reuse were excluded, regardless of the diluent sources. In the dilbit pathway, the emissions associated with the production and transportation of the diluent in dilbit were included. We examined the emission effects of diluents by comparing gasoline and diesel production from pure bitumen and from dilbit. Fugitive CH4 and CO2 Emissions. Surface mining of oil sands followed by bitumen extraction using water and hydrocarbon solvent produces oil sands tailings that are deposited in “tailing ponds,” which are microbiologically active 8221

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology

Bitumen from in situ projects produced WTR GHG emissions of 25.5 g CO2e/MJ, on average, and ranging from 20.9 to 30.0 g CO2e/MJ (95% CI). Compared to the M+B pathway, higher emissions from in situ projects were due to CSS and SAGD operations that were approximately 3-fold and 2-fold, respectively, more energy- and emission-intensive than surface mining (see Supporting Information, Table S5). This intensity resulted in an additional 7.3 g CO2e/MJ from the extraction and separation processes. Fugitive CH4 emissions from crude bitumen batteries and diluent production and transportation added 4.5 and 2.5 g CO2e/MJ, respectively, whereas the electricity emission credit subtracted 0.3 g CO2e/ MJ. Land disturbance accounted for about 3% of the WTR GHG emissions. SCO from upgrading of bitumen from in situ projects produced the highest WTR GHG emissions, 33.6 g CO2e/MJ, on average, and ranging from 28.3 to 38.9 g CO2e/MJ at the 95% CI. Intensive fossil fuel consumption for bitumen extraction and upgrading (see Supporting Information, Figure S1) accounted for about 41% and 30% of the emissions, respectively. Hydrogen production for bitumen upgrading contributed another 10%, followed by pipeline transportation of SCO and flaring emissions, which accounted for another 9% and 6%, respectively. Land disturbance GHG emissions were relatively small for the IS+SCO pathway, accounting for 2% of the WTR emissions. Combining the four pathways on a production energyweighted average basis resulted in an averaged WTR GHG emission of 25.9 g CO2e/MJ for Canadian oil sands products received by U.S. refineries, compared to about 7.1 g CO2e/MJ for U.S. conventional crudes simulated by GREET. Overall emissions from oil sands extraction and bitumen separation were responsible for about 41% of WTR emissions, while bitumen upgrading caused 17%. These major sources were followed by pipeline transportation (15%), fugitive CH4 and CO2 emissions (10%), and on-site hydrogen production (7%). Land disturbance accounted for 5% of the emissions, whereas production and transportation of diluents contributed 5% of the emissions, as shown in Figure 1. Electricity credits offset about 2% of the WTR emissions. The error bars in Figure 1 show larger absolute variability for the M+SCO pathway compared to the M+B pathway, primarily because of the larger absolute variability of the energy consumption of a larger number of M+SCO projects (see Supporting Information, Figure S1). The relative variabilities of both pathways, however, were about the same. The IS+B pathway exhibited a slightly larger magnitude of variability relative to average emissions than the IS+SCO pathway, which represented only two projects. Therefore, the variability of the combined Canadian oil sands pathways was due mostly to the M+SCO and IS+B pathways that dominate Canadian oil sands production. Impacts of Canadian Oil Sands Products on U.S. Refinery GHG Emissions. Refinery GHG emissions are calculated on the basis of the amounts and types of purchased and intermediate process fuel used (see Supporting Information, Equation S1).26 The distinct API and sulfur contents of oil sands products relative to U.S. conventional crudes led to different overall refinery efficiencies of Canadian oil sands products, which determine the amount of purchased process fuel used. Further, we estimated the gasoline- and diesel-specific refinery efficiencies on the basis of the ratios of the productspecific purchased energy consumption to the overall purchased

by rail may require heated railcars, with attendant emissions impacts. GHG Emissions from U.S. Refineries Supplied with Canadian Oil Sands Products. Bitumen, SCO, and dilbit differ from conventional crude sources in their API gravity and sulfur content (see Supporting Information, Table S4). These are two key drivers for the U.S. overall and product-specific refining efficiencies, which directly impact GHG emission intensities of various refined products.26 We applied a formula that correlates overall refinery efficiency with crude API gravity and sulfur content26 (see Supporting Information, Section 9) to predict the U.S. overall refinery efficiency in response to the share and quality of Canadian oil sands products. The formula is applicable to this analysis, since the API gravity and sulfur contents of SCO fall in or near the normal ranges (16.2−39.1 for API gravity and 0.5−3.5 for sulfur content) of crude supplied to U.S. refineries.26 The formula is also applicable to dilbit refining analysis because it covered 12 large U.S. refineries accepting Western Canadian Select, and other dilbit assays produced from Peace River and/or Cold Lake, as major crude inputs.26 We assumed that bitumen was refined at the lowest efficiency the formula would predict, with the lowest API gravity of the range and a sulfur content of 4.8%. We estimated the U.S. refinery gasoline- and diesel-specific GHG emissions according to their energy efficiencies and process fuel shares (see Supporting Information, Equation S1).26



RESULTS AND DISCUSSION GHG Emissions of Oil Sands Recovery and Transportation. The so-called WTR GHG emissions resulting from bitumen and SCO production varied significantly among the four production pathways but were much higher than those of the U.S. conventional crudes (see Figure 1). Despite variability, bitumen from surface mining projects showed the lowest WTR GHG emissions among the four pathways, averaging 15.0 g CO2e/MJ of bitumen and ranging from 12.3 to 17.6 g CO2e/ MJ (95% confidence interval [CI]). For this pathway, oil sands extraction and bitumen separation were the most energyintensive WTR processes (see Supporting Information, Table S5), resulting in the most WTR emissions, followed by pipeline transportation of dilbit to U.S. refineries and diluent production and transportation. Of the WTR GHG emissions, land disturbance and fugitive emissions from tailing ponds accounted for about 13% and 6%, respectively, whereas surplus electricity from cogeneration offset about 7% of the WTR emissions. SCO from upgrading of bitumen from surface mining projects produced WTR GHG emissions of 26.7 g CO2e/MJ SCO, on average, and ranging from 21.9 to 31.6 g CO2e/MJ (95% CI). The higher GHG emissions associated with SCO primarily resulted from bitumen upgrading and from oil sands extraction followed by bitumen separation, which accounted for about 35% and 27% of the WTR emissions, respectively. The upgrading process consumes a significant amount of hydrogen (sourced from fossil natural gas and fuel gas via steam methane reforming; see Supporting Information, Figure S1) to convert heavy (low API) and sour (high sulfur content) bitumen into refinery-grade light (high API) and sweet (low sulfur content) SCO. Pipeline transportation of SCO to U.S. refineries and land disturbance contributed about 12% and 7%, respectively, while fugitive CH4 and CO2 from tailing ponds contributed about 4% of the WTR emissions. 8222

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology

Figure 2. WTW GHG emissions of gasoline and diesel sourced from Canadian oil sands products produced from four pathways, in comparison to those of fuels sourced from U.S. conventional crudes. Error bars show the standard deviations of 1,000 random runs of stochastic simulation of WTW GHG emissions, by pathway, due to variations in energy consumption and fugitive emissions associated with operational activities of a variety of oil sands projects and to uncertainties in other GREET parameters. M+B, M+SCO, IS+B, and IS+SCO represent surface mined bitumen, surface mined and upgraded SCO, bitumen from in situ production, and SCO from in situ production, respectively.

energy consumption, which were 1.18 and 0.92, respectively.26 Supporting Information, Table S11, shows the overall and gasoline- and diesel-specific refinery efficiencies, the purchased energy consumption, and the resultant refinery GHG emissions when bitumen, SCO, dilbit, and conventional crudes were refined. The refinery efficiencies for producing gasoline and diesel from bitumen were noticeably lower than those for the U.S. conventional crudes, causing higher refinery GHG emissions. In contrast, production efficiency for gasoline and diesel from SCO was higher than for U.S. conventional crudes, owing to the higher API gravity and the lower sulfur content of SCO (see Supporting Information, Table S4). This higher SCO efficiency resulted in lower process energy consumption and lower refinery GHG emissions. Overall, gasoline and diesel derived from combined oil sands products had 5% and 9%, respectively, higher refinery GHG emissions than their counterparts derived from U.S. conventional crude. WTW GHG Emissions of U.S. Petroleum Fuels Sourced from Canadian Oil Sands Products. BC and POC contributed less than 1% of the WTW emissions of the four oil sands pathways and thus were excluded from the WTW GHG emissions discussed below. WTW GHG emissions of oil sands-derived gasoline and diesel in the U.S. averaged 100−115 and 99−117 g CO2e/MJ, respectively, depending on the type and production technology of the oil sands products, compared to 92 and 91 g CO2e/MJ for the respective fuels produced from U.S. conventional crudes. If gasoline and diesel were produced with an energy-based weighted-average mix of the four oil sands pathways, they would have 18% and 21% higher WTW GHG emissions than those of the respective fuels sourced from U.S. conventional crudes. Given the variation and uncertainty

associated with the emissions, gasoline sourced from oil sands products, regardless of their types and production technologies, had higher WTW GHG emissions than gasoline produced from U.S. conventional crudes at the 95% CI. This is also true for oil sands-derived diesel. Figure 2 shows a breakdown of the WTW GHG emissions of gasoline and diesel sourced from oil sands products, in comparison to those derived from U.S. conventional crudes. GHG emissions from fuel combustion were the same for all four oil sands pathways and conventional crudes and dominated the WTW GHG emissions. Diesel combustion produces higher GHG emissions than gasoline combustion, owing to diesel’s ∼2% higher carbon content per MJ. For gasoline sourced from oil sands products, refining, which was a more energy- and emission-intensive process for gasoline production than diesel production,26 was the second largest source, except for using SCO as the crude feedstock. When SCO produced from mining and in situ projects was the feedstock, oil sands extraction, bitumen separation, and subsequent bitumen upgrading were the largest sources after fuel combustion for gasoline and diesel. In particular, refinery coke combustion accounted for about 31% and 37% of the emissions of SCO upgrading for mining and in situ projects, respectively. In addition, emissions associated with transportation of dilbit, fugitive emissions from tailing ponds and crude bitumen batteries, and diluent production and transportation totaled about 6−8% and 7−9%, respectively, of the WTW GHG emissions associated with gasoline and diesel sourced from bitumen, while on-site hydrogen production was an important emission source for gasoline and diesel sourced from SCO, accounting for about 3−4% of their WTW GHG 8223

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology

Table 1. Average Gasoline Carbon Intensities (GCI, in g CO2e/MJ, Excluding BC and POC Emissions) and the Average Relative Changes in GCI for Oil Sands Pathways Compared to Those of the Conventional Crude Pathway M+B WTW WTPa WTR crude recovery crude transpor- tation refining fuel combustion a

M+SCO

IS+B

IS+SCO

oil sands weighted average

conv crudesb

GCI

change (%)

GCI

change (%)

GCI

change (%)

GCI

change (%)

GCI

change (%)

GCI

99.5 26.3 11.7 7.8 3.9 14.0 73.2

8 38 96 76 155 12 0

108.8 35.5 22.8 20.1 2.7 12.1 73.2

18 86 284 354 80 −3 0

108.5 35.3 20.7 16.8 3.9 14.0 73.2

18 85 248 280 155 12 0

114.7 41.5 28.7 26.0 2.7 12.2 73.2

24 118 383 487 80 −3 0

108.5 35.3 21.6 18.2 3.3 13.1 73.2

18 85 263 312 119 5 0

92.3 19.1 5.9 4.4 1.5 12.5 73.2

From oil wells to fuel pumps. bU.S. conventional crudes.

tankers,36 resulted in about 119% higher transportation emissions. In contrast, the difference in refinery emissions between oil sands and conventional-crude pathways was small, and refining of SCO produced about 3% lower emissions than refining of conventional crudes. Therefore, it is important to compare the emissions of the oil sands and conventional-crudes pathways within the same system boundaries. Without significant reduction in the energy intensity of extraction, separation, and upgrading of oil sands, which were the primary drivers leading to about 18% (for gasoline) and 21% (for diesel) more WTW emissions than for U.S. conventional crudes (see Table 1 and Supporting Information, Table S12), higher WTW emissions for gasoline and diesel production in the U.S. are expected when oil sands products become a larger fraction of the U.S. fuel mix. It is noted that different oil sands pathways and the conventional crudes pathway may pose distinctive environmental issues beyond GHG emissions, for example, air pollution and water consumption. Holistic consideration of such issues toward a “weighted environmental impact”37 may result in different rankings of their total environmental burdens. Impacts of Diluents. Using diluents “diluted” the WTW emissions by 3−5% and 4−6%, respectively, for gasoline and diesel produced from dilbit compared to those derived from bitumen, primarily because of the ∼36% lower amount of the bitumen component delivered to U.S. refineries per MJ of dilbit compared to that per MJ of bitumen delivered. WTW GHG emissions of gasoline and diesel derived from dilbit were in the range of 97−103 and 96−104 g CO2e/MJ, respectively, compared to 100−109 and 99−110 g CO2e/MJ for gasoline and diesel derived from bitumen (see Supporting Information, Table S13). In addition, we found that a shift of the diluent pool to more of NGC from U.S. shale gas wells in 2020 would result in negligible differences in WTW emissions for dilbitderived gasoline and diesel (see Supporting Information, Table S13). Comparison to Previous Studies. This study differed from previous work in some or all of such aspects as system boundaries; oil sands project coverages; data sources and qualities of process fuel mixes associated with oil sands extraction, separation, and upgrading; methodologies of allocating refining emissions to gasoline and diesel; treatment of electricity cogeneration; and treatment of coke.1,2,8,10,12−15,38−42 A key methodological difference is the collection and use of multiyear detailed energy consumption data for 27 operating oil sands projects at monthly time increments. These publicly reported data are a useful addition to the study of oil sands emissions impacts and provide greater

emissions. Flaring emissions accounted for less than 2% of the WTW GHG emissions of fuels derived from SCO produced from in situ projects. Flaring was an even smaller emission source when bitumen and SCO upgraded from surface-mined bitumen were used as crude feedstocks. Land disturbance accounted for less than 2% of the WTW GHG emissions of finished fuels sourced from bitumen and SCO that were produced from surface mining projects. Land disturbanceinduced emissions were even lower for finished fuels sourced from bitumen and SCO produced from in situ projects. Emission credits from surplus electricity subtracted 0.1−1.0% of the WTW emissions of the oil sands pathways. Compared to oil sands-derived gasoline, oil sands-derived diesel had higher emissions associated with crude feedstock production but lower emissions from refining,26 resulting in slightly higher WTW GHG emissions for diesel compared to gasoline, which agreed with previous studies.1,2 We conducted a sensitivity analysis to look at the impacts of variations in individual life-cycle stages/activities on the WTW GHG emissions for each oil sands pathway. See Supporting Information, Section 10, for details. The results indicate that reduction of process fuel consumption that would result in higher energy efficiencies of oil sands production is one key to mitigating GHG emissions of oil sands-derived fuels. It is noted that refineries processing oil sands would not replace the crude with conventional light-sweet crude without modifying the refinery configuration. Therefore, the comparison between oil sands- and conventional crudes-derived refining products, as shown in Figure 2, was based on different refineries with different configurations that can only process a respective narrow range of crude API gravity and sulfur content. See Supporting Information, Section 13, for a comparison of oil sands pathways to those with other heavy crudes of similar quality. The magnitude of differences in emissions of different lifecycle stages varied significantly between the oil sands pathways and the U.S. conventional crudes pathway (see Table 1). The crude recovery stage of the oil sands pathways was about 3.1 times higher in carbon emission intensity than that of the GREET-simulated U.S. conventional crudes pathway, with the difference ranging from 76% to 487% depending on the oil sands pathway (see Table 1). This difference was attributable to the dramatically more-energy-intensive oil sands extraction, separation, and upgrading processes relative to conventional crude recovery (see Supporting Information, Table S5). Besides, electricity-powered pipeline transportation of oil sands products, which was more energy- and emission-intensive than transportation of conventional crudes primarily by ocean 8224

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology

a decrease by about 6% compared to the Jacobs study.2 Our efforts to carefully characterize the energy intensities associated with process operations of oil sands projects, which were identified as a fundamental cause of the variability seen in previous results,43 improved the validity of the carbon intensity results. Different methods of allocating emissions to refinery finished products can result in different product-specific emissions. For example, with the same crude feedstock, the Jacobs study allocated emissions about equally between gasoline and diesel,2 while the TIAX study allocated emissions mostly to the gasoline stream.1 The allocation schemes in both studies relied on proprietary refining data. The Petroleum Refinery Life-cycle Inventory Model, which was developed with refinery processlevel energy requirements from the literature and was compared with confidential information for evaluation, estimates the overall refinery GHG emissions as a function of the crude distillation curve, the crude API, sulfur, hydrogen content, and the carbon residue and allocates the emissions to gasoline and diesel on a fuel hydrogen content basis.16 We applied a refinery process-level energy allocation method to derive the gasoline- and diesel-specific refinery energy efficiencies and emissions.26 Building on more representative and up-to-date refinery process-level energy and material balance data, the modeling of refinery GHG emissions of gasoline and diesel produced in U.S. refineries was improved in the present study. An important driver behind the various WTW GHG results in previous studies was the assumption about coke use in bitumen upgrading.41 We treated coke as an upgrading byproduct. No energy or emission burdens of the upgrading process were allocated to coke. However, we calculated the GHG emissions from any coke combustion for steam and/or electricity generation. Furthermore, use of surplus coke for offsite power generation or other purposes beyond the upgrader and refinery gates was beyond the scope of this analysis. Given that much of the excess coke produced at on-site upgraders is currently stockpiled, this is a reasonable assumption. Compared to previous studies, we have presented WTW GHG emissions of oil sands-derived finished fuels with improved transparency and representativeness and with operating data that represent the average practices and technological advances of the oil sands industry since 2008. In estimating WTW emissions of finished fuels derived wholly from oil sands products, we differentiated primary extraction technology and product output of oil sands projects; evaluated the impact of the quality of oil sands products on refinery operational efficiencies and emission intensities; and considered land disturbance and BC and POC emissions. Our results contribute to the existing knowledge in this field and shed light on evaluating the incremental GHG emissions of U.S. transportation fuels that are anticipated to come from increasing use of oil sands products.

transparency than previous models based on proprietary data. Besides, monthly data since 2008 allowed for evaluation of the variability of energy intensities and the emission implications associated with oil sands production pathways. Despite such methodological and parametric differences, our finding that gasoline derived from oil sands products had WTW GHG emissions of 100−115 g CO2e/MJ was within the range of 100−120 g CO2e/MJ from earlier studies (see Supporting Information, Table S14, for a comparison).41 The much greater energy intensity of extraction, separation, and upgrading of oil sands relative to conventional-crude recovery remained the major reason for the higher WTW GHG emissions of oil sands pathways than conventional-crude pathways in this and previous studies. Moreover, we found that gasoline and diesel derived from surface-mined bitumen had the lowest WTW GHG emissions, while those sourced from SCO produced from in situ projects exhibited the highest emissions, despite a significant degree of variability due to operating conditions. This finding agrees with previous work.2,14 We estimated that WTW emissions of gasoline derived wholly from oil sands products ranged from 9% to 24%, on average, higher than those of gasoline produced from U.S. conventional crudes (see Table 1), compared to an increase of 14% according to a metaanalysis of 12 studies.40 Inclusion of land disturbance-induced emissions in our analysis explained about 2% of the difference. Since comprehensive comparisons of previous studies have been done, focusing on the study scopes and parametric assumptions,13,40−43 our discussion focused on several key emission processes and assumptions that were emphasized as key drivers of the variation and uncertainties in WTW results in previous studies. Both surface mining and in situ projects consume a variety of process fuels for steam and electricity production needed for extraction and separation of bitumen and purchase and/or produce hydrogen for bitumen upgrading. The data quality and assumptions related to the process fuels consumed can lead to varied emissions estimates for mining and upgrading processes.43 While we included diesel/SCO used in mining and the coke used in upgrading according to the monthly statistics on oil sands plant operations, earlier versions of GREET omitted these process fuels, the TIAX study omitted the diesel/SCO consumption for mining,1 and the Jacobs study2 did not specify the use of these process fuels in a large unknown portion of the process fuel mix. As a result, we arrived at higher estimates than earlier GREET studies and the TIAX study.1 For example, we estimated that gasoline produced from the IS+B pathway produced about 5 and 8 g CO2e/MJ higher emissions than earlier GREET studies and the TIAX study, respectively. However, for this fuel pathway, we estimated about 4 g CO2e/MJ lower emissions than the Jacobs study, which might be attributable to improved energy efficiencies of recent industrial practices,15 which were not reflected earlier. Natural gas was the major process fuel consumed to generate steam for in situ production (see Supporting Information, Figure S1). The difference in the consumption of natural gas and other process fuels for steam production reflected different assumptions about the steam/oil ratio among studies. With the process fuel mix obtained from recent operating data for major in situ projects, our estimated WTP GHG emissions for finished fuels derived from the IS+SCO pathway, which were about 42 g CO2e/MJ, were in line with most earlier results,1,2,8 while the WTP GHG emissions for finished fuels derived from the IS+B pathway, which is the major in situ production pathway, showed



ASSOCIATED CONTENT

S Supporting Information *

Imported Canadian oil sands in the U.S. crude oil mix; GREET methodology for well-to-wheels greenhouse gas emission calculation; System boundary of WTW analysis; Energy consumption of the four Canadian oil sands pathways; Fugitive CH4 and CO2 emissions from tailing ponds; Variability and uncertainty analysis; Alberta average electricity generation mix; Sources, production, and logistics of diluents; Regression 8225

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology

Cycle Greenhouse Gas Emissions; DOE/NETL-2009/1362; National Energy Technology Laboratory: USA, 2009. (13) Charpentier, A. D.; Bergerson, J. A.; MacLean, H. L. Understanding the Canadian oil sands industry’s greenhouse gas emissions. Environ. Res. Lett. 2009, 4, 014005. (14) Bergerson, J. A.; Kofoworola, O.; Charpentier, A. D.; Sleep, S.; MacLean, H. L. Life Cycle Greenhouse Gas Emissions of Current Oil Sands Technologies: Surface Mining and In Situ Applications. Environ. Sci. Technol. 2012, 46, 7865−7874. (15) Englander, J. G.; Bharadwaj, S.; Brandt, A. R. Historical trends in greenhouse gas emissions of the Alberta oil sands (1970−2010). Environ. Res. Lett. 2013, 8, 044036. (16) Abella, J. P.; Bergerson, J. A. Model to Investigate Energy and Greenhouse Gas Emissions Implications of Refining Petroleum: Impacts of Crude Quality and Refinery Configuration. Environ. Sci. Technol. 2012, 46, 13037−13047. (17) MathPro Inc. Effects of possible changes in crude oil slate on the U.S. refining sector’s CO2 emissions; International Council on Clean Transportation: USA, 2013. (18) Oil Sands Today: GHG emissions. http://www.oilsandstoday. ca/topics/ghgemissions/Pages/default.aspx (accessed June 2, 2015). (19) Climate Change and Emissions Management Amendment Act. http://www.qp.alberta.ca/documents/Acts/C16P7.pdf (accessed June 2, 2015). (20) Englander, J. G.; Brandt, A. R. Oil Sands Energy Intensity Analysis for GREET Model Update; Argonne National Laboratory: Lemont, Illinois, USA, 2014. (21) Cai, H.; Wang, M. Q. Consideration of Black Carbon and Primary Organic Carbon Emissions in Life-Cycle Analysis of Greenhouse Gas Emissions of Vehicle Systems and Fuels. Environ. Sci. Technol. 2014, 48 (20), 12445−12453. (22) Alberta Oil Sands Bitumen Valuation Methodology. http:// www.capp.ca/publications-and-statistics/publications/261786 (accessed June 2, 2015). (23) Quality Guidelines for Western Canadian Condensate. http:// www.coqa-inc.org/docs/default-source/meeting-presentations/ Segato0608.pdf (accessed June 2, 2015). (24) Fifty Shades Of Eh? − The Canadian Market For Condensate. Part 1. https://rbnenergy.com/fifty-shades-of-eh-the-canadian-marketfor-condensate (accessed June 2, 2015). (25) Lighter Hydrocarbons Dilute Heavy Crude Oil to Transport Long Distances. http://bloximages.chicago2.vip.townnews.com/ chippewa.com/content/tncms/assets/v3/editorial/f/82/f827a0f051d9-5ca6-abe1-d6903d6b7003/54b0409a2f71d.pdf.pdf (accessed June 2, 2015). (26) Elgowainy, A.; Han, J.; Cai, H.; Wang, M.; Forman, G. S.; DiVita, V. B. Energy Efficiency and Greenhouse Gas Emission Intensity of Petroleum Products at U.S. Refineries. Environ. Sci. Technol. 2014, 48, 7612−7624. (27) Siddique, T.; Penner, T.; Klassen, J.; Nesbø, C.; Foght, J. M. Microbial Communities Involved in Methane Production from Hydrocarbons in Oil Sands Tailings. Environ. Sci. Technol. 2012, 46, 9802−9810. (28) Small, C. C.; Cho, S.; Hashisho, Z.; Ulrich, A. C. Emissions from oil sands tailings ponds: Review of tailings pond parameters and emission estimates. J. Pet. Sci. Eng. 2015, 127 (0), 490−501. (29) Yeh, S.; Jordaan, S. M.; Brandt, A. R.; Turetsky, M. R.; Spatari, S.; Keith, D. W. Land Use Greenhouse Gas Emissions from Conventional Oil Production and Oil Sands. Environ. Sci. Technol. 2010, 44, 8766−8772. (30) Rooney, R. C.; Bayley, S. E.; Schindler, D. W. Oil sands mining and reclamation cause massive loss of peatland and stored carbon. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 4933−4937. (31) Yeh, S.; Zhao, A.; Hogan, S. Spatial and Temporal Analysis of Land Use Disturbance and Greenhouse Gas Emissions from Canadian Oil Sands Production; Argonne National Laboratory: Lemont, Illinois, USA, 2014.

formula to predict U.S. overall refinery efficiency; Sensitivity analysis; Comparison of average carbon intensities of diesel produced from oil sands products compared to that derived from U.S. conventional crudes; Impacts of the use of diluents on WTW GHG emissions of gasoline and diesel produced from dilbit; Comparison of WTW GHG emissions of Canadian oil sands fuel pathways to those of other heavy crudes pathways; and comparison of WTW GHG emissions of Canadian oil sands fuel pathways. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b01255.



AUTHOR INFORMATION

Corresponding Author

* Phone: 1-6302522892; Fax: 1-6302523443; Email: hcai@anl. gov. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research effort was supported by the Vehicle Technologies Office and the Bioenergy Technologies Office of the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy under Contract DE-AC02-06CH11357. We thank the anonymous reviewers of this paper for their helpful comments. The authors are solely responsible for the contents and results of the paper.



REFERENCES

(1) TIAX LLC; MathPro Inc. Comparison of North American and Imported Crude Oil Lifecycle GHG Emissions; Final Report, TIAX Case No. D5595; Alberta Energy Research Institute: Calgary, Alberta, Canada, 2009. (2) Keesom, W.; Unnasch, S.; Moretta, J. Life Cycle Assessment Comparison of North American and Imported Crudes; File No.: AERI 1747; Alberta Energy Research Institute: Calgary, Alberta, Canada, 2009. (3) 2014 Crude Oil Forecast, Markets and Transportation. http:// www.capp.ca/publications-and-statistics/publications/247759 (accessed June 2, 2015). (4) Canada’s Emissions Trends. http://ec.gc.ca/ges-ghg/default. asp?lang=En&n=E0533893-1 (accessed June 2, 2015). (5) Low Carbon Fuel Standard Program. http://www.arb.ca.gov/ fuels/lcfs/lcfs.htm (accessed June 2, 2015). (6) Fuel Quality. http://ec.europa.eu/clima/policies/transport/fuel/ documentation_en.htm (accessed June 2, 2015). (7) GREET Model: The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model Web site. https://greet.es.anl.gov (accessed June 2, 2015). (8) GHGenius Web site. http://www.ghgenius.ca (accessed June 2, 2015). (9) OPGEE: the Oil Production Greenhouse gas Emissions Estimator Web site. https://pangea.stanford.edu/researchgroups/eao/research/ opgee-oil-production-greenhouse-gas-emissions-estimator (accessed June 2, 2015). (10) Charpentier, A. D.; Kofoworola, O.; Bergerson, J. A.; MacLean, H. L. Life Cycle Greenhouse Gas Emissions of Current Oil Sands Technologies: GHOST Model Development and Illustrative Application. Environ. Sci. Technol. 2011, 45, 9393−9404. (11) McKellar, J. M.; Charpentier, A. D.; Bergerson, J. A.; MacLean, H. L. A life cycle greenhouse gas emissions perspective on liquid fuels from unconventional Canadian and US fossil sources. Int. J. Glob. Warm. 2009, 1, 160−178. (12) Gerdes, K.; Skone, T. J. An Evaluation of the Extraction, Transport and Refining of Imported Crude Oils and the Impact on Life 8226

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227

Article

Environmental Science & Technology (32) There is no transportation crisis for Alberta oil.http://www. edmontonsun.com/2014/09/26/hicks-there-is-no-transportationcrisis-for-alberta-oil (accessed June 2, 2015). (33) Annual Energy Outlook Web site. http://www.eia.gov/forecasts/ archive/aeo14 (accessed June 2, 2015). (34) Wang, M.; Han, J.; Dunn, J. B.; Cai, H.; Elgowainy, A. Well-towheels energy use and greenhouse gas emissions of ethanol from corn, sugarcane and cellulosic biomass for US use. Environ. Res. Lett. 2012, 7, 045905. (35) Updated Vented, Flaring, and Fugitive Greenhouse Gas Emissions for Crude Oil Production in the GREET Model. https:// greet.es.anl.gov/publication-emissions-crude-oil-2014 (accessed June 2, 2015). (36) Update to Transportation Parameters in GREET. https://greet. es.anl.gov/publication-tansportation-distribution-13 (accessed June 2, 2015). (37) Yang, Y.; Bae, J.; Kim, J.; Suh, S. Replacing Gasoline with Corn Ethanol Results in Significant Environmental Problem-Shifting. Environ. Sci. Technol. 2012, 46, 3671−3678. (38) Brandt, A. R.; Farrell, A. E. Scraping the bottom of the barrel: greenhouse gas emission consequences of a transition to low-quality and synthetic petroleum resources. Clim. Change 2007, 84, 241−263. (39) Gerdes, K.; Skone, T. Development of Baseline Data and Analysis of Life Cycle Greenhouse Gas Emissions of Petroleum-Based Fuels; DOE/ NETL-2009/1346; National Energy Technology Laboratory, USA, 2008. (40) Oil Sands, Greenhouse Gases, and US Oil Supply: Getting the Numbers Right − 2012 Update. http://www.api.org/~/media/Files/ %20Oil-and-Natural-Gas/Oil_Sands/CERA_Oil_Sands_GHGs_US_ Oil_Supply.pdf (accessed June 2, 2015). (41) Canadian Oil Sands: Life-Cycle Assessments of Greenhouse Gas Emissions. https://www.fas.org/sgp/crs/misc/R42537.pdf (accessed June 2, 2015). (42) GHG Emission Factors for High Carbon Intensity Crude Oils. http://docs.nrdc.org/energy/files/ene_10070101a.pdf (accessed June 2, 2015). (43) Brandt, A. R. Variability and Uncertainty in Life Cycle Assessment Models for Greenhouse Gas Emissions from Canadian Oil Sands Production. Environ. Sci. Technol. 2011, 46, 1253−1261.

8227

DOI: 10.1021/acs.est.5b01255 Environ. Sci. Technol. 2015, 49, 8219−8227