Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Modelling the greenhouse gas emissions intensity of plug-in electric vehicles in Canada using short-term
and long-term perspectives
by George Kamiya
B.Sc., University of British Columbia, 2008
Research Project Submitted in Partial Fulfillment of the
Requirements for the Degree of
Master of Resource Management (Planning)
in the
School of Resource and Environmental Management
Faculty of Environment
© George Kamiya 2015
SIMON FRASER UNIVERSITY Fall 2015
ii
Approval
Name: George Kamiya Degree: Master of Resource Management (Planning) Title: Modelling the greenhouse gas emissions intensity of
plug-in electric vehicles in Canada using short-term and long-term perspectives
Examining Committee: Chair: Maxwell Sykes MRM Candidate
Jonn Axsen Senior Supervisor Assistant Professor
Curran Crawford Supervisor Associate Professor University of Victoria
Date Defended/Approved: September 22, 2015
Ethics Statement
iii
Abstract
Plug-in electric vehicles (PEVs) have the potential to achieve deep greenhouse gas
(GHG) emission reductions. However, the magnitude of these reductions depends
largely on the source(s) of electricity, which can vary regionally and over time, making it
unclear how policymakers should regulate PEVs over the short and long-term. To
contribute to this discussion, I model the short and long-term operational (source-to-
wheels) greenhouse gas emissions intensity of PEVs in the Canadian provinces of
British Columbia, Alberta, and Ontario. I use empirical data on vehicle preferences,
driving patterns, and potential recharge access from the Canadian Plug-in Electric
Vehicle Survey (n=1754) to construct a temporally explicit model of PEV usage and
emissions over the short-term. Fleet-wide emissions intensity of PEVs varies
substantially between the three regions studied, with the greatest reduction potential in
British Columbia (78-99%), followed by Ontario (58-92%) and Alberta (34-41%) relative
to conventional (gasoline) vehicles. I then model the potential long-term dynamics of
technology, behaviour, and emissions with the CIMS energy-economy model under
three policy scenarios. Emissions intensity of electricity decreases by at least one-third
by 2050 even under current policies, with the deepest reductions in Alberta (64%).
Consequently, by 2050, fleet average PEV emissions are 23-40% (BC), 51-68%
(Alberta), and 25-40% (Ontario) below 2015 levels. Despite the large range of emissions
intensities between regions and over time, PEVs offer substantial GHG emissions
benefits compared to conventional vehicles. Therefore, policy makers should look to
design policies that concurrently promote vehicle electrification and decarbonisation of
the electricity supply to help achieve long-term mitigation targets.
Keywords: greenhouse gas emissions; passenger vehicles; plug-in electric vehicles; source-to-wheel; Canada; consumer behaviour
iv
Acknowledgements
First and foremost, I thank Jonn Axsen for his enthusiasm, patience, and wisdom, and
Curran Crawford for providing unique and insightful perspectives throughout the project.
I thank key individuals and organisations that made this project possible:
• Joe Bailey, Paulus Mau, Steven Groves, Jeff Rambharack, and Grace Lau for
their assistance in designing, programming, and implementing the survey
instrument; survey respondents and testers for their time and input.
• Members of EMRG and IESVic, in particular, Maxi Kniewasser, Stephen Healey,
Suzanne Goldberg, and Sahand Behbhoodi for their help with CIMS and
emissions modelling.
• BC Hydro, AESO, and IESO for providing electricity generation data; NRCan for
providing access to their fuel economy database.
• BC Hydro, NRCan, PICS, and SSHRC for providing funding for the project. SFU,
CIEEDAC, and my previous employer, Health Canada, for financial support.
Thank you to Mark Jaccard, John Nyboer and Noory Meghji for their support throughout
my time in EMRG. Thanks to the great friends I made in REM. Thank you to my
colleagues at the IEA for their ongoing support in completing this project.
Finally, I thank my parents Fumiko and Thomas for their commitment to my education,
and my sister and Tony for their support.
v
Table of Contents
Approval ............................................................................................................................. iiAbstract ............................................................................................................................. iiiAcknowledgements ........................................................................................................... ivTable of Contents ............................................................................................................... vList of Tables .................................................................................................................... viiList of Figures ................................................................................................................. viiiList of Acronyms ................................................................................................................ x
Chapter 1. Introduction ............................................................................................... 11.1. Context ..................................................................................................................... 11.2. Modeling greenhouse gas emissions from plug-in electric vehicles ........................ 31.3. Approach and research objectives ......................................................................... 12
Chapter 2. Methodology ............................................................................................ 142.1. Consumer data collection ...................................................................................... 142.2. Sample ................................................................................................................... 202.3. Modeling vehicle use and recharge ....................................................................... 242.4. Short-run model of GHG emissions ....................................................................... 262.5. CIMS as a long-run energy-economy model ......................................................... 322.6. Informing CIMS with short-run modeling outputs ................................................... 39
Chapter 3. Results ..................................................................................................... 453.1. Driving activity and recharge potential ................................................................... 453.2. PEV electricity recharge profiles ............................................................................ 493.3. Short-run greenhouse gas impacts of PEVs .......................................................... 533.4. Long-run greenhouse gas impacts of PEVs .......................................................... 57
3.4.1. Emissions intensity of electricity ................................................................ 573.4.2. Fleet composition ...................................................................................... 593.4.3. Emissions intensity of the PEV fleet .......................................................... 613.4.4. Emissions intensity of PHEVs and BEVs .................................................. 64
Chapter 4. Discussion ............................................................................................... 674.1. Short-term emissions intensity of PEVs ................................................................. 674.2. Long-term emissions intensity of PEVs ................................................................. 694.3. Policy implications .................................................................................................. 714.4. Limitations .............................................................................................................. 71
vi
Chapter 5. Conclusions and future research .......................................................... 74
References .....................................................................................................................76Appendix A: Assumptions ............................................................................................... 83
vii
List of Tables
Table 1: Summary of previous leading studies estimating GHG intensity of PEVs ......................................................................................................... 7
Table 2: Sample demographics and representativeness ............................................ 22Table 3: Electricity consumption (kWh/km) by vehicle class and usable battery
capacity (kWh) for PEV designs and vehicle classes (adapted from Axsen & Kurani, 2013) .................................................................... 26
Table 4: Lifecycle GHG emissions intensity of electricity, by generation type (from Moomaw et al., 2014) .................................................................... 27
Table 5: Comparing emission factors over time and method ....................................... 29Table 6: Well-to-wheels gasoline emission factors (g CO2e/L) .................................... 29Table 7: Available generation types in CIMS by province and load ............................. 35Table 8: Allocation of hours and generation types by low, medium, and high
load hours ................................................................................................ 36Table 9: Modeled scenarios in CIMS ........................................................................... 38Table 10: Key CIMS parameters and assumptions and updates based on
CPEVS data and recent literature ........................................................... 40Table 11: Regional differences in passenger vehicle attributes by vehicle class ........ 41Table 12: CIMS passenger vehicle motors parameters and assumptions:
capital cost, declining intangible costs, and fixed intangible costs .......... 42Table 13: CIMS passenger vehicle motor technical assumptions: gasoline and
electricity consumption and electric utility factor. .................................... 43Table 14: Comparing technical parameters of available PEV models with CIMS
assumptions. ........................................................................................... 44Table 15: Average daily electricity demand, electric utility factor, and morning
and evening peak information, by scenario ............................................. 50Table 16: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of
conventional (CV), hybrid-electric (HEV), and plug-in electric (PEV) vehicles ......................................................................................... 56
viii
List of Figures
Figure 1: Hourly electricity generation mix in Alberta (9 September 2012) .................. 11Figure 2: Overview of survey (from Axsen et al., 2015). .............................................. 16Figure 3: Sample driving diary from CPEVS: Part 2 .................................................... 17Figure 4: Sample vehicle design game from CPEVS: Part 3 ....................................... 19Figure 5: Geographic distribution of sample (n=1,754) ................................................ 21Figure 6: Respondent vehicle design in the lower price design game.
Respondents who designed a PHEV and BEV were included in the potential “Early Mainstream” sample shown in Figure 7. .................. 23
Figure 7: Share of PEV vehicle designs of the potential “Early Mainstream” sample ..................................................................................................... 24
Figure 8: Sample estimation of marginal emissions factor .......................................... 28Figure 9: Marginal hourly and average hourly emissions factor using recent
generation data ....................................................................................... 31Figure 10: Motors available in “Passenger Vehicle Motors” in the CIMS
Transportation sector .............................................................................. 33Figure 11: Proportion of “Early Mainstream” respondents driving by time of day,
averaged across weekdays and weekends in British Columbia (n=201 respondents, 603 diary days), Alberta (n=102 respondents, 306 diary days), and Ontario (n=194 respondents, 582 diary days). ....................................................................................... 46
Figure 12: Proportion of “Early Mainstream” respondents driving on weekdays (460 diary days) vs. weekends (143 diary days) by time of day in British Columbia. ..................................................................................... 46
Figure 13: Existing home recharge access (i.e. outlets within 25 feet of the vehicle’s typical parking location) from Part 2: Home Recharge Assessment ............................................................................................. 47
Figure 14: Driving activity and recharge access by time of day, British Columbia potential Early Mainstream respondents from Part 2: Driving Diary (n=201, 603 diary days) .......................................................................... 48
Figure 15: Proportion of existing workplace recharge access reported by potential Early Mainstream respondents in BC (n=201), Alberta (n=102), and Ontario (n=194). ................................................................. 48
Figure 16: Electricity demand profiles under three scenarios in British Columbia (n=201; 603 diary days) .......................................................... 51
Figure 17: Electricity demand profiles in British Columbia: weekdays (460 diary days) vs. weekends (143 diary days) for Scenario 1: User informed .................................................................................................. 52
ix
Figure 18: Electricity demand profiles in British Columbia: weekdays (460 diary days) vs. weekends (143 diary days) for Scenario 2: User Vehicle + Enhanced Workplace L2 ...................................................................... 52
Figure 19: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of conventional, hybrid electric, and plug-in electric vehicles in British Columbia, Alberta, and Ontario, using hourly marginal emissions factors for electricity, (including trade). .................................................... 54
Figure 20: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of conventional, hybrid electric, and plug-in electric vehicles in British Columbia, Alberta, and Ontario using average hourly emissions factors for electricity (including trade). .................................... 55
Figure 21: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of conventional, hybrid electric, and plug-in electric vehicles in British Columbia, Alberta, and Ontario using average annual emissions factors for electricity (excluding trade). ................................... 55
Figure 22: GHG emissions intensity of electricity (g CO2e/kWh) in Alberta and Ontario, by policy scenario ...................................................................... 58
Figure 23: Share of PEVs by type and range (left axis) and PEV and hybrid shares of the total vehicle stock (right axis), 2015-2050, BC Reference (Current Policies) Scenario .................................................... 59
Figure 24: Levelised costs ($/km) of medium efficiency gasoline, hybrid, PHEV-32, and BEV-240 under the Current Policies Scenario in British Columbia. The initially depressed costs of PEVs are due to the provincial incentive that is removed in 2020. ........................................... 60
Figure 25: Comparison of PHEV to BEV ratio between the three scenarios in British Columbia ...................................................................................... 61
Figure 26: Fleet-average short-run and long-run GHG emissions intensity of vehicles in BC. Short-run PEV results are shown for Scenario 1 only. ......................................................................................................... 62
Figure 27: Fleet-average short-run and long-run GHG emissions intensity of vehicles in Alberta. Short-run PEV results are shown for Scenario 1 only. ...................................................................................................... 63
Figure 28: Fleet-average short-run and long-run GHG emissions intensity of vehicles in Ontario. Short-run PEV results are shown for Scenario 1 only. ...................................................................................................... 64
Figure 29: Emissions intensity of PHEV-32 and BEV under the Reference scenario ................................................................................................... 65
Figure 30: Emissions intensity of PHEV-32 and BEV in Alberta under the Reference (BAU) and Carbon Tax scenarios .......................................... 66
x
List of Acronyms
BAU Business as usual
BC British Columbia
BEV Battery electric vehicle (see also: EV)
CCS Carbon capture and storage
CO2e Carbon dioxide equivalent
CPEVS Canadian Plug-in Electric Vehicle Survey
EV Electric vehicle (see also: BEV)
GHG Greenhouse gas
HEV Hybrid electric vehicle
LCA Lifecycle analysis
NHTS National Household Travel Survey
PEV Plug-in electric vehicle
PHEV Plug-in hybrid electric vehicle
UF Utility factor
WtW Well-to-wheels
1
Chapter 1. Introduction
1.1. Context
The road transportation and power sectors are two of the largest sources of
greenhouse gas (GHG) emissions in Canada1, accounting for 31% of the total
(Environment Canada, 2015a). While emissions from electricity generation have fallen
by about one-third over the past 10 years, emissions from road transportation continue
to increase. Though regulations to improve fuel efficiency have been effective to some
degree, growth in vehicle travel has outpaced these efforts. In this context, continued
decarbonisation of electricity and transportation electrification through plug-in electric
vehicles may play a critical role in achieving long-term climate targets (IEA, 2015;
McCollum et al., 2013; Williams et al., 2012).
This project focuses on the electrification of passenger vehicles. Plug-in electric
vehicles (PEVs) are motor vehicles with batteries that can be recharged by plugging into
the electricity grid, and include plug-in hybrid electric vehicles (PHEVs), which can run
on electricity or gasoline, and battery-electric vehicles (BEVs), which operate solely on
electricity. While PEVs are touted as “clean energy technologies” (International Energy
Agency, 2015a), the true emissions reduction potential of PEVs depends largely on the
electricity supply mix (Elgowainy et al., 2009). In Canada, PEVs are a particularly
promising technological solution to reduce emissions from light-duty vehicles, given the
country has one of the cleanest electricity systems in the world (151 g CO2e/kWh vs.
world average of 533 g CO2e/kWh) (International Energy Agency, 2014a). However,
several Canadian provinces, notably Alberta, Saskatchewan, and Nova Scotia, remain
heavily coal and gas-dependent, with emission intensities (2013) of 820 g CO2e/kWh, 1 By IPCC sector, the leading subsectors are: “Road Transportation”: 137 Mt CO2e; “Mining and Upstream
Oil and Gas Production”: 94 Mt CO2e; and “Public Electricity and Heat Production”: 88 Mt CO2e.
2
770 g CO2e/kWh, and 740 g CO2e/kWh respectively (Environment Canada, 2015).
Some studies estimate that PEVs in such coal-based regions may not actually reduce
emissions compared to conventional and hybrid-electric vehicles (e.g. Jochem et al.,
2015; Ma et al., 2012; Raykin et al., 2012; Ribberink & Entchev, 2013). With regionally
diverse power grids, complex electricity trade networks, numerous and potentially
conflicting environmental and health objectives (e.g. climate, air quality, hazardous
waste), policy makers face challenges in designing “optimal” PEV policy in conjunction
with planning of electricity grids.
Even with a singular objective of climate change mitigation, policy makers face
many difficult issues when designing and implementing policy. In particular, the
emissions intensity of electricity is likely to change over time. In Ontario, for example, the
emissions intensity of electricity has fallen by two thirds over the past 15 years (from
320 gCO2e/kWh to ~100 gCO2e/kWh), and the grid is expected to continue to
decarbonise over the next 10 years, with about half of the installed generation capacity
expected to come from renewable sources (Ontario Ministry of Energy, 2013). Similarly,
the coal-based grid of Alberta is expected to change to a gas-based grid with increased
wind capacity. The Government of Alberta (2015) recently announced plans to phase out
all coal-fired power generation in the province by 2030.
This potential for significant dynamics in electricity systems raises an important
question for policy makers: should emissions from vehicles, in particular, PEVs, be
evaluated based on the current electricity grid, or from a longer-term perspective as
grids decarbonise? In this study, I model the emissions intensity of PEVs from short-term
and long-term perspectives in three Canadian regions to explore these questions and
provide insights for policy makers.
3
1.2. Modeling greenhouse gas emissions from plug-in electric vehicles
There are a wide variety of studies that estimate the greenhouse gas emission
impacts of PEVs compared to conventional vehicles. Some previous GHG impact
studies of PEVs use a lifecycle analysis (LCA) approach, and consider emissions from
vehicle production and disposal, as well as those from fuel production and distribution
(e.g. Garcia et al., 2015; Ma et al., 2012; Rangaraju et al., 2015; Samaras & Meisterling,
2008; Silva et al., 2009). Most studies focus exclusively on emissions from the “use”
phase, following a “source-to-wheels” approach, i.e. fuel production and combustion and
upstream electricity production (e.g. Axsen et al., 2011; Elgowainy et al., 2009; Nguyen
et al., 2013; Raykin et al., 2012; Shen et al., 2012). Because the use phase accounts for
the majority of lifecycle emissions for passenger vehicles (e.g. 60-90% in EVs)
(Hawkins, Gausen, & Strømman, 2012; Nealer & Hendrickson, 2015), my own study
also takes a source-to-wheels approach.
Quantifying the greenhouse gas emissions impacts of plug-in electric vehicles is
difficult because these impacts depend on numerous technical and behavioural factors
that can differ between regions and change over time, such as the grid mix, the number
of vehicles and their characteristics (including PEV type), and how vehicles are driven
and recharged by consumers. As a result, estimates of PEV GHG impacts can range
widely, from as little as 3 g/km for a BEV in a hydro-based region (Ribberink & Entchev,
2013) to as much as 280 g/km for PHEV in a coal-based region driven in heavy traffic
(Raykin et al., 2012). Methodological differences in how GHG emissions are attributed
can also affect emissions estimates. Previous studies employ a range of assumptions
and methodologies; Table 1 compares these approaches and results of selected studies.
In summary, the studies differ in a number of key aspects: geographic location, time
frame, carbon intensity of electricity, emissions allocation method, vehicles, behavioural
assumptions and data sources.
Geographic location and timeframe: Previous studies have estimated
emissions impacts at national and subnational levels in the United States (e.g.
Elgowainy et al., 2009), Canada (e.g. Ribberink & Entchev, 2013), China (Shen et al.,
4
2012), Germany (Jochem et al., 2015), and Portugal (Rolim et al., 2012), among others.
A few studies include inter-regional comparisons at the national and subnational levels:
Doucette & McCulloch (2011) compare France, US, and China; Ma et al. (2012)
compare impacts in the UK and California; Plötz et al. (2015) compare impacts in
Canada, US, and Germany. The majority of the studies reviewed consider present-day
or near-term electricity grids (e.g. Doucette & McCulloch, 2011; Moorhouse &
Laufenberg, 2010; Samaras & Meisterling, 2008); only a few consider longer-term
changes to the power sector (e.g. Axsen et al., 2011; EPRI, 2007; Nguyen et al., 2013).
Axsen et al. (2011) use a dispatch model (LEDGE-CA) to consider power plant
retirements and capacity expansions to estimate PEV emissions based on the modeled
grid in 2020. EPRI (2007) uses the US National Energy Modeling System (NEMS) to
calculate energy supply and demand to 2050, and then use price and electricity load
outputs to model the evolution of the electric grid using their National Electric System
Simulation Integrated Evaluator (NESSIE). Nguyen et al. (2013) derive electricity grid
mix assumptions for 2035 from the US Energy Information Administration’s Annual
Energy Outlook 2012, which is also modeled using NEMS.
Emissions intensity of electricity: The GHG emissions impacts of PEVs are
arguably most dependent on the emissions intensity of the electricity supply
(gCO2/kWh), which is closely associated with a study’s geographic location and
timeframe (discussed above). Previous studies have investigated PEV impacts in many
regions and timeframes, and as a result, have covered everything from near-zero carbon
power systems (e.g. British Columbia, 10 gCO2/kWh (Moorhouse & Laufenberg, 2010) to
coal-dependent regions in Canada and the United States with electricity emissions
intensity exceeding 1 000 gCO2/kWh (Onat, Kucukvar, & Tatari, 2015; Ribberink &
Entchev, 2013).
Electricity emissions allocation methodology: Previous GHG impact studies
use a range of approaches to allocate and attribute electricity emissions (Yang, 2013).
First, emissions can be allocated on an average or marginal basis. An average approach
treats the electricity demand from PEVs to be the same as other loads, while a marginal
approach considers the load from PEVs to be additional to the current grid, demanding
specific power plants to generate to meet this incremental demand. Therefore, GHG
5
impacts are calculated based on the emissions intensity of the marginal generator (the
last generator to “turn on” to meet the new demand presented by the addition of PEVs),
which are often higher than the average generation mix that typically includes base load
nuclear, hydro, and other renewables. Both perspectives can provide useful insights.
The majority of previous studies (e.g. Moorhouse & Laufenberg, 2010; Ribberink &
Entchev, 2013; Samaras & Meisterling, 2008) use an average approach, while a few
allocate emissions on a marginal basis (e.g. Axsen et al., 2011; EPRI, 2007; Elgowainy
et al., 2009; Ma et al., 2012). Second, emissions intensity of electricity can be averaged
across a long period (e.g. annual) or allocated based on the time of day of demand (e.g.
hourly). This distinction is important because PEV electricity demand profiles are highly
variable across a 24-hour period. While many studies employ an annual average, some
studies will use more detailed data to represent specific generators for each hour of the
year (e.g. Axsen et al., 2011).
Vehicles: Earlier studies (pre-2011) focused predominantly on PHEVs, given the
higher likelihood of their near-term adoption. Some studies, such as Elgowainy et al.
(2009) and Samaras & Meisterling (2008), consider multiple PHEV models with different
electric ranges (e.g. PHEV-16 and PHEV-64, indicating 16 and 64 km of all-electric
range respectively) and consequently, different recharging patterns and electric utility
factors. Electric utility factor (UF) refers to the proportion of PHEV driving that is powered
by electricity rather than gasoline. More recent studies (e.g. Nguyen et al., 2013; Onat,
Kucukvar, & Tatari, 2015; EPRI, 2015) have shifted in focus to estimate emission
impacts of both PHEVs and BEVs. Samaras & Meisterling (2008) compare PEV
emissions to a baseline conventional vehicle, while studies such as Axsen et al. (2011)
and Nguyen et al. (2013) use future fleet-wide fuel economy standards as a baseline.
Such differences in baseline vehicle assumptions, including the emissions factor of fuels
(e.g. oil sands vs. conventional crude) are potentially large sources of variability in
estimating the emission reduction potential of PEVs.
Behavioural assumptions and data sources: GHG impacts of PEVs depend
on a number of behavioural and technical parameters, such as vehicle design
(discussed above) and driving and recharging patterns, which are important predictors of
PEV electricity use and timing of demand. The sources of such data can have effects on
6
quality and therefore the behavioural realism of the model. To simulate driving and
potential recharging patterns, numerous US-based studies use data from the National
Household Travel Survey (NHTS) (e.g. Samaras & Meisterling, 2008), while Axsen et al.
(2011) use data from driving diary of consumers, matched to their vehicle preferences.
Other studies simply simulate standardised driving cycles used to evaluate fuel economy
ratings (e.g. Doucette & McCulloch, 2011; Ma et al., 2012). Recharge profiles may also
be simulated based on likely charging scenarios in the presence of time-of-use pricing
(e.g. late night to early morning).
7
Table 1: Summary of previous leading studies estimating GHG intensity of PEVs
8
9
10
As Table 1 shows, few studies have estimated the GHG emissions of PEVs in
Canada, and the results range substantially by region:
• In British Columbia, the Pembina Institute found that BEVs (11 g/km) presently reduce source-to-wheels emissions by 96% compared to an average gasoline vehicle2, while a PHEV with 50 km of electric range (PHEV-50) (111 g/km) can reduce emissions by 63% (Moorhouse & Laufenberg, 2010). These results are similar to Ribberink & Entchev's (2013) estimated reductions in Quebec, which has a similar hydro-dominated electricity system.
• In Alberta, where electricity generation is predominantly coal-based, a study by Natural Resources Canada (NRCan) estimates that BEVs could have 4% higher emissions than conventional vehicles in 2025, while PHEVs could have 13% lower emissions (Ribberink & Entchev, 2013).
• In Ontario, a peer-reviewed study by Raykin et al. (2012) estimated that a PHEV-32 (60-170 g/km) can reduce emissions by 24-80%, depending on driving conditions. PEV advocacy group Plug’n Drive (2015) estimated that with the current Ontario grid, PHEVs (65 g/km) can reduce emissions by 74% compared to conventional vehicles, while BEVs (12 g/km) may achieve a 95% reduction. These estimates are more optimistic than Ribberink & Entchev (2013) who estimated a 49% reduction for PHEVs and a 82% reduction for BEVs in 2025.
While these estimates illustrate the possible range of emission impacts in
Canada, the analyses rely on four main methodological limitations. First, all four studies
use average emission factors for electricity, which does not consider the additional PEV
load as incremental, demanding power plants on the margin to supply the required
electricity. Depending on the region, the marginal grid emissions intensity might be
higher or lower than the average intensity. In Alberta for example, marginal emissions
will more likely be generated by natural gas plants (~450 g CO2e/kWh), whereas an
estimate of carbon intensity across the entire grid would be higher due to a reliance on
coal for the base load (~1000 g CO2e/kWh). Raykin et al. (2012) model emission
impacts from 100% hydro, natural gas, and coal to simulate such conditions where these
generation types would be on the margin.
2 Baseline fuel efficiency of 10.2L/100km
11
Second, these Canadian studies use annual (average) emission factors, which
do not consider variations in electricity emissions intensity over the course of a given
day. Varying demand and availability of intermittent renewables at different times of day
can result in high variability in the generation mix, and in particular, in the marginal
supply. For example, Figure 1 shows the variability in generation mix and trade flows in
Alberta over the course of a sample day in September 2012.
Figure 1: Hourly electricity generation mix in Alberta (9 September 2012)
Third, by using retrospective or present day emissions factors for electricity,
longer-term (prospective) changes to the generation mix over time may not be captured.
While retrospective data offer a higher degree of certainty, accuracy, and detail, it only
provides a limited view of the emissions intensity of electricity and PEVs of the future. As
greater numbers of PEVs are adopted, longer-term emissions intensity of electricity
becomes more important in shaping policy decisions.
Finally, on the demand side (vehicle use), these studies use simplified
behavioural assumptions of vehicle use, such as one electric utility factor for PHEVs.
While these assumptions may be reasonably accurate, the use of empirical data could
improve the behavioural realism of these studies, in particular where driving behaviours
vary among vehicle types and regions.
-800
-600
-400
-200
0
200
400
600
800
0
1000
2000
3000
4000
5000
6000
7000
8000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Tr
ade (
MWh)
Gene
ratio
n (M
Wh)
Time of Day (hour ending)
Net Trade
Other
Wind
Hydro
Gas
Coal
Exports (right axis)
Imports (right axis)
12
1.3. Approach and research objectives
In this study, I attempt to address the limitations above by modelling two
complementary perspectives: a detailed analysis over the short-term informed by
empirical data and multiple emissions allocation approaches (average annual and hourly
marginal), and a broader, dynamic analysis over the longer term to consider
technological dynamics. Each approach offers its own advantages and limitations.
The short-term (static) perspective provides a rich and detailed perspective on
PEVs using empirical data on PEV preferences and usage collected from potential PEV
buyers, and present-day electrical grid information. Data availability and certainty allows
me to consider electricity demand and generation on an hourly basis, and also model
GHG impacts on a “marginal” basis (based on the electricity generation plants that would
be used to recharge PEVs at the time of demand. The marginal approach is used under
the benchmark Greenhouse Gas Protocol when quantifying emission reductions from
grid-connected projects (WRI, 2005). However, this short-term perspective does not
represent how the electrical system may change over time, which is an important
limitation.
The long-term (dynamic) perspective considers changes to the electricity grid
mix and vehicle fleet over the longer term. This is an important perspective because the
emissions intensity of electricity supply will change in the future as new, cleaner
generation comes online and older power plants are retired. This dynamic perspective is
particularly important when considering the impact of emerging technologies such as
PEVs due to their limited uptake over the near-term, but potentially substantial
contributions to emission reductions over the medium to long-term (2035-2050).
However, this perspective inevitably includes more uncertainty about the future world
(behaviour and technology), and thus more assumptions are required. For example, it
becomes increasingly difficult to maintain the high temporal resolution of the short-term
model (e.g. hourly PEV demand and electricity generation mixes, as well as an
understanding of marginal generators), which is a limitation of this approach.
13
My research objectives are to:
1. Develop a behaviourally realistic model of the short-term greenhouse gas (GHG) emissions impacts of plug-in electric vehicles using data from the Canadian Plug-in Electric Vehicle Study (Axsen et al., 2015);
2. Use insights from the short-term model to model long-term technological dynamics of the transport and power sectors and resulting GHG emissions intensity of plug-in electric vehicles (PEVs) to 2050; and
3. Compare GHG intensity results from three different Canadian regions (British Columbia, Alberta, Ontario), and demonstrate the effects of climate policy on PEV GHG intensity.
The remainder of the paper is organised as follows. Chapter 2 describes the
project methodology, including data collection, the modelling approaches, and
assumptions. Chapter 3 summarises the results. Chapter 4 compares key results to
literature, and identifies some of the limitations of this study. Chapter 5 concludes with a
discussion of policy implications and suggests areas for future research.
14
Chapter 2. Methodology
This chapter describes the methods I used to collect, analyse and model the data to
address my research objectives. First, I describe the survey instrument used to collect
consumer data to improve the behavioural realism of my analysis. Next, I discuss how I
modeled this consumer data to construct PEV recharge profiles. The remaining sections
provide detailed descriptions of the approaches used to model the short and long-term
GHG emissions intensity of PEVs, including key model assumptions.
2.1. Consumer data collection
I collected consumer data using the 2013 Canadian Plug-in Electric Vehicle
Survey (CPEVS)3, a three-part, mixed-mode survey of new vehicle buyers in Canada.
Additional details of the survey instrument and methodology are available in Axsen et al.
(2015). I define “new vehicle buyers” as households who had purchased a new vehicle
in the past five years and use a vehicle regularly (referred to the “Mainstream” market in
CPEVS). This target group contains the potential “Early Mainstream” PEV buyer market,
whose acceptance of the technology will be crucial to broader market acceptance and
adoption (Axsen et al., 2015). A market research company (Sentis) recruited a
representative sample of new vehicle buying households, deployed the survey, and
provided incentives to respondents for survey completion. Data were collected between
April and October 2013. Note that other elements of CPEVS also collected interview
data from Mainstream respondents, as well as survey and interview data from a sample
of actual PEV owners (or “Pioneers”) – however, I presently only draw from Mainstream
respondents’ survey data.
3 Survey documents are available for download at: http://www.rem.sfu.ca/people/faculty/jaxsen/cpevs/.
15
The survey included three parts, delivered over the web and by mail (depicted in
Figure 1):
• Part 1: a 30-minute online survey to collect background information, including sections on the respondent’s vehicle fleet, home electricity conditions, and general lifestyle.
• Part 2: a two-component survey delivered by mail. First, a Home Recharge Assessment evaluated respondents’ potential to install a charger at home. Respondents then completed a three-day driving diary, which I use to anticipate potential PEV driving and recharging patterns. Respondents were also provided with a short introductory guide to vehicle technologies and green electricity in preparation for Part 3.
• Part 3: a 35-minute online survey to elicit consumer vehicle and recharging preferences through PEV discrete choice experiments and design exercises adapted from Axsen & Kurani (2009).
16
Figure 2: Overview of survey (from Axsen et al., 2015).
My project primarily uses data from Parts 2 and 3, particularly the respondent
data regarding home recharge potential, driving and parking activity (simulating potential
PEV use and recharge) and vehicle design preferences. I collected respondents’ driving
and parking activities using a three-day driving diary (Figure 3) stratified across the week
(respondents were randomly assigned to start their diary across all days of the week to
ensure coverage of weekdays and weekends). Using the conventional (gasoline
powered) vehicle they intended to replace next, respondents recorded trip time,
distance, and purpose, as well as information about each parking location (type,
potential recharge access). I used a three-day diary to overcome the inherent limitation
of one-day driving diaries (e.g. the commonly used US National Household Travel
Survey) that cannot represent driving patterns across multiple, sequential driving days
17
(Davies and Kurani, 2013). Design of this diary was informed by past travel surveys and
literature (Axhausen, 1995; Axsen & Kurani, 2010, 2013; Butcher & Eldridge, 1990;
Santos, McGuckin, Nakamoto, Gray, & Liss, 2011; Stopher, 2008), and was pre-tested
to assure that respondents would understand the instructions.
Figure 3: Sample driving diary from CPEVS: Part 2
18
The survey also elicited vehicle design preferences using vehicle design space
exercises, adapted from Axsen & Kurani (2008, 2013). In the design space exercises,
respondents were presented with four versions of the next new vehicle they intended to
purchase: the conventional (market) version, along with theoretical hybrid, plug-in hybrid
(PHEV), and battery electric (BEV) versions of the same vehicle (Figure 4).
Respondents could alter key PEV attributes at varying additional cost: electric range and
speed of home recharge (see Appendix A for vehicle upgrade and charger costs).
Respondents completed two versions of the exercise under two price scenarios; this
study uses results from the second, lower-priced scenario, which simulates costs under
subsidies or lower battery costs. Recharge options and upgrade costs were personalised
to each respondent using data from the Home Recharge Assessment from Part 2 of the
CPEVS survey.
19
Figure 4: Sample vehicle design game from CPEVS: Part 3
20
While internet-based surveys offer advantages in design flexibility, data accuracy
and cost compared to mail or telephone surveys, they have been criticised due to
potential coverage error by excluding those that do not have computer/internet access,
thereby biasing the realised sample (e.g. towards younger, more higher educated
respondents) (Dillman, Smyth, & Christian, 2009). I believe that access concerns are
minimal because internet penetration has increased to 87% in Canada in 2015 (ITU,
2015). Further, given that this study attempts to characterise potential buyers of new
PEVs, a sample potentially biased towards younger, higher-income respondents living in
metropolitan areas, internet access is likely to be nearly universal in this target
population. The realised sample is evaluated and discussed below.
2.2. Sample
1,754 respondents representing new vehicle buyers across Canada (excluding
Quebec) completed all three parts of the survey (Figure 5). I oversampled respondents
residing in British Columbia and Alberta to facilitate an inter-regional comparison of
contrasting electricity grids and potential differences in vehicle use. Initially, 3,179
respondents completed Part 1, with 1,823 completing Part 2, of which 1,754 finished
Part 3 of the survey. I find that the recruited sample is generally representative of new
car buyers, which are typically older (Harris-Decima, 2013), earn higher income, are
more highly educated (Busse, Knittel, & Zettelmeyer, 2013), and are more likely to own
their own home, relative to the general population (Table 2).
21
Figure 5: Geographic distribution of sample (n=1,754)
22
Table 2: Sample demographics and representativeness
Region British Columbia Alberta Ontario Canada a Sample Census Sample Census Sample Census Sample Census
Sample Size 538 4.4m 326 3.6m 616 12.9m 1 754 33.5m
Household Size
1 15.1% 28.3% 11.0% 24.7% 10.9% 25.2% 13.1% 27.6%
2 42.2% 34.8% 44.5% 34.3% 33.3% 32.4% 40.0% 34.1%
3 18.8% 15.0% 20.9% 16.2% 24.4% 16.4% 20.8% 15.6%
4+ 24.0% 22.0% 23.6% 24.9% 31.5% 26.0% 26.2% 22.7%
Household income (pre-tax)
Less than $40,000 16.5% 25.8% 11.4% 19.2% 11.6% 21.8% 14.8% 24.9%
$40,000 to $59,999 26.7% 32.2% 18.2% 25.6% 19.7% 28.2% 23.6% 32.4%
$60,000 to $89,999 35.4% 36.8% 28.6% 31.0% 29.5% 32.8% 33.1% 37.4%
$90,000 to $124,999 40.8% 37.4% 34.7% 33.8% 36.2% 34.5% 39.0% 38.1%
Greater than $125,000 41.9% 36.1% 34.7% 33.5% 39.9% 33.9% 40.1% 36.4%
Residence ownership
Own 75.8% 69.7% 79.1% 73.1% 80.4% 71.0% 77.9% 68.4%
Residence type
Detached House 61.7% 53.8% 71.1% 71.2% 66.6% 61.8% 66.7% 61.9%
Attached House (townhouse, duplex, etc.) 14.8% 23.2% 14.3% 13.7% 17.9% 18.7% 15.3% 17.0%
Apartment 21.3% 20.8% 12.4% 12.0% 14.9% 19.3% 16.4% 20.0%
Sex*
Female 60.8% 51.0% 66.9% 49.9% 55.5% 51.3% 58.4% 51.0%
Age*
under 35 25.8% 30.1% 30.1% 35.8% 33.6% 31.2% 30.0% 31.2%
35-44 18.8% 16.0% 17.5% 17.5% 19.3% 16.6% 18.2% 16.1%
45-54 20.4% 18.9% 19.0% 18.9% 19.8% 19.3% 19.5% 19.1%
55-64 19.5% 16.5% 20.9% 14.0% 16.2% 15.3% 19.2% 15.8%
65+ 15.4% 18.5% 12.6% 13.7% 11.0% 17.6% 13.1% 17.7%
Primary Work Status*
Employed 59.1% 61.6% 58.6% 70.8% 63.6% 62.8% 60.9% 62.3%
Highest level of education completed*
Other b 29.1% 58.6% 28.4% 60.5% 19.2% 57.0% 24.6% 60.2%
College, CEGEP, some university 34.0% 22.1% 37.4% 22.0% 36.3% 22.5% 36.8% 21.7%
University degree (Bachelor) 26.5% 14.2% 24.1% 13.5% 28.9% 15.2% 26.2% 13.5%
Graduate or professional degree 10.5% 5.1% 10.2% 4.0% 15.6% 5.3% 12.4% 4.6%
Note: Data on household size, sex, age, and residence type are from the 2011 Canada Census. Data on work status, education, income, and home ownership are from the 2006 Canada Census. * Sex, age, primary work status, and highest level of education completed are of the person completing the survey. a Overall Canada sample is unweighted. Survey data includes only English-speaking Canada – Quebec was excluded due to language translation costs. Census data includes Quebec. b Less than high school; High school certificate or equivalent; Apprenticeship or trades certificate or diploma
23
More than one-third of respondents designed a PEV in the low-price design
space exercises (Figure 6), ranging from 34% of Alberta respondents to 40% of British
Columbia respondents. I refer to these respondents as the potential “Early Mainstream”
PEV market in each province – respondents that are most likely to characterise the early
PEV market over the near-term (the next 10-15 years). These “Early Mainstream”
respondents have a strong preference for PHEVs; nearly 90% designed a PHEV rather
than a BEV. Respondents residing in British Columbia and Ontario had similar
preferences for PEV designs (Figure 7), while the Alberta sample included a slightly
higher proportion that preferred longer range PEVs, with 63% designing a PHEV-64. My
analysis focuses on the potential PEV usage patterns of “Early Mainstream” respondents
in each province.
Figure 6: Respondent vehicle design in the lower price design game. Respondents who designed a PHEV and BEV were included in the potential “Early Mainstream” sample shown in Figure 7.
Data quality was evaluated at several points of data collection. At the end of
Parts 1 and 2 of the survey, I removed low quality and repeat responders (i.e. multiple
respondents per household) before inviting respondents to complete the next part of the
survey. I identified low quality respondents as those with very short survey completion
23 11 25
192 102 194
0%
20%
40%
60%
80%
100%
BC (n=538)
AB (n=326)
ON (n=616)
% P
refe
rred CV
HEV
PHEV
BEV
24
times and lack of variation in response patterns for a given section. For the driving diary,
missing or inappropriate values were imputed where possible, e.g. AM/PM errors, typos
in data entry, and odometer decimal errors. Through this data cleaning process, the
sample size of “Early Mainstream” respondents decreased slightly in the three regions to
the final samples of 201 in British Columbia, 102 in Alberta, and 194 in Ontario (Figure
7). I use data from these subsample for the analyses presented in this research project.
Figure 7: Share of PEV vehicle designs of the potential “Early Mainstream” sample
2.3. Modeling vehicle use and recharge
I did not collect data on actual usage of PEVs – rather, I collected data on what
types of PEVs the potential “Early Mainstream” PEV buyers may want and how these
households currently operate their current conventional vehicles. Using respondents’
driving diary data and preferred vehicle designs, I developed a vehicle use and recharge
model in Microsoft Excel to simulate each respondent’s vehicle usage patterns and
corresponding electricity and gasoline demand.
0%
20%
40%
60%
80%
100%
BC (n=201)
AB (n=102)
ON (n=194)
% P
refe
rred
PHEV-16
PHEV-32
PHEV-64
BEV-80
BEV-120
BEV-160
BEV-200
BEV-240
25
To anticipate a range of potential usage patterns of “Early Mainstream” PEV
buyers, I model three scenarios to illustrate the impact of recharge access rates (home
and work) and vehicle design preferences:
• Scenario 1: “User Informed” is informed by data collected from the respondents. Respondents are assumed to drive the vehicles they selected in the lower priced PEV design space exercise (Figure 7), predominantly PHEVs. Recharge availability is based on the Home Recharge Assessment and driving diary data. Respondents without any home recharge access are assigned Level 1 access at home for the modeling exercise. I assume that respondents drive their designed PEV exactly as they drive their driving diary vehicle, and plug in to recharge immediately whenever they are parked within 8m (25 ft.) of a recharge opportunity.
• Scenario 2: “User + enhanced workplace access” is identical to Scenario 1, but also assumes that all workplace locations have Level 2 charging access, and that all respondents use these chargers when they are parked at work.
• Scenario 3: “BEV-240 + extended access” models a theoretical “high electrification” scenario where all respondents are driving a battery electric version with 240km of range (BEV-240) of their vehicle and Level 2 recharge access is available at all homes and workplaces.
In all three scenarios, I also assume recharge rates of 1 kW for Level 1 and 6 kW
for Level 2, and usable battery capacities and PEV electricity consumption per km (fuel
efficiency) as shown in Table 3 I do not alter public charging availability in any of the
scenarios. Increased availability of public chargers could encourage within-day charging
for PHEVs in particular, decreasing gasoline consumption by 30% and energy use by
10% in one demonstration in Austin, Texas (Dong & Lin, 2012). However, the effect on
peak demand would likely be small, considering public charging is likely to occur mid-
day when demand is low to moderate.
26
Table 3: Electricity consumption (kWh/km) by vehicle class and usable battery capacity (kWh) for PEV designs and vehicle classes (adapted from Axsen & Kurani, 2013)
Compact Sedan Mid-SUV Full-SUV Consumption (kWh/km) 0.163 0.188 0.252 0.297
Usable Battery Capacity (kWh) PHEV-16 2.6 3.0 4.0 4.7 PHEV-32 5.2 6.0 8.1 9.5 PHEV-64 10.4 12.0 16.1 19.0 BEV-80 13.0 15.0 20.2 23.7 BEV-120 19.5 22.5 30.2 35.6 BEV-160 26.0 30.0 40.3 47.4 BEV-200 32.5 37.5 50.4 59.3 BEV-240 39.0 45.0 60.5 71.2
2.4. Short-run model of GHG emissions
To achieve my first research objective, I calculate potential short-term source-to-
wheel PEV emissions using region-based survey and electricity generation data to
represent the provinces of British Columbia, Alberta, and Ontario. Previous studies use a
number of different approaches to estimate GHG impacts and allocate and attribute
emissions (Yang, 2013). Emissions allocation and attribution, discussed in Section 1.2,
refers to how the net change in electricity demand due to PEVs is attributed to specific
fuels or technologies used to generate the incremental electricity. For this study I
calculate both marginal and average grid emissions intensity. Because the electricity
demand profile of PEVs varies by time of day and day of the week, I use both hourly
marginal and hourly average emissions factors for electricity in my short-run analysis. I
also use an annual average emissions factor for electricity to aid comparison with other
Canadian studies (e.g. Moorhouse & Laufenberg, 2010; Raykin et al., 2012; Ribberink &
Entchev, 2013).
For each province, I construct models of PEV usage and emissions using several
pieces of information. I modeled PEV usage and electricity demand using profiles of
PEV electricity demand over the course of 24 hours (built from the empirical survey data
27
and assumptions described above). I represent the provincial electricity grid mix using
hourly electricity generation and trade data from BC Hydro (April 2012-March 2013),
Alberta Electric Service Operator (AESO) (April 2011-March 2013), and Ontario’s
Independent Electricity System Operator (IESO) (April 2012-March 2014). I assign GHG
emissions factors for each electricity source using median values from the IPCC’s
literature review of lifecycle analyses of electricity emission factors (Table 4).
Table 4: Lifecycle GHG emissions intensity of electricity, by generation type (from Moomaw et al., 2014)
Emissions intensity (g CO2e/kWh)
Coal 1,001 Oil 840 Gas 469 Solar PV 46 Geothermal 45 Solar CSP 22 Biomass 18 Nuclear 16 Wind 12 Ocean 8 Hydro 4
To represent provincial marginal GHG emissions, I followed a “linear-regression”
approach used in Ma et al. (2012) and Siler-Evans, Azevedo, & Morgan (2012). First, I
plot hourly generation/demand (MWh, x-axis) against its corresponding hourly GHG
emissions (kg, y-axis) for a given hour over the study period (e.g. 3 AM over one year).
Each point on the plot represents that hour’s emission factor (kg/MWh); the slope of the
linear regression of a large collection of points is the average marginal emission factor
for that data set (Figure 8). Hourly imports from other jurisdictions (e.g. Saskatchewan,
United States) are included in these calculations. However, due to a lack of data on
28
hourly generation mix of imported power, I assumed the annual average generation mix
for the originating region 4.
Figure 8: Sample estimation of marginal emissions factor
I also estimated hourly average and annual average GHG emissions factors for
each province using the historical generation data referenced above. The annual
average emission factor (which assumes no trade) is comparable to the annual average
emission factors reported in national GHG inventories, e.g. Environment Canada
(2015a). Table 5 compares emission factors by source and method.
4 For imports into Alberta from Saskatchewan, we assumed an average intensity of 730g/kWh (Nyboer, 2014). For imports into BC from the US (WECC), we assumed an average intensity of 384 g/kWh (US EPA, 2014). For imports into Ontario, we assume the following emission factors: Manitoba (2.8 g/kWh), Minnesota (701 g/kWh –MROW), Michigan (743 g/kWh – RFCM), New York (249 g/kWh – NYUP), Quebec (2.9 g/kWh) (Nyboer, 2014; US EPA, 2014).
y = 644.17x + 604880 R² = 0.546
4000
4500
5000
5500
6000
6000 6500 7000 7500
GHG
Emiss
ions
(k
g CO
2e)
Thou
sand
s
Hourly Demand (MWh)
29
Table 5: Comparing emission factors over time and method
Source Environment Canada (2015a) This study Method Annual Average Average
(domestic) Hourly Average (with trade)
Hourly Marginal (with trade)
2000 2005 2010 2012 2011-2014a BC 38 27 26 9.3 11 33 (24-54) 33 (10-64) Alberta 1000 990 1100 930 860 819 (800-839) 964 (781-1264) Ontario 320 240 150 110 100 97 (66-116) 365 (216-472) a Most recent available data was used: BC (April 2012-March 2013), Alberta (April 2011-March 2013), and Ontario (April 2012-March 2014).
GHG emissions from gasoline use for conventional vehicles, HEVs and PHEVs
are estimated using a “well-to-wheels” (WtW) approach, which include emissions
associated with fuel production, transportation, and use. I use recent literature values
(Table 6) to estimate appropriate emissions factors from combustion and upstream
(production and distribution) of gasoline in Canada. Based on the recently updated
Canadian Association of Petroleum Producers (2015) projections for domestic
production, I assume a 1:1 oil sands to conventional oil blend for gasoline production.
Based on these assumptions, I use a WtW emissions factor of 3,461g CO2e/L for
gasoline (2,537 g/L combustion from Cai et al. (2015) and 924 g/L upstream emissions
(using an average of Lattanzio (2014) and Cai et al. (2015).
Table 6: Well-to-wheels gasoline emission factors (g CO2e/L)
Study Crude Combustion Upstream Total Samaras & Meisterling, 2008 US 2 322 659 2 981 Ma et al., 2012 US - - 3 327 Lattanzio, 2014 US 2 516 645 3 161 Cai et al., 2015 US 2 537 662 3 199 Lattanzio, 2014 Canada conventional 2 516 624 3 140 Lattanzio, 2014 Canada oil sands 2 516 1 168 3 684 Cai et al., 2015 Canada oil sands 2 537 1 224 3 761 This study 1:1 oil sands /
conventional blend 2 537 924 3 461
30
Figure 9 summarises the hourly emissions factors for electricity in British
Columbia, Alberta, and Ontario, which can be explained as follows:
• British Columbia’s electricity system is dominated by large hydro and as a result, has the lowest emissions intensity of the three regions. Higher imports of coal-fired electricity from Alberta overnight result in higher emissions intensity during the early morning. Imports and exports to the US (absolute average 675 kW) and Alberta (absolute average 284 kW) are about 14% compared to domestic consumption (6976 kW).
• In Alberta, emissions intensity of electricity is high due to the province’s dependence on coal-fired electricity (Environment Canada, 2014). I find that the calculated hourly marginal emissions factor fluctuates more than the hourly average, with higher emissions during low load hours (peaking at around 4 AM) and lower emissions during the early afternoon. This is due to the peak hour marginal demand being met by peaking natural gas and imported hydro electricity from British Columbia.
• Ontario’s policies to phase-out coal generation and increase renewables have contributed to a considerable decrease in average emissions intensity over the past five years, decreasing nearly 40% since 2009. The majority of present generation is from nuclear, with the balance supplied by natural gas, hydro, solar, and other renewables. Emissions factors are highest around noon, because the marginal demand is being met largely through natural gas-fired plants, which have higher emissions intensity compared to the other generation sources in Ontario.
31
Figure 9: Marginal hourly and average hourly emissions factor using recent generation data
Alberta - Marginal
Ontario - Marginal
BC - Marginal
Alberta - Average
Ontario - Average
BC - Average
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
0 3 6 9 12 15 18 21 24
Hour
ly Em
issio
ns F
acto
r (g
CO2
e/kW
h)
Hour Ending
32
2.5. CIMS as a long-run energy-economy model
To achieve my second and third research objectives, I estimate the GHG
emissions associated with PEV usage in Canada using a long-term perspective, where
the stock of vehicles and electricity generation plants are changing over time (in a multi-
decadal time scale). To do so, I use the CIMS energy-economy model to simulate the
effect of energy policies on GHG emissions in five-year steps from 2000 to 2050,
focusing on the electricity and personal transportation sectors.
CIMS is a hybrid energy-economy model developed at Simon Fraser University
that simulates capital stock turnover through time, as technologies are acquired,
retrofitted, and retired and replaced (Jaccard, 2009). It improves on traditional
technology-specific “bottom-up” models by incorporating aspects of full economy “top-
down” models such as macroeconomic feedbacks and behavioural realism informed by
empirical research (Axsen, Mountain, & Jaccard, 2009; Mau, Eyzaguirre, Jaccard,
Collins-Dodd, & Tiedemann, 2008). CIMS has been used to inform energy and
environmental policies at municipal, provincial, federal and international (US, China)
levels. Further details of the model and its parameters, specifically on the transportation
sector, are described in Fox (2013) and Murphy (2000). CIMS represents the uptake of
new technologies through the market share algorithm represented below:
The equation considers the primary factors that affect vehicle purchase decisions
and resulting market share allocation for technology j among k technologies: capital
costs (CC), maintenance costs (MC), energy costs (EC), intangible costs (i), and private
discount rate (r) (Fox, 2013). Briefly, CIMS includes the following functions:
• Stock turnover: over time, old vehicles, energy-using equipment and power plants are retired. To meet new demand and replace retired stock, CIMS determines market shares for new technologies through the above market share algorithm.
�
���
�
𝑀𝑆 =𝐶𝐶 ∗ 𝑟
1 − (1 + 𝑟) +𝑀𝐶 + 𝐸𝐶 + 𝑖
∑ 𝐶𝐶 ∗ 𝑟1 − (1 + 𝑟) +𝑀𝐶 + 𝐸𝐶 + 𝑖
�
Equation 1
�
,�WDNH�WKLV�IXQFWLRQ�GLUHFWO\�IURP�WKH�IXOO�&,06�86�PRGHO� (TXDWLRQ�� DFFRXQWV�IRU�
WKH�SULQFLSOH�IDFWRUV�DIIHFWLQJ�FRQVXPHUV¶�YHKLFOH�SXUFKDVH�GHFLVLRQV��FDSLWDO�FRVWV��
PDLQWHQDQFH�FRVWV��HQHUJ\�FRVWV��WKH�SHUFHLYHG�LQWDQJLEOH�FRVWV�RI�QHZ�WHFKQRORJLHV�
ZLWK�ZKLFK�FRQVXPHUV�KDYH�OLWWOH�XVHU�H[SHULHQFH��DQG�D�SULYDWH�GLVFRXQW�UDWH�IRU�HDFK�
GHFLVLRQ�PDNHU��7KH�DOJRULWKP�DOORFDWHV�QHZ�PDUNHW�VKDUHV��MS��DPRQJVW�K
WHFKQRORJLHV��CCj��MCj��DQG�ECj�UHSUHVHQW�WKH�FDSLWDO��PDLQWHQDQFH��DQG�HQHUJ\�FRVWV�RI�
WHFKQRORJ\�j, UHVSHFWLYHO\��7KH�FDSLWDO�FRVW�LV�WKH�SXUFKDVH�SULFH�WKDW�FRQVXPHUV�SD\�IRU�
WKH�SURSXOVLRQ�V\VWHP��7KH�PDLQWHQDQFH�FRVW�LV�DQ�HVWLPDWH�RI�WKH�DQQXDO�DPRXQW�
FRQVXPHUV�ZLOO�KDYH�WR�SD\�WR�PDLQWDLQ�WKH�YHKLFOH�WHFKQRORJ\�RYHU�LWV�OLIHVSDQ��)LQDOO\��
WKH�HQHUJ\�FRVW�LV�DQ�HVWLPDWH�RI�WKH�DQQXDO�FRVW�RI�IXHO���
7KH�WHFKQRORJ\�FRPSHWLWLRQ�IXQFWLRQ�DOVR�FRQWDLQV�WKUHH�EHKDYLRXUDO�SDUDPHWHUV��
i��r, DQG�v�WKDW�GHILQH�FRQVXPHU�SUHIHUHQFHV�RYHU�FRPSHWLQJ�YHKLFOH�SURSXOVLRQ�
WHFKQRORJLHV��7KH�YDOXHV�IRU�WKHVH�SDUDPHWHUV�DUH�GUDZQ�IURP�SUHYLRXV�HPSLULFDO�
UHVHDUFK��$[VHQ��0RXQWDLQ��DQG�-DFFDUG�������0DX�HW�DO��������+RUQH��-DFFDUG��DQG�
7LHGHPDQQ��������,�GLVFXVV�PRGHOLQJ�FRQVXPHU�SUHIHUHQFHV�IXUWKHU�LQ�6HFWLRQ�������7KH�
ij SDUDPHWHU�LV�WKH�SHUFHLYHG�LQWDQJLEOH�FRVW�RI�WHFKQRORJ\�j��,Q�WKH�&,06�PRGHO��
LQWDQJLEOH�FRVWV�UHSUHVHQWV�IDFWRUV�WKDW�DUH�QRW�UHIOHFWHG�LQ�WKH�ILQDQFLDO�FRVWV�RI�D�
SDUWLFXODU�WHFKQRORJ\�EXW�QRQHWKHOHVV�DIIHFW�WKH�FKRLFHV�RI�FRQVXPHUV��
�7KH�r�SDUDPHWHU�LV�WKH�SHUFHLYHG�SULYDWH�GLVFRXQW�UDWH��ZKLFK�GHWHUPLQHV�KRZ�
FRQVXPHUV�SHUFHLYH�WKH�IXWXUH�FRVWV�WKDW�DUH�DVVRFLDWHG�ZLWK�D�WHFKQRORJ\�LQ�WKH�FXUUHQW�
WLPH�SHULRG��7UDLQ��������,Q�ERWK�&,06�86�39�DQG�WKH�IXOO�&,06�86�PRGHO��WKH�r IRU�
YHKLFOHV�LV��������ZKLFK�LV�D�UHODWLYHO\�KLJK�GLVFRXQW�UDWH�FRPSDUHG�ZLWK�WKH�GLVFRXQW�
UDWHV�IRU�RWKHU�HQHUJ\�WHFKQRORJLHV��$[VHQ��0RXQWDLQ��DQG�-DFFDUG�������+RUQH��
-DFFDUG��DQG�7LHGHPDQQ��������)DFWRUV�WKDW�FRQWULEXWH�WR�WKLV�KLJK�GLVFRXQW�UDWH�DUH�
OLPLWHG�WLPH�IRU�FRQVXPHUV�WR�PDNH�GHFLVLRQV��WKH�SHUFHSWLRQ�RI�D�JUHDWHU�ULVN�RI�IDLOXUH�
33
• Technology dynamics: for new and emerging technologies such as PEVs, solar panels, wind turbines, capital costs (CC) decline as more of the technology is adopted (also known as “learning curves”).
• Behavioural realism: the CIMS market share algorithm represents how the decision-maker perceives the financial costs of new technologies, the intangible costs (i), the time value of money (r), and the heterogeneity of the market (v). These behavioural parameters are derived from previous empirical research (e.g. Axsen et al., 2009; Mau et al., 2008).
• Behavioural dynamics: a declining intangible cost function represents how consumers’ preferences (i) may change as new technologies become more prevalent in the market (as new market share increases).
While I model energy use across the Canadian economy, I only report results on
the transportation and electricity sectors in British Columbia, Alberta, and Ontario. In the
personal transportation sector, I focus largely on passenger vehicles, specifically the
“competition node” in CIMS that determines new market share for passenger vehicle
motors (Figure 10).
Figure 10: Motors available in “Passenger Vehicle Motors” in the CIMS Transportation sector
34
CIMS includes detailed cost and fuel consumption characteristics for a range of
vehicle propulsion types. I created six PEV model types to represent a range of PEV
models available in the market (numbers indicate all-electric range in km): PHEV-16
(Toyota Prius PHEV), PHEV-32, PHEV-64 (Chevrolet Volt), BEV-120 (Nissan LEAF),
BEV-240, and BEV-480 (Tesla Model S). These model types (except the BEV-480) were
developed to align with the available model types in the short-run model (based on the
CPEVS survey design exercises).
The Electricity sector in CIMS is region-specific, and generation is exogenously
allocated into base (80%), shoulder (15%), and peak (5%) loads. Table 7 shows the
available generation types by region, reflecting current and likely availability of
technologies for each region and load type. For example, nuclear and coal are not
available in British Columbia, while intermittent renewable power is assumed to only be
able to provide base load power in all regions.
35
Table 7: Available generation types in CIMS by province and load
British Columbia Alberta Ontario
Base Load
Coal X X
Diesel X X
Hydro (Large) X X
Hydro (Small) X X X
Natural Gas X X X
Nuclear X X
Biomass X X X
Fuel Cell X X X
Geothermal X X X
Solar PV X X X
Solar Thermal X X X
Wind X X X
Shoulder Load
Coal X X
Diesel X X
Heavy Fuel Oil X
Hydro (Large) X X
Hydro (Small) X X X
Natural Gas X X X
Nuclear X X
Peak Load
Coal X X
Diesel X X X
Heavy Fuel Oil X
Hydroelectric (Large) X X
Hydroelectric (Small) X X X
Natural Gas X X X
36
The generation types compete for market share to meet the required demand for
base, shoulder, and peak loads. Emissions are calculated by multiplying the total
generation (kWh) of each type with its respective emission factor (g CO2e/kWh) to obtain
load-specific and total average emission factors. To calculate load-explicit PEV impacts,
I allocated an average 24-hour day into low, medium, and high load hours based on
historical residential electricity demand profiles in Alberta and Ontario (Toronto Hydro,
2011), and assumed the percentage of each generation type summarised in Table 8.
These are then matched with hourly PEV electricity demand profiles modeled from
driving diary data, described in Section 2.3.
Table 8: Allocation of hours and generation types by low, medium, and high load hours
Hours % Base % Shoulder % Peak
Low Load Hours (LLH) 6 (0h-6h)
100% 0% 0%
Medium Load Hours (MMH) 12 (6h-16h; 22h-24h)
84% 16% 0%
High Load Hours (HLH) 6 (16h-22h)
67% 16% 17%
I use to the CIMS model to estimate GHG emissions from PEVs in the long-term,
as PEV adoption increases and the electricity sector evolves. The evolution of
technology adoption in these sectors can be strongly shaped by the climate policies put
in place by governments. To explore the importance of policy, I model three policy
scenarios in CIMS, described below and in detail in Table 9:
1. Reference (Current Policies): existing federal and provincial climate policies affecting the transportation and electricity sectors, including federal and provincial (Ontario) regulations to phase-out coal, BC’s low carbon fuel standard and carbon tax, as well as provincial incentives on the purchase of PEVs in BC and Ontario;
2. Carbon Tax: escalating, economy-wide carbon tax starting in 2016 at $25/t and increasing to $250/t by 2050. The price
37
schedule is based on the International Energy Agency’s CO2 price assumptions in the World Energy Outlook “450 Scenario”5 for the United States and Canada (International Energy Agency, 2014b);
3. Strong Standards (ZEV mandate and Clean Electricity): standards that require the deployment of zero-emissions electricity and zero-emissions vehicles over the medium term (2025), including a minimum PEV share of new vehicles of 12.75% in 2025 (9.25% PHEV; 3.5% BEV).
5 The 450 Scenario illustrates the policies and technologies needed to achieve an emissions trajectory
consistent with limiting global average long-term temperature increase to 2°C.
38
Table 9: Modeled scenarios in CIMS
Scenario Details
1. Reference (Current Policies Scenario) Existing federal and provincial climate policies affecting transportation and/or electricity sectors
Canada-wide: Reduction of Carbon Dioxide Emissions from Coal-Fired Generation of Electricity Regulations6: no new conventional coal plants after 2015; 45-year economic lifespan (following methods described in Kniewasser, 2014).
British Columbia Alberta Ontario
• Clean Energy Vehicle Program7: purchase subsidy on PHEV ($2500), BEV ($5000) and chargers ($500) (2016-2020 period) • Renewable and Low Carbon Fuel Requirements Regulation8, Renewable Fuel Requirement (Part 2): 5% renewable content in gasoline; 4% renewable content in diesel • Carbon Tax9: $30/t from 2011-2050
None10 • Electric Vehicle Incentive Program: purchase subsidy on PHEV ($5000 and up), BEV (up to $8500) and chargers ($1000) from for a 5-year period (2015-2020) • Ending Coal for Cleaner Air Act
2. Carbon Tax Economy-wide carbon tax starting at $25/t in 2016 and increasing $5/t/yr from 2016-35 and $10/t/yr from 2036-50
2000-2015: $0/t CO2e 2016-2020: $25/t CO2e 2021-2025: $50/t CO2e 2026-2030: $75/t CO2e 2031-2035: $100/t CO2e 2036-2040: $150/t CO2e 2041-2045: $200/t CO2e 2046-2050: $250/t CO2e
3. Strong standards (PEV + Clean Electricity Standard) Policy package of a PEV standard and clean electricity standard
• PEV share of new vehicles reach a minimum of 12.75% in 2025 (9.25% PHEV; 3.5% BEV), which aligns with California’s 2012 ZEV Program amendments (California Air Resources Board, 2012)). The minimum market share is linear over time: 6.4% in 2020, 12.75% in 2025, 19.1% in 2030, etc.) • No new fossil fuel generation without CCS after 2025
6 https://www.ec.gc.ca/lcpe-cepa/eng/regulations/detailreg.cfm?intReg=209 7 http://www.cevforbc.ca/ 8 http://www.empr.gov.bc.ca/RET/RLCFRR/Pages/default.aspx 9 http://www.fin.gov.bc.ca/tbs/tp/climate/carbon_tax.htm 10 Due to the complex nature and limited applicability of Alberta’s Specified Gas Emitters Regulation (large emitters, $15/t, applying to only 12% of emissions for an average maximum cost of $1.8/t), we opt not to model this policy.
39
2.6. Informing CIMS with short-run modeling outputs
In practice, parameters in the CIMS model are informed by empirical research
and literature reviews, which are updated and maintained by CIMS users depending on
the specific application. For my project, I used a March 2012 version of CIMS (“NRTEE
Investment”) as my base model. I could not access more recent versions of the model
because these versions contained proprietary or confidential data that could not be
released. For the purposes of this project, I revised several model parameters in the
Transportation sector, informed by results from the CPEVS survey and recent literature,
to reflect regional differences in vehicle preferences and driving patterns.
Table 10 provides an overview of key model parameters and describes the
rationale and sources for any updates. The most notable updates are the inclusion of
intangible costs (varying by region and PEV type) and changes to motor (gasoline and
electricity) efficiency assumptions based on the results from the CPEVS vehicle design
games and driving diary. The intangible costs were estimated by calibrating the relative
market shares of the PEV models in 2015 to the relative market shares observed in the
CPEVS vehicle design results reported shown in Figure 7 (e.g. large share of PHEV-64).
However, based on current trends in PEV market penetration, I did not calibrate the new
market share of PEVs overall to the “latent demand” (~40% of new vehicles) which
assumes no real world constraints. The differences between latent demand and actual
market penetration is described and analysed in Axsen & Wolinetz (2014; 2015).
40
Table 10: Key CIMS parameters and assumptions and updates based on CPEVS data and recent literature
Parameter Assumption Updates
Average vehicle lifespan 16 years No change.
Capital cost Varies by vehicle class (Table 11) and motor type (Table 12)
Capital costs for PEVs were updated to better align with CPEVS cost assumptions while considering the updates to intangible costs and progress ratio.
Progress ratio (experience curve)
HEV: 0.93; PHEV: 0.9; BEV: 0.87
Updated based on recent estimates in Nykvist & Nilsson (2015) on battery costs.
Heterogeneity (v) 15 (motor) No change. See Fox (2013) and Mallory (2007) for sensitivity analyses.
Private discount (r) 24-27%
Minor updates to account for regional differences in vehicle fuel consumption seen in CPEVS.
Intangible costs (i): vehicle class
Varies by region (Table 11)
Reduced intangible benefit of large cars/trucks to calibrate to actual market shares observed in Canada.
Intangible costs (i): motors
Varies by motor type (e.g. electric range) and region (Table 12)
Previously, CIMS did not have any intangible costs for PEVs. Upfront fixed and declining intangible costs were added based on observed preferences in the vehicle design experiment in CPEVS.
Technology share (class) Varies by region (Table 11)
Updated based on vehicle shares from the most recent Canada Vehicle Survey (Statistics Canada, 2010).
Motor efficiency Varies by PEV type and region (Table 13)
Updated based on CPEVS. Diary data and electricity demand models used to estimate electric utility factors for different PHEV ranges. Aligned fuel and electricity consumption efficiencies with CPEVS assumptions.
Table 11 summarises the assumptions for select parameters related to vehicle
classes (e.g. small cars vs. large trucks). Initial capital costs reflect relative differences in
costs by vehicle (size) class, annual operating costs reflect ongoing costs such as
maintenance, fixed intangible costs reflect the relative differences in preference between
the classes in each region (a negative value indicates intangible value, i.e. preference),
and class shares describe the market share of each vehicle class type within a region.
Capital costs are highest for large trucks, while ongoing operating costs (excluding fuel
costs) are assumed to be equal across classes. The fixed intangible costs vary by
vehicle class and region, reflecting regional differences in preferences. For example, the
41
large positive value (i.e. high intangible cost) for small cars reflects consumer aversion to
small cars, particularly in Alberta. On the other hand, a negative value for large trucks
implies an intangible benefit or a positive preference. Class share reflects the relative
market shares of different vehicle classes in each region (e.g. more trucks than cars in
Alberta) based on results from the vehicle design exercises. The shares are in line with
the relative shares of cars and trucks in the most recent Canada Vehicle Survey
(Statistics Canada, 2010).
Table 11: Regional differences in passenger vehicle attributes by vehicle class
Type Capital Cost
Operating Cost
Fixed Intangible Costs Class Share
(initial) (per year) BC Alberta Ontario BC Alberta Ontario
Small car $16 217 $2 139 $11 000 $12 750 $12 750 19% 20% 28%
Large car $24 952 $2 139 $5 000 $5 000 $2 750 37% 24% 28%
Small truck $28 434 $2 139 $500 $250 $2 750 22% 28% 22%
Large truck $37 169 $2 139 -$11 750 -$13 500 -$10 000 22% 28% 22%
Table 12 shows the assumptions for select parameters related to motor types.
Capital cost assumptions are based on previous versions of CIMS, with minor
adjustments to PEVs to align with the short-term model assumptions in the vehicle
design exercises. Declining intangible costs (i.e. intangible costs that decline with
increased market penetration) were introduced to most PEV models, and are higher for
shorter range BEVs due to their limited electric range. I calibrated the fixed intangible
costs for each region to reflect differences in vehicle preference among provinces based
on initial shares of PEV model types from the vehicle design exercises described earlier
in Figure 7.
42
Table 12: CIMS passenger vehicle motors parameters and assumptions: capital cost, declining intangible costs, and fixed intangible costs
Capital Cost Declining Intangible
Cost
Fixed Intangible Cost
BC Alberta Ontario
Gas Std Eff $5 000 $0 $0 $0 $0
Gas Med Eff $5 200 $0 $0 $0 $0
Gas High Eff $6 000 $0 $0 $0 $0
Diesel Std $8 000 $0 $0 $0 $0
Diesel Med $8 200 $0 $0 $0 $0
Diesel High $9 000 $0 $0 $0 $0
Ethanol $10 581 $0 $0 $0 $0
Hydrogen $150 437 $0 $0 $0 $0
Hybrid $12 000 $0 $0 $0 $0
PHEV-16 $16 000 $1 000 $4 000 $5 500 $3 500
PHEV-32 $18 000 $1 000 $1 500 $3 000 $1 500
PHEV-64 $21 000 $1 000 -$500 -$600 -$750
BEV-120 $20 000 $4 000 $4 500 $4 000 $5 000
BEV-240 $24 000 $1 500 $3 750 $3 500 $3 750
BEV-480 $30 000 $0 -$750 -$2 500 -$500
Table 13 summarises my technical assumptions. The relative differences among
the four vehicle size classes are based on previous versions of CIMS. Gasoline
efficiencies for conventional, hybrid, and plug-in hybrid vehicles were based on prior
versions of CIMS while incorporating assumptions from the short-term model, e.g. HEVs
are one-third more efficient than medium efficiency gasoline vehicles, PHEVs have the
same gasoline efficiency as HEVs once the battery has been depleted to its minimum
state of charge. Electricity efficiencies for BEVs were based on assumptions from the
short-term model. PHEV electricity efficiencies for each model and region were
calculated using the results from the short-term model (Scenario 1: user-informed
43
vehicle design and charging availability modeled with driving diary data). For PHEVs, the
gasoline and electricity consumption rates represent the average between charge
depleting and charge sustaining modes based on the assumed utility factor for each
model. I find that these assumptions generally align with the technical parameters for
PEVs available on the market today (Table 14).
Table 13: CIMS passenger vehicle motor technical assumptions: gasoline and electricity consumption and electric utility factor.
Gasoline Consumption (L/100km)
Electricity Consumption (kWh/km)
Utility Factor
Motor Small car
Large car
Small truck
Large truck
Small car
Large car
Small truck
Large truck
Gas Std Eff 7.1 8.1 10.9 12.9 - - - - -
Gas Med Eff 6.0 6.9 9.3 10.9 - - - - -
Gas High Eff 5.1 5.9 7.9 9.3 - - - - -
Hybrid 4.0 4.6 6.2 7.3 - - - - -
PHEV-16 2.0 2.4 3.2 3.7 0.065 0.075 0.101 0.119 0.40
PHEV-16 (ON) 2.4 2.8 3.7 4.3 0.049 0.056 0.075 0.089 0.30
PHEV-32 1.8 2.1 2.8 3.3 0.078 0.089 0.120 0.141 0.48
PHEV-64 1.2 1.4 1.8 2.2 0.106 0.122 0.164 0.193 0.65
BEV-120 - - - - 0.186 0.214 0.288 0.339 1
BEV-240 - - - - 0.186 0.214 0.288 0.339 1
BEV-480 - - - - 0.186 0.214 0.288 0.339 1
44
Table 14: Comparing technical parameters of available PEV models with CIMS assumptions.
PEV Model Consumption CIMS Archetype Consumption
Type Gasolinea (l/100km)
Electricity (kWh/km)
Motor class Vehicle class Gasolinea (l/100km)
Electricity (kWh/km)
Chevrolet Volt PHEV-61 6.4 0.22 PHEV-64b Large car 4.61 0.188
Nissan LEAF BEV-135 - 0.18 BEV-120 Small car - 0.19 a Gasoline consumption (if applicable) is reported for charge-sustaining operation. b The CIMS PHEV-64 efficiencies are calculated based on the assumed utility factor of 0.475.
45
Chapter 3. Results
This chapter presents the results of the analyses. First, I summarise the
behavioural data collected from the survey and construct consumer-informed recharge
profiles for the short-term model. I then show the key results of this work: the short and
long-term emissions intensity of PEVs.
3.1. Driving activity and recharge potential
Trip timing and distance traveled per day could be important determinants of
electricity use and GHG emissions. For example, final home arrival times could have
important electricity load implications for PEVs, as it generally predicts peak load if
charging is unconstrained.
In the CPEVS diary data, the timing of driving activities among the “Early
Mainstream” respondents between the three provinces is similar (Figure 11). On
weekends, initial trips start later in the day, and driving activity levels peak during mid-
day (Figure 12). Peak arrival time is around 5 PM in all three provinces, which is
comparable to analyses using US national travel data (Weiller, 2011). On average,
Ontario respondents arrive home a bit later, with 65% arriving home by 8PM, compared
to 71% and 72% in BC and Alberta respectively. Average daily driving distances (for
driving days) are highest in Ontario (61 km), followed by Alberta (53 km) and BC (49
km). These regional differences are comparable to the 2009 Canada Vehicle Survey
(Natural Resources Canada, 2011), and similar to previous analyses of 2001 national
US data (Jaramillo, Samaras, Wakeley, & Meisterling, 2009; Weiller, 2011).
46
Figure 11: Proportion of “Early Mainstream” respondents driving by time of day, averaged across weekdays and weekends in British Columbia (n=201 respondents, 603 diary days), Alberta (n=102 respondents, 306 diary days), and Ontario (n=194 respondents, 582 diary days).
Figure 12: Proportion of “Early Mainstream” respondents driving on weekdays (460 diary days) vs. weekends (143 diary days) by time of day in British Columbia.
BC
AB
ON
0%
2%
4%
6%
8%
10%
12%
14%
16%
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
% D
rivin
g
Time of Day
BC AB ON
BC - Weekday
BC - Weekend
0% 2% 4% 6% 8%
10% 12% 14% 16% 18% 20%
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
% D
rivin
g
Time of Day
BC - Weekday
BC - Weekend
47
Using data from the CPEVS Home Recharge Assessment (summarised in
Section 2.1), I find that access to existing 120V outlets (Level 1) at home parking
locations is highest in Alberta, while access to 240V outlets (Level 2) is similar among
the three regions (Figure 13). I define home recharge “access” as regularly parking their
diary vehicle within 25 feet (~8 metres) of an outlet. Housing type and parking space
type are predictors of home recharge access, particularly Level 1 access. For example,
British Columbia as a whole has a higher proportion of apartments compared to Alberta
or Ontario, which is likely one reason that BC has the lowest rates of Level 1 access.
Parking spaces typically associated with detached and attached homes (driveways and
garages) are more likely to be located closer to electrical outlets than parking lots. I also
collected data on recharge access at home and other locations through the driving diary,
applying the same proximity threshold for “access” of 25 feet.
Figure 13: Existing home recharge access (i.e. outlets within 25 feet of the vehicle’s typical parking location) from Part 2: Home Recharge Assessment
Generally, recharge access decreases during the daytime due to a higher
proportion of active/driving vehicles and generally lower access to outlets away from
home (Figure 14). 16-30% of respondents reported having some form of recharge
access at work, including 120-V outlets (Figure 15). While existing EV recharge stations
are rare (less than 5%), over 25% of the early mainstream respondents in Alberta report
having existing outlets (likely for block heaters) at their typical parking location at work.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Level 1 Level 2
% o
f res
pond
ents
with
exist
ing
acce
ss British Columbia
Alberta
Ontario
48
Figure 14: Driving activity and recharge access by time of day, British Columbia potential Early Mainstream respondents from Part 2: Driving Diary (n=201, 603 diary days)
Figure 15: Proportion of existing workplace recharge access reported by potential Early Mainstream respondents in BC (n=201), Alberta (n=102), and Ontario (n=194).
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
100%
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00
% o
f Veh
icles
Time of Day
Driving
Parked No Recharge Access
Parked Recharge Access (Other)
Parked Recharge Access (Work)
Parked Recharge Access (Home)
0%
20%
40%
60%
80%
100%
BC AB ON
% o
f the
Ear
ly Ma
inst
ream
No Charge Access
Yes - 120V
Yes - EV Charging Station
49
3.2. PEV electricity recharge profiles
I model the three PEV usage scenarios described in Section 2.3 for each
province, considering different vehicle designs and recharge availability at home and
work. As a reminder, these scenarios are:
1. User informed: representing survey respondents’ selected PEV designs, driving behaviour, and present recharge access.
2. User + enhanced workplace access: same as Scenario 1, but with enhanced workplace recharge access (i.e. assuming Level 2 access is universally available at all workplaces).
3. BEV-240 + extended access: using respondents’ driving data, but assuming each “Early Mainstream” respondent is driving a pure BEV-240, and that Level 2 recharge access is universally available at all homes and workplaces.
Table 15 summarises the modeled PEV electricity usage by respondents in each
province, for each scenario. Average daily electricity demand ranges from 6.9 to
14.3 kWh/day, per vehicle. Of the three provinces modeled, British Columbia has the
lowest average electricity demand per PEV under all three scenarios, mostly due to the
shorter driving distances. Alberta has higher electricity demand per PEV than Ontario
under Scenario 1 due to the higher frequency of existing workplace recharge access
(likely due to the existence of block heaters, due to the colder climate). However, in
Scenario 3 where all drivers are assigned a 240-km BEV, average daily electricity
demand in Ontario exceeds Alberta’s because on average, drivers in Ontario drive
further each day. The “electric utility factor” refers to the proportion of total driving
distance that is powered by electricity (instead of gasoline), where increasing charging
access and electric range achieve higher electricity utility factors for PHEV designs
(whereas BEV designs always have 100% utility factors).
50
Table 15: Average daily electricity demand, electric utility factor, and morning and evening peak information, by scenario
Scenario Avg. daily demand
(kWh/vehicle)
Utility factor (across
respondents)
Morning Peak Demand Evening Peak Demand Time Demand, kW
(vs. Scenario1) Time Demand, kW
(vs. Scenario 1) BC 6.9 61% 9:20 0.16 18:00 0.73
1 AB 8.7 64% 9:18 0.35 17:55 0.86 ON 8.0 57% 9:47 0.18 16:46 0.77 BC 8.2 74% 8:27 0.72 (+350%) 17:44 0.80 (+10%)
2 AB 9.5 67% 8:58 0.68 (+94%) 18:01 0.95 (+10%) ON 9.5 65% 8:50 0.84 (+354%) 16:46 0.73 (-5%) BC 12.2 100% 8:26 0.71 (+344%) 17:20 1.27 (+74%)
3 AB 13.7 100% 9:00 0.75 (+114%) 18:01 1.37 (+59%) ON 14.3 100% 8:50 0.91 (+392%) 18:00 1.27 (+65%)
The three scenarios produce very different time of day demand profiles for PEVs.
Figure 16 illustrates the three scenarios using the BC sample. Across the week,
Scenarios 1 and 2 follow similar electricity demand profiles in the afternoon with peaks in
the early evening. Due to enhanced workplace access in Scenario 2, there is an
additional peak in the morning as vehicles arrive at work—this effect is more pronounced
among BC and Ontario respondents due to their presently lower recharge potential.
Scenario 3 has a large morning peak and even larger evening peak due to enhanced
home recharge access (with universal Level 2 access), and the assumption of universal
BEV usage (where electricity powers 100% of all vehicle km). Comparing weekday and
weekend loads in British Columbia, we observe a slightly earlier evening (and additional
late afternoon peak) in Scenario 1 (Figure 17). In Scenario 2 there is a more pronounced
difference between weekdays and weekends, particularly the heightened morning peak
during weekdays (Figure 18).
51
Figure 16: Electricity demand profiles under three scenarios in British Columbia (n=201; 603 diary days)
Scenario 1: User Informed
Scenario 2: User Vehicle + Enhanced Workplace (L2)
Scenario 3: BEV-240 + Home/Work L2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Aver
age L
oad
(kW/P
EV)
Time of Day
Scenario 1: User Informed
Scenario 2: User Vehicle + Enhanced Workplace (L2)
Scenario 3: BEV-240 + Home/Work L2
52
Figure 17: Electricity demand profiles in British Columbia: weekdays (460 diary days) vs. weekends (143 diary days) for Scenario 1: User informed
Figure 18: Electricity demand profiles in British Columbia: weekdays (460 diary days) vs. weekends (143 diary days) for Scenario 2: User Vehicle + Enhanced Workplace L2
Weekdays
Weekends
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Aver
age L
oad
(kW
/PEV
)
Time of Day
Weekdays
Weekends
Weekdays Weekends
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Aver
age L
oad
(kW
/PEV
)
Time of Day
Weekdays
Weekends
53
3.3. Short-run greenhouse gas impacts of PEVs
I simulated short-run source-to-wheel greenhouse gas emissions using the
above recharge profiles, and the electricity generation data for each province as
described in Section 2.4.
Figure 19 shows the results using marginal emission factors for electricity. This
represents the emissions impacts of PEVs resulting from the incremental (marginal)
electricity generated to meet the additional load demanded by PEVs. There is
substantial regional variation in the fleet-wide emissions intensity of PEV travel
compared to conventional gasoline vehicles and HEVs. Across the scenarios, British
Columbia shows the greatest potential emission reduction from PEVs due to its very
clean electricity generation system, with Scenario 3 (all BEVs) achieving a 98%
reduction. Under Scenario 1, which reflects the most likely short-term scenario (with a
fleet of primarily PHEV designs using only existing recharge access), PEVs can reduce
fleet-average GHG emissions intensity by 78% in BC, 37% in Alberta, and 58% in
Ontario relative to conventional (gasoline) vehicles. The three scenarios achieve
relatively modest reductions in Alberta (less than 5% compared to HEVs), due largely to
their emissions-intense electricity grid.
54
Figure 19: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of conventional, hybrid electric, and plug-in electric vehicles in British Columbia, Alberta, and Ontario, using hourly marginal emissions factors for electricity, (including trade).
To compare the effects of different emissions allocation approaches, I also
calculated emissions intensity of PEVs using average hourly (with trade) and domestic
average annual emission factors from the same data sets (Figure 20 and Figure 21).
0
50
100
150
200
250
300
350
BC AB ON
Emiss
ions
inte
nsity
(g C
O2e/k
m) Conventional (CV)
Hybrid-Electric (HEV)
PEV - Scenario 1: User Informed
PEV - Scenario 2: Enhanced workplace charging
PEV - Scenario 3: BEV-240 with home and work L2
55
Figure 20: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of conventional, hybrid electric, and plug-in electric vehicles in British Columbia, Alberta, and Ontario using average hourly emissions factors for electricity (including trade).
Figure 21: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of conventional, hybrid electric, and plug-in electric vehicles in British Columbia, Alberta, and Ontario using average annual emissions factors for electricity (excluding trade).
0
50
100
150
200
250
300
350
BC AB ON
Emiss
ions
inte
nsity
(g C
O2e/k
m) Conventional (CV)
Hybrid-Electric (HEV)
PEV - Scenario 1: User Informed
PEV - Scenario 2: Enhanced workplace charging
PEV - Scenario 3: BEV-240 with home and work L2
0
50
100
150
200
250
300
350
BC AB ON
Emiss
ions
inte
nsity
(g C
O2e/k
m) Conventional (CV)
Hybrid-Electric (HEV)
PEV - Scenario 1: User Informed
PEV - Scenario 2: Enhanced workplace charging
PEV - Scenario 3: BEV-240 with home and work L2
56
The results in British Columbia are generally consistent across all three
emissions allocation approaches, with slightly lower emissions intensity calculated in the
annual average approach due to the lack of consideration for imported coal-fired power
from Alberta. The average hourly and average annual approaches estimate lower PEV
emissions intensity in Alberta and substantially lower emissions intensity in Ontario
compared to the hourly marginal approach. This is a result of differences in the mix of
marginal generation sources during PEV charging. For example, peaking natural gas
fired plants, which have higher emissions intensity compared to the other generation
sources in Ontario, result in considerably higher emissions intensity estimates using the
marginal approach. Using an hourly average approach in Ontario results in an estimate
that is 30% lower than the marginal hourly approach (Scenario 1). Full results for all
scenarios and emissions allocation approaches are shown in Table 16.
Table 16: Source-to-wheel greenhouse gas emissions intensity (g CO2e/km) of conventional (CV), hybrid-electric (HEV), and plug-in electric (PEV) vehicles
Scenario Province CV HEV PEV Annual average
PEV Hourly average
PEV Hourly marginal
PEV % Reduction vs. CV
BC 310 206 65 67 67 78-79% 1 Alberta 323 215 196 189 205 37-41% Ontario 302 201 88 88 126 58-71% BC 310 206 51 53 53 83-84%
2 Alberta 323 215 201 193 208 36-40% Ontario 302 201 77 78 124 59-74% BC 310 206 3 7 7 98-99%
3 Alberta 323 215 203 192 215 34-41% Ontario 302 201 24 24 92 70-92%
57
3.4. Long-run greenhouse gas impacts of PEVs
A primary of objective of this study is to compare short-run perspectives on GHG
intensity of PEVs (g/km), as summarised above, with a long-run perspective that
considers technological dynamics in the transportation and electricity sector. I used the
CIMS model (summarised in Section 2.5) to simulate the effects of three policy
scenarios over the long-run:
1. Reference (Current Policies): existing federal and provincial climate policies affecting the transportation and electricity sectors, e.g. regulations to phase-out coal, low carbon fuel standard, carbon tax, PEV purchase incentives;
2. Carbon Tax: escalating, economy-wide carbon tax starting in 2016 at $25/tonne and increasing to $250/tonne by 2050;
3. Strong Standards (ZEV mandate and Clean electricity): standards that require the deployment of zero-emissions electricity and zero-emissions vehicles over the medium term.
The estimated intensity values are highly dependent on changes to the electricity
generation mix (g/kWh), and to an extent, the shares of various PEV types/designs (e.g.
electric range).
3.4.1. Emissions intensity of electricity
Figure 22 shows the medium to long-term changes in electricity emissions
intensity (gCO2e/kWh) to 2050 in Alberta and Ontario. BC is omitted from this figure
because its power system has very low GHG emissions intensity under all policy
scenarios (<10 g/kWh by 2050). In Alberta, even under the Reference (Current Policies)
Scenario, emissions intensity falls substantially by 2050 (288 g/kWh) to less than one-
third of 2000 levels (949 g/kWh) due to the phase-out of new conventional coal-fired
electricity through federal regulations. My simulation indicates an even greater reduction
in emissions intensity over the near-to-medium-term with a carbon tax compared to a
2025 zero-emissions electricity policy. This may be explained by: i) later implementation
of the zero-emissions electricity policy (2026 compared to the 2016 start date of the
carbon tax); and ii) the zero-emissions electricity policy targeting new generation, rather
than existing generation.
58
Figure 22: GHG emissions intensity of electricity (g CO2e/kWh) in Alberta and Ontario, by policy scenario
In Ontario, due largely to the recent provincial coal phase out (Ending Coal for
Cleaner Air Act), emissions intensity falls by nearly three-quarters from 2000 levels to
50-65 g CO2e/kWh by 2020. More obvious differences between the three Ontario
scenarios emerge post-2020, when emissions intensity begins to rise under the Current
Policies Scenario. This rise is driven primarily by increased use of natural gas-fired
generation to meet increasing demand (in particular, shoulder and peak load demand).
The Ontario ban on coal-fired generation, largely motivated by air quality objectives,
includes a ban on coal with carbon capture and storage (CCS). Under the modelled
Clean Electricity Standard, which restricts new natural gas generation (and all other
fossil fuel generation) after 2025, emissions intensity falls gradually to 46 g CO2e/kWh by
2050 – below that of the Carbon Tax scenario, which remains near 52 g CO2e/kWh over
the 2020-2050 period.
0
100
200
300
400
500
600
700
800
900
1000
2000 2010 2020 2030 2040 2050
GHG
Emiss
ions
Inte
nsity
of E
lectri
city
(g C
O2e/k
Wh)
AB - Reference
AB - Carbon Tax
AB - Standards
ON - Reference
ON - Carbon Tax
ON - Standards
59
3.4.2. Fleet composition
Because I report PEV emissions intensity as a fleet-based average (across a mix
of PHEVs and BEVs of varying electric ranges), PEV fleet composition is a key driver for
differences in PEV emissions intensities. For instance, over time, BEVs and longer-
range PHEVs increase their relative market share (Figure 23) compared to shorter-range
PHEVs, due to declining capital costs and intangible costs (Figure 24). This in turn
increases the electric utility factor and helps to drive further GHG reductions.
Figure 23: Share of PEVs by type and range (left axis) and PEV and hybrid shares of the total vehicle stock (right axis), 2015-2050, BC Reference (Current Policies) Scenario
PEV Share
Hybrid Share
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
100%
2015 2020 2025 2030 2035 2040 2045 2050
Shar
e of a
ll veh
icles
New
Mark
et S
hare
(of P
EVs)
BEV-480
BEV-240
BEV-120
PHEV-64
PHEV-23
PHEV-16
PEV Share
Hybrid Share
60
Figure 24: Levelised costs ($/km) of medium efficiency gasoline, hybrid, PHEV-32, and BEV-240 under the Current Policies Scenario in British Columbia. The initially depressed costs of PEVs are due to the provincial incentive that is removed in 2020.
Empirical data from the survey design exercises and subsequent long-term
modeling results suggest that PHEVs will dominate the PEV market initially due to higher
intangible costs and capital costs of BEVs, some of which decline over time. Regional
differences in PEV types and ranges are also apparent – for instance, longer-range
PHEVs are particularly preferred in Alberta. The evolution of PEV fleet composition (i.e.
PHEV to BEV ratio) differs among the three policy scenarios (Figure 25). Notably, due to
the forced PEV fleet composition (9.25% PHEV and 3.5% BEV in 2025) in the Standards
scenario, the PHEV to BEV ratio declines to less than 3:1 from 2020 onward in BC. In
contrast, the PHEV to BEV ratio increases to almost 8:1 in 2020 under the Current
Policies and Carbon Tax scenarios when the provincial incentives are removed, and
higher capital costs and intangible costs of BEVs take an effect on consumer
preferences among PEVs. This ratio increases further to 11:1 in 2025, until BEV costs
begin to fall and become competitive against PHEVs. Under the Carbon Tax scenario,
the ratio drops to just 2:1 by 2050, driven by lower capital and intangible costs of BEVs
and lower carbon intensity of BEVs.
Gasoline (Med Eff) Hybrid
PHEV-32
BEV-240
$0.00
$0.05
$0.10
$0.15
$0.20
$0.25
$0.30
2005 2010 2015 2020 2025 2030 2035 2040 2045 2050
Leve
lised
Cos
t ($/k
m)
61
Figure 25: Comparison of PHEV to BEV ratio between the three scenarios in British Columbia
3.4.3. Emissions intensity of the PEV fleet
In British Columbia, I find that the policies have a modest effect on emissions
intensity relative to the Reference (Current Policies) scenario, largely due to its already
low emissions electricity grid (Figure 26). The renewable fuel content standard for
petroleum-based transportation fuels (part of the Renewable and Low Carbon Fuel
Requirement Regulation) has an effect in reducing all emissions intensities, including
gasoline and hybrid-electric vehicles. Over the medium-term (to 2030), the PEV
Standard and Clean Electricity Standard achieves the lowest emissions intensity (60
g/km), while the Carbon Tax achieves the lowest intensity by 2050 of 42 g CO2e/km,
69% lower than hybrids and 79% lower than conventional gasoline vehicles.
Current Policies
Carbon Tax Standards
0
2
4
6
8
10
12
2015 2020 2025 2030 2035 2040 2045 2050
PHEV
:BEV
Rat
io
62
Figure 26: Fleet-average short-run and long-run GHG emissions intensity of vehicles in BC. Short-run PEV results are shown for Scenario 1 only.
Alberta is an interesting case to consider because of the substantial reduction in
emissions intensity of electricity by 2050. Figure 27 shows the fleet-average long-run
GHG intensity of vehicles types in Alberta, overlaid with the short-run results for
comparison (Scenario 1: user-designed vehicles + current charging access). PEVs
achieve the deepest emission reductions through the Carbon Tax scenario: in 2035,
PEV emissions intensity under the Carbon Tax scenario is 109 g/km (24% below Current
Policies scenario and 22% below the Standards scenario). By 2050, PEVs under the
Carbon Tax scenario (61 g/km) achieve an emissions intensity of less than one-quarter
of the level of conventional vehicles in that year (270 g/km). Across all policy scenarios,
the fleet of PEVs starts to offer obvious GHG benefits (compared to hybrid-electric
vehicles) after 2025. While the emissions intensity of electricity is similar between the
Carbon Tax and Standards scenarios in 2045-2050, the Carbon Tax scenario results in
26% lower PEV emissions intensity in 2050 compared to the Standards scenario. This is
due to the higher share of pure-electric BEVs among PEVs (27% under Carbon Tax
compared to 10% under Standards) using low-carbon electricity. By 2050, PEV
emissions intensity is 61-94 g/km in Alberta, 65-77% below conventional vehicles and
48-66% below HEVs.
Short-run, Gasoline
Short-run, Hybrid
Short-run, PEV
0
50
100
150
200
250
300
350
2015 2020 2025 2030 2035 2040 2045 2050
Emiss
ions
inte
nsity
(g
CO2
e/km)
Long-run, Gasoline
Long-run, Hybrid
Long-run, PEV (Current Policies)
Long-run, PEV (Standards)
Long-run, PEV (Carbon Tax)
63
Figure 27: Fleet-average short-run and long-run GHG emissions intensity of vehicles in Alberta. Short-run PEV results are shown for Scenario 1 only.
Figure 28 shows the results for Ontario, where the provincial phase-out of coal
has had a major effect in decreasing emissions intensity of electricity in the province in
the decade prior to the simulation period. The simulated policies have limited impact on
further reductions in emissions intensity of the grid past 2020. The “Standards” scenario
in particular has limited impact because its design (which restricts new fossil fuel
generation past 2025) is unable to reduce emissions from the existing stock of natural
gas fired generation. The Carbon Tax achieves the lowest GHG emissions intensity by
2050 of 56 g/km, 69% lower than hybrids and 79% lower than conventional gasoline
vehicles.
Short-run, Gasoline
Short-run, Hybrid
Short-run, PEV
0
50
100
150
200
250
300
350
2015 2020 2025 2030 2035 2040 2045 2050
Emiss
ions
inte
nsity
(g
CO2
e/km)
Long-run, Gasoline
Long-run, Hybrid
Long-run, PEV (Current Policies)
Long-run, PEV (Standards)
Long-run, PEV (Carbon Tax)
64
Figure 28: Fleet-average short-run and long-run GHG emissions intensity of vehicles in Ontario. Short-run PEV results are shown for Scenario 1 only.
It is important to note the decreasing emissions intensity of baseline vehicles
(conventional and hybrid electric) over time as more high-efficient conventional vehicles
are introduced over the short to medium-term, replacing less efficient conventional
vehicles that are retired. Because only one hybrid archetype exists in CIMS, I elected to
benchmark the emissions intensity of hybrids at two-thirds of the fleet-average
conventional vehicle to reflect improving efficiency over time.
3.4.4. Emissions intensity of PHEVs and BEVs
To isolate differences in intensity of PHEVs and BEVs, Figure 29 compares the
differences in emissions intensity between a PHEV-32 and a BEV in the three regions.
Under the reference scenario, we see that BEVs in Alberta have the highest emissions
intensity (233 g/km) in 2015, but decline to levels below PHEVs in all regions by 2050
(79 g/km). In Alberta, emissions intensity of a BEV falls below that of a PHEV-32 in the
region after around 2035. This crossover happens slightly sooner in the policy scenarios
due to a more rapid decarbonisation of the power sector; in the Carbon Tax Scenario,
the crossover occurs between 2025 and 2030 (Figure 30). BEVs have substantially
Short-run, Gasoline
Short-run, Hybrid
Short-run, PEV
0
50
100
150
200
250
300
350
2015 2020 2025 2030 2035 2040 2045 2050
Emiss
ions
inte
nsity
(g
CO2
e/km)
Long-run, Gasoline
Long-run, Hybrid
Long-run, PEV (BAU)
Long-run, PEV (Standards)
Long-run, PEV (Carbon Tax)
65
lower emissions in British Columbia and Ontario compared to PHEV-32 in all years
simulated (93-96% in British Columbia; 57% to 82% in Ontario). As expected, policies
that decarbonise the electricity sector have a greater effect on BEV emissions intensity
compared to PHEVs.
Figure 29: Emissions intensity of PHEV-32 and BEV under the Reference scenario
AB-PHEV-32
ON-PHEV-32 BC-PHEV-32 AB-BEV
ON-BEV BC-BEV 0
50
100
150
200
250
2015 2020 2025 2030 2035 2040 2045 2050
Emiss
ions
inte
nsity
(g
CO2
e/km)
66
Figure 30: Emissions intensity of PHEV-32 and BEV in Alberta under the Reference (BAU) and Carbon Tax scenarios
PHEV-32 BAU
PHEV-32 Carbon Tax
BEV BAU
BEV Carbon Tax
0
50
100
150
200
250
2015 2020 2025 2030 2035 2040 2045 2050
Emiss
ions
inte
nsity
(g
CO2
e/km)
67
Chapter 4. Discussion
This analysis considered two perspectives: a detailed, static analysis over the
short-term and a broader, dynamic analysis over the longer term to answer three
research objectives introduced in Section 1.3:
1. Develop a behaviourally realistic model of the short-term greenhouse gas (GHG) emissions impacts of plug-in electric vehicles using behavioural data from the Canadian Plug-in Electric Vehicle Study (Axsen et al., 2015);
2. Use insights from the short-term model to model long-term technological dynamics of the transport and power sectors and resulting GHG emissions intensity of plug-in electric vehicles (PEVs) to 2050; and
3. Compare GHG intensity results from three different Canadian regions (British Columbia, Alberta, Ontario), and demonstrate the effects of climate policy on PEV GHG intensity.
This chapter summarises my key results, compares these values to the literature,
and extends the insights from the short and long-term approaches to a policy-maker’s
context. I also discuss the key limitations of this work and its implications in interpreting
the study’s findings.
4.1. Short-term emissions intensity of PEVs
The short-term model, due to its use of rich behavioural and technical data, fine
temporal resolution, and consideration of marginal generation and electricity trade,
reveals important micro-level considerations when estimating the GHG impacts of PEVs
such as charge timing and recharge availability.
Fleet-wide GHG emissions intensity of PEVs varies substantially between the
three regions studied (Table 16). I find that over the short-term, British Columbia has the
greatest emission reduction potential (78-99%) compared to conventional vehicles due
to its low carbon power system. Under the baseline scenario, I estimate the fleet
68
average PEV emits 65-67 gCO2e/km, which lies in the middle of the Pembina Institute’s
range of estimates from 11 gCO2e/km (BEV) to 111 gCO2e/km (PHEV-50) (Moorhouse
& Laufenberg, 2010). Adjusting Moorhouse & Laufenberg's (2010) estimate to the
predominantly PHEV fleet composition in my sample (89% PHEV, 11% BEV) would
result in a fleet average PEV emissions intensity of 100 gCO2e/km. This difference may
be a result of different behavioural assumptions (e.g. lower electric utility factor based on
the electric range of PHEVs and how PEVs are driven and charged) and motor efficiency
assumptions.
Despite Alberta’s coal and gas-based generation system, I find that a fleet of
PEVs in Alberta emits 189-215 g/km and can therefore reduce emissions by at least
one-third (34-41%) compared to conventional vehicles and up to 12% (0.3-12%)
compared to HEVs. The reductions are negligible for a fully battery-electric (BEV) fleet
(Scenario 3) and when allocating emissions using hourly marginal emissions factors.
Some previous studies estimate that PEV usage results in higher GHG emissions than
HEVs in Alberta (Ribberink & Entchev, 2013) and coal-based regions like Alberta (e.g.
Samaras & Meisterling, 2008). The primary cause for this difference can be traced to
differences in the lifecycle GHG emission factor for gasoline. I assume a 50/50 oil sands
and conventional oil blend and use lifecycle emission factors from recent literature. The
emission factor I use in this study (3,461 g/L) is therefore 16% higher than the emission
factor used in Samaras & Meisterling (2008).
In Ontario, I estimate an average fleet-wide PEV emissions intensity of 124 g/km
under Scenario 1 (using hourly marginal emission factors), which represents the most
likely short-term scenario. This represents a reduction of 59% compared to conventional
vehicles, which is approximately mid-way between the emissions intensity range
(depending on driving conditions) estimated by Raykin et al.'s (2012) Ontario-based
analysis: 60-170 g/km representing a 24-80% reduction. Using annual average emission
factors, I find that emissions intensity of 77-88 g/km for a mixed PEV fleet and 24 g/km
for BEVs. My results are slightly higher than the results in Plug’n Drive (2015) 65 g/km
(PHEVs) and 12 g/km (BEVs).
69
The differences between the three regions studied can be largely attributed to
differences in the emissions intensity of the electricity supply, differences in vehicle
preferences (e.g. larger vehicles and longer-range PHEVs in Alberta), driving patterns
(e.g. longer distances in Ontario), and recharge access (e.g. lower access in BC).
One interesting aspect of this study is its use of multiple electricity emissions
allocation approaches: hourly marginal, hourly average, and annual average. Given the
frequent use of annual average emission factors in literature, I compare the two other
approaches to this common approach. The use of hourly average emission factors,
which considers electricity trade, estimates slightly higher emissions intensity in British
Columbia (4%) and Ontario (0.5%) and 4% lower emissions intensity in Alberta under an
unconstrained charging scenario (evening peak). The use of hourly marginal emission
factors results in slightly higher emissions in British Columbia (4%) and Alberta (5%),
and much higher emissions intensity in Ontario (44%). In the early evening when a large
proportion of charging is occurring, Ontario’s peak generation being supplied by
emissions-intense natural gas.
In comparing the results the three recharge scenarios, introducing workplace
Level 2 charging access (Scenario 2) has a minor effect on emissions intensity in Alberta
(2-3% increase) and Ontario (2-12% decrease) while in BC, emissions are reduced by
21-22%. Workplace charging would peak from around 7-9AM; this corresponds with
higher hourly marginal emissions in Alberta compared to the evening. A slower charging
rate (Level 1) in Alberta could reduce emissions in this scenario by using lower
emissions electricity during the late morning. In Ontario, the highest marginal emissions
occur during the workday, dampening the emissions benefits of a higher electric utility
factor. Scenario 3, a high electrification scenario (BEV-240 with home and work Level 2
access), results in the lowest emissions intensity in all three regions. GHG emissions
intensity is particularly low in BC (3-7 g/km) and Ontario (24-92 g/km).
4.2. Long-term emissions intensity of PEVs
In contrast to the static short-term model, the long-run (CIMS) model captures
technology dynamics over the longer term, such as declining capital and intangible costs
70
of PEVs and the decarbonisation of the power sector. The results highlight the
importance of reducing the GHG intensity of the electricity grid to complement PEV
market penetration to maximize GHG emission reduction from passenger vehicles.
As with the short-term results, long-term fleet-wide GHG emissions intensity of
PEVs varies substantially between the three regions (Figure 29), largely a reflection of
differences in the emissions intensity of electricity (Figure 22). The emissions intensity of
electricity in Ontario and Alberta decreases by at least two-thirds by 2050 (from 2000
levels) due to the phase out of conventional coal-fired electricity through federal
regulations and Ontario’s Ending Coal for Cleaner Air Act. Emissions intensity of
electricity in 2050 is 8 g/kWh in BC, 76 g/kWh in Ontario, and 288 g/kWh in Alberta in
the Reference scenario. The two alternative climate policy scenarios result in differential
impacts on the emissions intensity of the three types of load: base, shoulder, and peak.
Generally, while the Carbon Tax and Standards scenarios result in similar average
emissions intensities in 2050, the Carbon Tax has a greater effect on reducing base load
emissions intensity while the Standard has a greater effect on shoulder and peak load
emissions intensity. Therefore, alternative PEV demand profiles (e.g. overnight charging
vs. evening peak charging) would result in substantially different PEV emissions impacts
under similar average electricity emissions intensities.
As expected, emissions intensity of PEVs over time largely follows the trends in
emissions intensity of electricity. By 2050, fleet average PEV emissions are 42-94 g/km,
representing reductions of 65-84% below conventional vehicles and 48-76% below
HEVs across all policy scenarios and regions. These results are comparable with the
most recent modelling results from EPRI (2015) which estimated fleet average PEV
emissions of 37-53 g/km in 2050 in the United States (100-340 g/kWh), representing a
reduction of 58-70% below efficient conventional vehicles. In Alberta, PEVs and HEVs
have comparable emissions intensity until 2025, but decarbonisation of the grid and an
increased share of BEVs further decreases emissions intensity by 2050 to as low as
66% below HEVs in the Carbon Tax scenario. Climate policies have a smaller effect on
PEV emissions intensity in BC and Ontario. Changes to the PEV fleet composition
towards longer-range PHEVs and BEVs (due to declining capital and intangible costs)
drives the decrease in emissions intensity towards 2050, and are lowest under the
71
Carbon Tax scenario due to the higher share of BEVs compared to the Standards
scenario.
4.3. Policy implications
The results show a large range in PEV emissions intensity between regions,
across scenarios, and over time. While the short-term analysis offers useful insights
supported by rich empirical data, policy makers are encouraged to focus on long-term
implications of policy given the long operational lifetimes of vehicles and power plants.
The results from both the short-term and long-term perspectives highlight the
importance of reducing the emissions intensity of the power sector to complement PEV
market penetration to maximise emission reductions from passenger vehicles over time.
While targeted standards and regulations of the power and transport sectors can have
the desired effect on PEV emissions intensity, I find that electricity regulations in
particular should focus on targeting existing stock in addition to new generation to have
greater impacts in the short to medium-term (i.e. forcing early retirement of coal plants).
Ontario’s coal phase-out provides a good example of how quickly this type of phase-out
can reduce electricity emissions intensity. Regions with emissions intensive power
sectors such as Alberta can still achieve emission reductions through PEV deployment,
but must focus on decarbonising their grids to maximise emission reduction of PEVs.
The recent announcement from the Government of Alberta (2015) to phase out all coal
by 2030 makes a further case for electric vehicles in this region.
4.4. Limitations
While this study attempts to address critiques of previous research in this area,
this analysis also contains several important limitations.
In terms of emissions scope, this study uses a source-to-wheels approach that
estimates operational (use-phase) emissions, and therefore does not consider emissions
from vehicle manufacture and disposal, including battery disposal, which would be
72
covered by a lifecycle analysis (LCA). A review of 51 LCA studies of PEVs found that
use-phase emissions accounts for approximately 60-90% of the total lifecycle GHG
emissions, with battery production accounting for about 12 g/km (Hawkins et al., 2012).
Lucas et al. (2012) conducted an LCA of energy supply infrastructure for vehicles (e.g.
gasoline fuelling stations, EV chargers) and found that energy supply infrastructure
accounts for no more than 8% of total vehicle lifecycle GHG impacts. Therefore, on a
lifecycle basis, PEVs are likely to have slightly lower emission reductions compared to
the source-to-wheels results reported in this study.
In my short-term models, I estimated emissions intensity of PEVs using average
annual, hourly average, and hourly marginal emission factors from historical generation
data. While the use of multiple approaches and historical data improves realism and
certainty, historical generation may not necessarily reflect future near-term generation
patterns. For example, the electricity generation mix in Ontario has changed
considerably in three years with the coal phase-out, and emissions intensity has been
nearly halved since 2010 (Environment Canada, 2015). An alternative, but considerably
more complex approach would be to build an electricity supply model that considers
capacity factors of existing power plants as well as proposed/planned plants, e.g.
through the use of a dispatch model.
I also implicitly assume that current driving patterns (in conventional vehicles)
accurately reflect the driving patterns of the same respondents if they were to buy a
PEV. Given the limited number of existing PEV users, it is not yet clear how driving
behaviours may actually differ between conventional vehicles and PEVs. On one hand,
the cheaper cost of operating a PEV could result in a rebound effect, resulting in
increased driving activity. However, a study of new Toyota Prius HEV buyers in
Switzerland found no evidence of rebound effects (de Haan, Mueller, & Peters, 2006).
Another idea is that with range-limited or low-carbon vehicles, drivers may be prompted
to reduce total driving distance or improve driving efficiency, reducing overall energy
consumption and GHG emissions. For example, some drivers in a PHEV trial reported
altering their driving habits (e.g. acceleration rate) to improve fuel economy, viewing
driving a PHEV as a “game” with objectives like maximizing electric-powered range
(Caperello & Kurani, 2011). To verify the representativeness of typical driving patterns,
73
respondents were shown a summary of their driving diary. 81% reported that their diary
was typical.
While I use empirical data on driving, I do not consider driving behaviours or
conditions (speed, acceleration, urban/rural, weather) that could impact efficiency
assumptions. For instance, cold weather can have substantial impacts on the battery
and affect electricity consumption and utility factor (Zahabi et al., 2014), while the use of
air conditioning can reduce electric range (Ma et al., 2012; Rangaraju et al., 2015).
Finally, there are numerous limitations of my long-term model. At the macro-
level, rapidly changing energy markets (e.g. low oil prices, faster than anticipated cost
reductions in renewables) and emerging climate policies could evolve in ways that
diverge from the modeled scenarios. For instance, in the lead-up to 21st UNFCCC
Conference of the Parties (COP21) in Paris, nearly all countries submitted national
pledges (intended nationally determined contributions), with many also announcing
supporting climate change policies. In Canada, major climate change strategies were
announced in Alberta, and the new federal government elected in October is likely to
change the landscape of climate policy in Canada. While this project does not
incorporate these latest and emerging policies, these policies will only serve to further
the case for electric vehicles in Canada. From a more detailed, methodological
perspective, a key limitation of my model is its inability to account for inter-regional
electricity trade, and the potential for alternative PEV recharge profiles (e.g. overnight
charging, vehicle-to-grid).
74
Chapter 5. Conclusions and future research
Greenhouse gas emissions impacts of PEVs vary between regions, across
scenarios, and over time. This work combined two complementary perspectives: a
detailed, short-term perspective supported by rich empirical data, and a long-term
analysis of the evolution of technologies in the electricity and personal transportation
sectors. Over the short-term, substantial emission reductions are observed in British
Columbia (78-99%), Ontario (58-92%), and Alberta (34-41%) relative to conventional
(gasoline) vehicles. Over the long-term, emissions intensity of electricity decreases by at
least one-third by 2050. Consequently, fleet average PEV emissions are 23-40% (BC),
51-68% (Alberta), and 25-40% (Ontario) below 2015 levels in 2050, and about a quarter
of the emissions of efficient conventional vehicles.
Despite the large range of emissions intensities between regions and over time,
PEVs offer substantial GHG emissions benefits compared to conventional vehicles.
Therefore, policy makers should look to design policies that concurrently promote
vehicle electrification and decarbonisation of the electricity supply to help achieve long-
term mitigation targets.
Future research could build on this and other work by delving deeper into a
number of areas. First, this work uses emissions intensity of PEV (g/km) as the metric
for comparison. While this provides a useful frame to compare emissions impacts of
different vehicle types (and over time), it fails to account for the cumulative emissions
impact of PEVs over time, an arguably more important indicator from a climate goal
perspective (i.e. planetary carbon budget). Therefore, improved market penetration
modeling, such as expanding on the work of Axsen & Wolinetz (2015) to other regions,
is critical in translating emissions intensities into cumulative GHG reductions over time,
especially when comparing a range of policy scenarios over the longer term. Second,
improved understanding of intangible costs of PEV types and models may help to
75
improve market penetration modelling and the potential fleet-wide emission reduction
potential of PEVs. Third, extending the models to include data from existing PEV users,
for instance, incorporating real-world electric utility factors, could improve behavioural
realism. Finally, developing additional recharge scenarios that consider vehicle users’
willingness to adopt utility-controlled charging could improve understanding of the
emission reduction potential of alternative recharging scenarios, e.g. overnight charging,
time-of-use charging. For example, Rangaraju et al. (2015) found that time-of-use
charging of BEVs could reduce source-to-wheel CO2 emissions by 13% in Belgium.
76
References
Axhausen, K. W. (1995). Travel Diaries: An Annotated Catalogue.
Axsen, J., Bailey, H., & Kamiya, G. (2013). The Canadian Plug-in Electric Vehicle Survey (CPEVS 2013): Anticipating Purchase, Use, and Grid Interactions in British Columbia. Retrieved from http://rem-main.rem.sfu.ca/papers/jaxsen/CPEVS-2013_BC_Preliminary_Report_Oct-31-13.pdf
Axsen, J., Goldberg, S., Bailey, J., Kamiya, G., Langman, B., Cairns, J., … Miele, A. (2015). Electrifying Vehicles: Insights from the Canadian Plug-in Electric Vehicle Study. Vancouver, Canada.
Axsen, J., & Kurani, K. S. (2008). The early US market for PHEVs: Anticipating consumer awareness, recharge potential, design priorities and energy impacts. UC Davis, (August). Retrieved from http://escholarship.org/uc/item/4491w7kf.pdf
Axsen, J., & Kurani, K. S. (2009). Early U.S. Market for Plug-In Hybrid Electric Vehicles. Transportation Research Record: Journal of the Transportation Research Board, 2139(-1), 64–72. doi:10.3141/2139-08
Axsen, J., & Kurani, K. S. (2010). Anticipating plug-in hybrid vehicle energy impacts in California: Constructing consumer-informed recharge profiles. Transportation Research Part D: Transport and Environment, 15(4), 212–219. doi:10.1016/j.trd.2010.02.004
Axsen, J., & Kurani, K. S. (2013). Hybrid, plug-in hybrid, or electric—What do car buyers want? Energy Policy, 61, 532–543. doi:10.1016/j.enpol.2013.05.122
Axsen, J., Kurani, K. S., McCarthy, R., & Yang, C. (2011). Plug-in hybrid vehicle GHG impacts in California: Integrating consumer-informed recharge profiles with an electricity-dispatch model. Energy Policy, 39(3), 1617–1629. doi:10.1016/j.enpol.2010.12.038
Axsen, J., Mountain, D. C., & Jaccard, M. (2009). Combining stated and revealed choice research to simulate the neighbor effect: The case of hybrid-electric vehicles. Resource and Energy Economics, 31(3), 221–238. doi:10.1016/j.reseneeco.2009.02.001
Axsen, J., & Wolinetz, M. (2015). How policy can build the plug-in electric vehicle market: Insights from the Respondent-Based Preferences and Constraints
77
(REPAC) model, 6926.
Busse, M. R., Knittel, C. R., & Zettelmeyer, F. (2013). Are consumers myopic? Evidence from new and used car purchases. American Economic Review, 103(1), 220–256. doi:10.1257/aer.103.1.220
Butcher, R., & Eldridge, J. (1990). The Use of Diaries in Data Collection. Journal of the Royal Statistical Society. Series D (The Statistician), 39(1), 25–41. doi:10.2307/2348192
Cai, H., Brandt, A. R., Yeh, S., Englander, J. G., Han, J., Elgowainy, A., & Wang, M. Q. (2015). Well-to-Wheels Greenhouse Gas Emissions of Canadian Oil Sands Products: Implications for U.S. Petroleum Fuels. Environmental Science & Technology, 150625082709001. doi:10.1021/acs.est.5b01255
California Air Resources Board. (2012). The Zero Emission Vehicle (ZEV) Regulation. Sacramento, CA.
Canadian Association of Petroleum Producers. (2015). Crude Oil: Forecast, Markets & Transportaton. Calgary, AB.
Caperello, N. D., & Kurani, K. S. (2011). Households’ Stories of Their Encounters With a Plug-In Hybrid Electric Vehicle. Environment and Behavior. doi:10.1177/0013916511402057
de Haan, P., Mueller, M. G., & Peters, A. (2006). Does the hybrid Toyota Prius lead to rebound effects? Analysis of size and number of cars previously owned by Swiss Prius buyers. Ecological Economics, 58(3), 592–605. doi:10.1016/j.ecolecon.2005.08.009
Dillman, D. A., Smyth, J. D., & Christian, L. M. (2009). Internet, Mail and Mixed-Mode Surveys: The Tailored Design Method (3rd ed.). Hoboken, NJ: John Wiley.
Dong, J., & Lin, Z. (2012). Within-day recharge of plug-in hybrid electric vehicles: Energy impact of public charging infrastructure. Transportation Research Part D: Transport and Environment, 17(5), 405–412. doi:10.1016/j.trd.2012.04.003
Doucette, R. T., & McCulloch, M. D. (2011). Modeling the prospects of plug-in hybrid electric vehicles to reduce CO2 emissions. Applied Energy, 88(7), 2315–2323. doi:10.1016/j.apenergy.2011.01.045
Duvall, M., Knipping, E., Alexander, M., Tonachel, L., & Clark, C. (2007). Environmental Assessment of Plug-In Hybrid Electric Vehicles. Volume 1: Nationwide Greenhouse Gas Emissions. Electric Power Research Institute, Palo Alto, CA, 1015325.
Elgowainy, A., Burnham, A., Wang, M., Molburg, J., & Rousseau, A. (2009). Well-to-Wheels Energy Use and Greenhouse Gas Emissions Analysis of Plug-in Hybrid
78
Electric Vehicles. Argonne National Laboratory. Retrieved from http://www.osti.gov/bridge/product.biblio.jsp?osti_id=951259
Environment Canada. (2014). National Inventory Report.
Environment Canada. (2015). National Inventory Report: Greenhouse Gas Sources and Sinks in Canada (1990-2013). Gatineau, Quebec.
EPRI. (2015). Environmental Assessment of a Full Electric Transportation Portfolio: Volume 2: Greenhouse Gas Emissions (Vol. 3002006876). Palo Alto, CA. doi:10.1017/CBO9781107415324.004
Fox, J. (2013). Picking winners: assessing the costs of technology-specific climate policy for US passenger vehicles. Retrieved from http://summit.sfu.ca/item/13533
Garcia, R., Gregory, J., & Freire, F. (2015). Dynamic fleet-based life-cycle greenhouse gas assessment of the introduction of electric vehicles in the Portuguese light-duty fleet. The International Journal of Life Cycle Assessment, 20(9), 1287–1299. doi:10.1007/s11367-015-0921-8
Government of Alberta. (2015). Climate Leadership Plan: Coal and electricity. Retrieved December 13, 2015, from http://alberta.ca/climate/coal-electricity.cfm
Harris-Decima. (2013). What does 2013 hold for the auto industry in Canada?
Hawkins, T. R., Gausen, O. M., & Strømman, A. H. (2012). Environmental impacts of hybrid and electric vehicles-a review. International Journal of Life Cycle Assessment, 17(8), 997–1014. doi:10.1007/s11367-012-0440-9
International Energy Agency. (2014a). CO2 Emissions from Fuel Combustion (2014th ed.). Paris. Retrieved from www.iea.org/statistics/
International Energy Agency. (2014b). World Energy Outlook 2014. Paris: International Energy Agency. doi:10.1787/weo-2014-en
International Energy Agency. (2015a). Clean energy technologies. Retrieved September 1, 2015, from http://www.iea.org/topics/cleanenergytechnologies/
International Energy Agency. (2015b). Energy Technology Perspectives 2015. Paris: International Energy Agency.
ITU. (2015). Percentage of Individuals using the Internet. Retrieved August 17, 2015, from http://www.itu.int/en/ITU-D/Statistics/Documents/statistics/2015/Individuals_Internet_2000-2014.xls
Jaccard, M. (2009). Combining top-down and bottom-up in energy economy models. In
79
International Handbook on the Economics of Energy (pp. 311–331).
Jaramillo, P., Samaras, C., Wakeley, H., & Meisterling, K. (2009). Greenhouse gas implications of using coal for transportation: Life cycle assessment of coal-to-liquids, plug-in hybrids, and hydrogen pathways. Energy Policy, 37(7), 2689–2695. doi:10.1016/j.enpol.2009.03.001
Jochem, P., Babrowski, S., & Fichtner, W. (2015). Assessing CO2 emissions of electric vehicles in Germany in 2030. Transportation Research Part A: Policy and Practice, 78, 68–83. doi:10.1016/j.tra.2015.05.007
Kniewasser, M. (2014). Achieving Canada’s climate targets and the impacts on Alberta's oil sands industry.
Lattanzio, R. K. (2014). Canadian Oil Sands: Life-Cycle Assessments of Greenhouse Gas Emissions. Congressional Research Service, 31.
Lucas, A., Alexandra Silva, C., & Costa Neto, R. (2012). Life cycle analysis of energy supply infrastructure for conventional and electric vehicles. Energy Policy, 41, 537–547. doi:10.1016/j.enpol.2011.11.015
Ma, H., Balthasar, F., Tait, N., Riera-Palou, X., & Harrison, A. (2012). A new comparison between the life cycle greenhouse gas emissions of battery electric vehicles and internal combustion vehicles. Energy Policy, 44, 160–173. doi:10.1016/j.enpol.2012.01.034
Mallory, J. (2007). Assessing policies for reducing greenhouse gas emissions from passenger vehicles. Simon Fraser University.
Mau, P., Eyzaguirre, J., Jaccard, M., Collins-Dodd, C., & Tiedemann, K. (2008). The “neighbor effect”: Simulating dynamics in consumer preferences for new vehicle technologies. Ecological Economics, 68(1-2), 504–516. doi:10.1016/j.ecolecon.2008.05.007
McCollum, D., Krey, V., Kolp, P., Nagai, Y., & Riahi, K. (2013). Transport electrification: A key element for energy system transformation and climate stabilization. Climatic Change, 123(3-4), 651–664. doi:10.1007/s10584-013-0969-z
Moomaw, W., Burgherr, P., Heath, G., Lenzen, M., Nyboer, J., & Verbruggen, A. (2014). Annex II: Methodology. IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation (Vol. 16). doi:10.1093/ntr/ntt227
Moorhouse, J., & Laufenberg, K. (2010). Backgrounder Electric Vehicles Powering the Future, (September).
Murphy, R. (2000). Analysis of measures for reducing transportation sector greenhouse gas emissions in Canada, (271). Retrieved from
80
http://www.emrg.sfu.ca/media/publications/Rose Murphy 699.pdf
Natural Resources Canada. (2011). Canadian Vehicle Survey: 2009 Summary Report.
Nealer, R., & Hendrickson, T. P. (2015). Review of Recent Lifecycle Assessments of Energy and Greenhouse Gas Emissions for Electric Vehicles. Current Sustainable/Renewable Energy Reports. doi:10.1007/s40518-015-0033-x
Nguyen, T., Ward, J., & Johnson, K. (2013). Well-to-Wheels Greenhouse Gases Emissions and Petroleum Use for Mid-Size Light-Duty Vehicles (Vol. 13005).
Nyboer, J. (2014). Energy Use and Related Data: Canadian Electricity Generation Industry 1990 to 2012. Burnaby, BC. Retrieved from http://www2.cieedac.sfu.ca/media/publications/Electricity_Report_2013__2012__Final.pdf
Nykvist, B., & Nilsson, M. (2015). Rapidly falling costs of battery packs for electric vehicles. Nature Climate Change, 5(April), 100–103. doi:10.1038/nclimate2564
Onat, N. C., Kucukvar, M., & Tatari, O. (2015). Conventional, hybrid, plug-in hybrid or electric vehicles? State-based comparative carbon and energy footprint analysis in the United States. Applied Energy, 150, 36–49. doi:10.1016/j.apenergy.2015.04.001
Ontario Ministry of Energy. (2013). 2013 LTEP Consolidated Figures and Data Tables List of Figures. Retrieved from http://www.energy.gov.on.ca/docs/en/LTEP_2013_Consolidated_Figures_and_Data_Tables.pdf
Plötz, P., Funke, S., & Jochem, P. (2015). Real-world fuel economy and CO2 emissions of plug-in hybrid electric vehicles. Retrieved from http://www.econstor.eu/handle/10419/106656
Plug’n Drive. (2015). Electric Vehicles: Reducing Ontario’s Greenhouse Gas Emissions, (May).
Rangaraju, S., De Vroey, L., Messagie, M., Mertens, J., & Van Mierlo, J. (2015). Impacts of electricity mix, charging profile, and driving behavior on the emissions performance of battery electric vehicles: A Belgian case study. Applied Energy, 148(X), 496–505. doi:10.1016/j.apenergy.2015.01.121
Raykin, L., MacLean, H. L., & Roorda, M. J. (2012). Implications of driving patterns on well-to-wheel performance of plug-in hybrid electric vehicles. Environmental Science and Technology, 46(11), 6363–6370. doi:10.1021/es203981a
Ribberink, H., & Entchev, E. (2013). Electric vehicles – a “one-size-fits-all” solution for
81
emission reduction from transportation? In EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium (pp. 1–7). IEEE. doi:10.1109/EVS.2013.6914837
Rolim, C. C., Gonçalves, G. N., Farias, T. L., & Rodrigues, Ó. (2012). Impacts of Electric Vehicle Adoption on Driver Behavior and Environmental Performance. Procedia - Social and Behavioral Sciences, 54, 706–715. doi:10.1016/j.sbspro.2012.09.788
Samaras, C., & Meisterling, K. (2008). Life cycle assessment of greenhouse gas emissions from plug-in hybrid vehicles: implications for policy. Environmental Science & Technology, 42(9), 3170–6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18522090
Santos, A., McGuckin, N., Nakamoto, H. Y., Gray, D., & Liss, S. (2011). Summary of Travel Trends: 2009 National Household Travel Survey. Washington, D.C. Retrieved from http://nhts.ornl.gov/2009/pub/stt.pdf
Shen, W., Han, W., Chock, D., Chai, Q., & Zhang, A. (2012). Well-to-wheels life-cycle analysis of alternative fuels and vehicle technologies in China. Energy Policy, 1–12. doi:10.1016/j.enpol.2012.06.038
Siler-Evans, K., Azevedo, I. L., & Morgan, M. G. (2012). Marginal emissions factors for the u.s. Electricity system. Environmental Science & Technology, 46(9), 4742–8. doi:10.1021/es300145v
Silva, C., Ross, M., & Farias, T. (2009). Evaluation of energy consumption, emissions and cost of plug-in hybrid vehicles. Energy Conversion and Management, 50(7), 1635–1643. doi:10.1016/j.enconman.2009.03.036
Statistics Canada. (2010). Canadian Vehicle Survey : Annual.
Stopher, P. R. (2008). The Travel Survey Toolkit: Where to from here? In 8th International Conference on Travel Survey Methods.
Toronto Hydro. (2011). Responses to Interrogatories of Vulnerable Energy Consumers Coalition.
US EPA. (2014). eGRID 9th edition Version 1.0 Year 2010 Summary Tables, (February), 0–12.
Weiller, C. (2011). Plug-in hybrid electric vehicle impacts on hourly electricity demand in the United States. Energy Policy, 39(6), 3766–3778. doi:10.1016/j.enpol.2011.04.005
Williams, J. H., DeBenedictis, A., Ghanadan, R., Mahone, A., Moore, J., Morrow, W. R., … Torn, M. S. (2012). The technology path to deep greenhouse gas emissions cuts by 2050: the pivotal role of electricity. Science, 335(6064), 53–9.
82
doi:10.1126/science.1208365
Wolinetz, M., & Axsen, J. (2014). Research Brief : Plug-in Electric Vehicle Sales in British Columbia to 2020 Introduction and Summary of the Methodology.
WRI. (2005). Guidelines for Quantifying GHG Reductions from Grid-Connected Electricity Projects. Retrieved from http://www.ghgprotocol.org/files/ghgp/electricity_final.pdf
Yang, C. (2013). A framework for allocating greenhouse gas emissions from electricity generation to plug-in electric vehicle charging. Energy Policy, 60, 722–732. doi:10.1016/j.enpol.2013.05.013
Zahabi, S. A. H., Miranda-Moreno, L., Barla, P., & Vincent, B. (2014). Fuel economy of hybrid-electric versus conventional gasoline vehicles in real-world conditions: A case study of cold cities in Quebec, Canada. Transportation Research Part D: Transport and Environment, 32, 184–192. doi:10.1016/j.trd.2014.07.007
83
Appendix A: Assumptions
Table A-1: Vehicle design space exercise options and incremental (upgrade) costs
Higher price Lower price Vehicle type Compact Sedan Mid-SUV Full-SUV Compact Sedan Mid-SUV Full-SUV HEV $1 380 $1 740 $2 050 $2 470 $930 $1 070 $1 200 $1 370 PHEV-16 $2 230 $2 720 $3 130 $3 690 $1 690 $1 910 $2 100 $2 360 PHEV-32 $2 680 $3 230 $3 810 $4 500 $1 910 $2 170 $2 440 $2 770 PHEV-64 $3 560 $4 260 $5 190 $6 120 $2 350 $2 680 $3 130 $3 580 BEV-80 $6 500 $7 880 $10 150 $12 150 $3 220 $3 620 $4 600 $5 300 BEV-120 $8 940 $10 690 $13 930 $16 600 $4 440 $5 030 $6 490 $7 520 BEV-160 $11 380 $13 500 $17 710 $21 050 $5 660 $6 440 $8 380 $9 750 BEV-200 $13 820 $16 310 $21 490 $25 500 $6 880 $7 840 $10 270 $11 970 BEV-240 $16 260 $19 130 $25 260 $29 940 $8 100 $9 250 $12 160 $14 200
Table A-2: Price model for Level 2 installation
Obstacle Cost Base cost: distance from parking spot to electricity panel <25 feet $1,000 26-50 feet $1,500 > 50 feet $2,000 Additional costs: obstacles Multiple walls + $500 Paved space + $500 Building floors + $500