Emission Impacts of Electric Vehicles in the US Transportation Sector

Subscriber access provided by Fudan University

Policy Analysis

Emission Impacts of Electric Vehicles in the US Transportation Sector Following Optimistic Cost and Efficiency Projections Azadeh Keshavarzmohammadian, Daven K. Henze, and Jana B Milford Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b04801 • Publication Date (Web): 11 Apr 2017 Downloaded from http://pubs.acs.org on April 15, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

Environmental Science & Technology

1

Emission Impacts of Electric Vehicles in the US

2

Transportation Sector Following Optimistic Cost

3

and Efficiency Projections

4

Azadeh Keshavarzmohammadian†, Daven K. Henze†, Jana B. Milford*,†

5

†

6

80309-0427, United States

7

KEYWORDS: Electric vehicles, MARKAL, EPA US 9-region database, well-to-wheels analysis

8

TOC ART

9

10

2010

Compact ICEV

Compact BEV200

Use Phase Processing Upstream

Compact BEV100

Compact ICEV

Optimistic Scenario

Compact BEV200

400 350 300 250 200 150 100 50 0

Compact BEV100

WTW GHG (g CO2-eq mi -1 )

Department of Mechanical Engineering, University of Colorado at Boulder, Boulder, CO

2030

ABSTRACT

ACS Paragon Plus Environment

1


Page 2 of 33

11

This study investigates emission impacts of introducing inexpensive and efficient electric

12

vehicles into the US light duty vehicle (LDV) sector. Scenarios are explored using the

13

ANSWER-MARKAL model with a modified version of the Environmental Protection Agency’s

14

(EPA) 9-region database. Modified cost and performance projections for LDV technologies are

15

adapted from the National Research Council (2013) optimistic case. Under our optimistic

16

scenario (OPT) we find 15% and 47% adoption of battery electric vehicles (BEVs) in 2030 and

17

2050, respectively. In contrast, gasoline vehicles (ICEVs) remain dominant through 2050 in the

18

EPA reference case (BAU). Compared to BAU, OPT gives 16% and 36% reductions in LDV

19

greenhouse gas (GHG) emissions for 2030 and 2050, respectively, corresponding to 5% and 9%

20

reductions in economy-wide emissions. Total nitrogen oxides, volatile organic compounds, and

21

SO2 emissions are similar in the two scenarios due to inter-sectoral shifts. Moderate, economy-

22

wide GHG fees have little effect on GHG emissions from the LDV sector, but are more effective

23

in the electricity sector. In the OPT scenario, estimated well-to-wheels GHG emissions from full-

24

size BEVs with 100-mile range are 62 gCO2-e mi-1 in 2050, while those from full-size ICEVs are

25

121 gCO2-e mi-1.

26

INTRODUCTION

27

A variety of alternative fuels, vehicle technologies, and policy options have been considered to

28

reduce dependence on oil and emissions from the transportation sector.1-4 In particular,

29

greenhouse gas (GHG) emissions from light duty vehicles (LDVs) can be reduced by improving

30

efficiency of vehicles with conventional internal combustion engines (ICEVs) and hybrid electric

31

vehicles (HEVs) via load reductions and powertrain improvements; however, deeper reductions

32

may require electrification. 5


2

Page 3 of 33


33

This study evaluates potential emissions implications of future use of electric vehicles (EVs) in

34

the LDV sector, following optimistic assumptions about improvements in vehicle cost and

35

efficiency. We use an integrated energy system model to evaluate how the US energy system

36

might respond to increased EV penetration. Although EVs, both plug-in hybrid (PHEV) and all-

37

electric (BEV), have low or no tailpipe emissions, they may lead to indirect emissions,

38

depending on the technology used to generate electricity for battery charging.5-9 Moreover,

39

increased penetration of EVs could shift emissions in sectors beyond electricity generation.10 For

40

example, increased EV penetration could draw natural gas from the industrial sector to the

41

electricity sector and push the industrial sector to use more carbon-intensive fuels.

42

Prior studies have examined potential reductions in fuel use and GHG emissions from the

43

LDV sector, including through introducing EVs.5-8,11-15 Some of these studies examined how to

44

reduce emissions to a set target level;5,8,14,15 others examined reductions achievable under

45

specified policies.6,7,11-13 The prior studies examined specific pathways such as demand

46

reductions or use of alternative fuels and technologies; some investigate electrification in

47

particular.5-7,10-14 However, recent advances in EV technology have outpaced vehicle cost and

48

performance assumptions used in earlier assessments, so consideration of more optimistic

49

projections is needed. Furthermore, most prior studies of implications of EV introduction have

50

focused on the transportation and electricity sectors alone without considering implications for

51

energy use in other parts of the economy. 5-8,11-15 Although some studies have accounted for “life

52

cycle” emissions upstream of the electric sector,5,11-15 they still lack the inter-sectoral

53

connections of an integrated model and a feasibility check for their exogenous assumptions such

54

as the assumed penetration rates of EVs. Integrated assessments are needed to develop

55

coordinated policies covering LDVs within the whole energy sector.5


3


Page 4 of 33

56

In one of the few previous cross-sectoral studies, Yeh et al. (2008) applied the integrated

57

energy modeling system ANSWER-MARKAL with the 2008 US EPA 9-Region (US9R)

58

database to evaluate CO2 emission reductions and fuel use in the LDV sector considering PHEVs

59

and ethanol flex-fuel vehicles, but not BEVs, as alternative technologies.9 Their study suggests

60

that a tight economy-wide cap is required to be able to sharply reduce CO2 emissions from the

61

transportation sector.9 Loughlin et al.10 included deploying PHEVs and BEVs among the

62

pathways they investigated for reducing NOx emissions, using the 2014 EPA US9R database.

63

They find the EV pathway to be complicated due to offsetting emissions from inter-sectoral

64

shifts.10 Rudokas et al.16 used MARKAL with the 2012 US9R database and examined the

65

influence of system-wide CO2 emissions fees or a 70% CO2 emissions cap on the transportation

66

sector, looking out to the year 2050. They find that EV penetration increases with a

67

transportation sector CO2 emissions cap, but see little influence of economy-wide emissions fees

68

on that sector.16 Babaee et al.17 developed a US dataset for the MARKAL-EFOM (TIMES)

69

model and investigated combined HEV and EV penetration and associated emissions under

70

numerous combinations of assumptions about natural gas and oil prices, CO2 emissions cap,

71

renewable portfolio standards and battery costs. They find low battery costs to be an important

72

driver of EV and HEV adoption.17 However, their national-level model ignores regional trades

73

and fuel supply curves, transmission constraints and energy conversion and processing

74

technologies other than power plants. Furthermore, while Babaee et al.17 investigated the effect

75

of battery cost on EV penetration, their study ignores the effect of other technology

76

improvements on efficiency and cost, not only for EVs but also for other LDV technologies.

77

Unlike these prior studies, here we develop scenarios representing consistent optimistic


4

Page 5 of 33


78

technology advances across ICEVs and EVs to investigate the effect of these advances on

79

emissions not only from the LDV sector, but also from the whole US energy system.

80

In this study, we use ANSWER-MARKAL in connection with the EPA US9R database

81

(version v1.1; 2014) to examine the impacts of greater EV penetration in the US LDV fleet from

82

now to 2055. Compared to prior studies, the model includes a relatively comprehensive suite of

83

available and viable forthcoming technologies, focusing on improved ICEVs and EVs, for

84

meeting demand for energy services in all economic sectors, including the LDV portion of the

85

transportation sector. Rather than pre-specifying mixes of energy sources and shares of

86

technologies in any of the sectors of the economy, we use an optimization model to determine

87

the least costly choices for meeting demand, with key cost and performance assumptions detailed

88

below. We modeled an optimistic scenario by adapting cost and efficiency estimates for ICEV

89

and EV technologies from the optimistic case developed by the National Research Council

90

Committee on Transitions to Alternative Vehicles and Fuels (NRC, 2013).5 As shown below, this

91

case projects greater improvements in vehicle efficiency and cost than assumed in previous

92

analyses. We compare results from this optimistic scenario with those from EPA’s original 2014

93

US9R database. We also examine how the energy system responds to GHG fees with and

94

without the optimistic LDV assumptions, and examine the sensitivity of the results with respect

95

to the future level of LDV travel demand, oil prices, and fees. In addition, we estimate well-to-

96

wheel (WTW) GHG, SO2, and NOx emissions for BEV and gasoline ICEV technologies based

97

on our OPT scenario results.

98

METHODS

99

ANSWER-MARKAL


5


Page 6 of 33

100

MARKAL uses linear programming to estimate energy supply shifts over a multi-decadal

101

timeframe, finding the least-cost means to supply specified demands for energy services subject

102

to user-defined constraints, assuming a fully competitive market.18 The model computes energy

103

balances at all levels of an energy system from primary sources to energy services, supplies

104

energy services at minimum total system cost, and balances commodities in each time period.18

105

Outputs consist of the penetration of various energy supply technologies at both regional and

106

national levels, technology-specific fuel use by type, and conventional air pollutant and GHG

107

emissions.

108

EPA US9R Database

109

The 2014 EPA US9R database provides inputs to the MARKAL model for nine US regions

110

(shown in Figure S1) and an international import/export region. The database specifies technical

111

and cost features of current and future technologies at five-year intervals, with a structure that

112

connects energy carriers (e.g., output of mining or importing technologies) to conversion or

113

process technologies (e.g., power plants and refineries) and in turn to the transportation,

114

residential, industrial, and commercial end-use sectors. Demand for energy services of end-use

115

sectors is given from 2005 through 2055. Primary energy supplies are specified via piece-wise

116

linear supply curves. The database includes comprehensive treatment of air pollutant emissions

117

from energy production, conversion, and use, accounting for existing control requirements and

118

including a range of additional control options. Current regulations including Renewable

119

Portfolio Standards and biofuel mandates are included as constraints. The database includes joint

120

Corporate Average Fuel Economy (CAFE) and GHG emission standards for LDVs, requiring

121

average fuel economy for passenger cars and light trucks of 34.1 mpg by 2016, rising to 54.5

122

mpg by 2025.19,20 The 2015 Clean Power Plan (CPP) requirements are not included, since they


6

Page 7 of 33


123

were not finalized at the time of the 2014 release.21 All costs are presented in 2005 US dollars,

124

with deflator factors from the Department of Commerce Bureau of Economic Analysis used to

125

adjust prices to the base year. More details are provided in the US9R database documentation.22

126

Future LDV transportation demand in the US9R database is specified based on Annual Energy

127

Outlook (AEO) 2014 projections,23 allocated to the model’s nine regions and to seven vehicle

128

size classes ranging from mini-compacts to light trucks. Demand rises from 2687 billion vehicle

129

miles traveled (bVMT) in 2005 to 3784 bVMT in 2055. This future demand can be met with

130

gasoline (conventional), diesel, ethanol, CNG, and liquefied petroleum gas (LPG) or flex-fueled

131

ICEVs; HEVs; PHEVs; fuel cell vehicles; and BEVs. PHEVs have 20 or 40-mile ranges; BEVs

132

have 100 and 200-mile ranges. All technology and fuel combinations are available in all size

133

classes, except that mini-compact cars have limited options. That is, mini-compact cars are only

134

available for ICEVs, and BEVs. Technology-specific hurdle rates are used to reflect customer,

135

market, and infrastructure barriers. Hurdle rates range from 18% for gasoline and diesel ICEVs

136

to 28% for BEVs. Cost and efficiency of vehicles other than BEVs are taken from AEO2014

137

projections and data from EPA’s Office of Transportation and Air Quality. Cost and efficiency

138

estimates for BEVs are based on expert judgment.24 Emissions factors are calculated from the

139

EPA MOVES model.22 For EV charging, the year is partitioned into summer, winter, and

140

intermediate seasons and the day into am, pm, peak, and nighttime hours. The fraction of EV

141

charging in each time period is fixed as an input in the model and is uniform across regions.

142

Most charging happens at night in all seasons.

143

Changes to the EPA US9R Database and Scenarios Modeled

144

Table 1. Scenarios and sensitivity analyses included in this study. Scenario Name

Description


7


BAU

OPT

BAUFEE/ OPTFEE BAUHIFEE/

Page 8 of 33

Reference scenario with unmodified 2014 EPA US9R database, including EPA’s efficiency and cost estimates for future gasoline ICEV, HEV, PHEV, BEV, and ethanol vehicles. Substitutes optimistic efficiency and cost improvement for gasoline ICEV, HEV, PHEV, BEV, and ethanol vehicles from NRC (2013); adds and refines upstream emissions and refines cost and performance estimates for other sectors. Moderate CO2, and CH4 fees are applied to BAU and OPT scenarios, starting in 2020, based on social cost of carbon and methane.31

OPTHIFEE

CO2 fees are 52% higher in 2020 and 41% higher in 2050 compared to moderate fees; CH4 fees are 36% higher in 2020 and 21% higher in 2050.31

BAULOFEE/ OPTLOFEE

CO2 fees are 69% lower in 2020 and 63% lower in 2050 compared to moderate fees; CH4 fees are 50% lower in 2020 and 48% lower in 2050.31

BAUHIDMD/ OPTHIDMD

LDV demand is increased by 0% in 2005 to 6% in 2040 relative to BAU and OPT scenarios, based on AEO2014 high LDV demand projections. See Figure S2 for complete high demand projections.

BAULODMD/ OPTLODMD

LDV demand is reduced by 0% in 2005 to 19% in 2040 relative to BAU and OPT scenarios, based on AEO2014 low LDV demand projections. (See Figure S2).

BAUHIOIL/ OPTHIOIL BAULOOIL/ OPTLOOIL

Oil prices are increased by 0% in 2005 to 78% in 2040 relative to BAU and OPT scenarios, based on AEO2015 high North Sea Brent crude oil price projections. For complete high oil price projections refer to Figure S3. Oil prices are reduced by 0% in 2005 to 47% in 2040 relative to BAU and OPT scenarios, based on AEO2015 low North Sea Brent crude oil price projections. (See Figure S3).

145 146

Table 1 lists the scenarios examined in this study. The reference case (BAU) uses the

147

unmodified 2014 EPA US9R database, including EPA’s efficiency and cost estimates for all

148

LDV. In contrast, the optimistic scenario (OPT) uses cost and efficiency estimates for gasoline

149

and ethanol ICEV, HEV, PHEVs, and BEVs based on the optimistic projections given by NRC

150

(2013).5 Costs and efficiencies for other fuels and technologies, including FCEV and CNG


8

Page 9 of 33


151

vehicles, were not modified for the OPT case as they were little used in the BAU case (as shown

152

below) and were not the focus of this study.

153

The NRC Committee developed their projections considering both technology-specific

154

powertrain technology improvements and common improvements via reductions in weight and

155

rolling resistance and aerodynamic drag, and improved energy efficiencies for accessories across

156

all technologies. Learning curves were considered for technology improvements and costs,

157

which are calculated based on mass production assumptions. According to the NRC Committee,

158

meeting the optimistic projections would entail significant research and development, but no

159

fundamental technology innovations.5 The Committee also found that meeting these projections

160

would require significant incentives or regulatory requirements to spur development and increase

161

production levels. Compared to other available projections, the NRC estimates provide a more

162

consistent accounting of improvements across different technologies,2,15,23,25 are more readily

163

extended across vehicle sizes,2,11,15,25 and extend further into the future.11,15 For details of the

164

assumptions and calculations see the NRC report and its appendix F.5 Additional changes to the

165

LDV segment in the OPT scenario are summarized in Table 2, and discussed further in the SI

166

Section 4. The SI also explains how the NRC projections are extrapolated to the other vehicle

167

size classes included in the US9R database.

168 169

Table 2. Major changes to LDV sector inputs or constraints for the OPT scenario. See SI Sections 2 and 4 for more information. Parameter/ Description of modifications/ changes Constraint LDV cost

Estimated based on optimistic case of NRC (2013) for gasoline ICEV, HEV, PHEV, BEV, and ethanol vehicle technologies for seven classes of vehicles for the entire time horizon. For example, in 2050 the BEV100 vehicle costs range from 9% to 33% lower than in the BAU scenario.

LDV

Same as cost. For example, in 2050 the BEV100 vehicle fuel economies range


9


Page 10 of 33

efficiency

from 30% to 90% higher than in the BAU scenario.

LDV demand

Updated based on 2014 US Census population projections, which results in lower projections for LDV demand. LDV demand is reduced by 2% in 2050 relative to BAU.

Hurdle rate

Assumed to be uniform at 24% for all technologies to reflect a fully competitive market without customer and infrastructure barriers, compared to 18% for ICEVs, 24% for HEVs and 28% for EVs in the BAU scenario.

CAFE constraint

Recalculated based on LHV (Lower Heating Value), which tightens the CAFE representation in the model by 7-9% in each year.

Investment constraints

Removed for years after 2025, except for BEV with 100-mile range. Constraints updated for BEV100, reflecting limited market for shorter-range vehicles.

170 171

In addition to the changes made for LDVs, the OPT scenario also incorporates refinements for

172

other sectors as described in the SI Section 5, Brown et al.26-28 and McLeod et al.29,30 Overall, the

173

changes provide for a more comprehensive treatment of upstream emissions, an expanded set of

174

emissions control options, especially in the industrial sector, and updates and refinements to cost

175

and equipment lifetime estimates in the electricity sector.

176

The fees applied in the GHG fee scenarios are adapted from EPA (2012),31 which provides

177

estimates for the social cost of methane from an integrated assessment model, rather than more

178

simplistic scaling based on global warming potentials (GWPs).31 The EPA (2012) projections of

179

social cost of CO2 (SCC) are almost equal to those reported by the Interagency Working Group.32

180

The moderate CO2 (CH4) fees are 40 2005 US $/metric tonne of CO2 (1036 $/ tonne of CH4) in

181

2020, escalating to 80 (3107) $/tonne in 2055. High fees are a factor of 1.5 (1.3) times greater,

182

and low fees are 2.9 (1.9) times lower. The full set of values is given in Table S2. CO2-eq is

183

calculated considering CO2 and methane as the main GHG contributors, using both a 20-year

184

GWP of 84 and a 100-yr GWP of 28 for CH4,33 as we are interested in opportunities for

185

reductions over a range of time horizons.


10

Page 11 of 33


186

The BAU and OPT scenarios were also run with high and low LDV demand and high and low

187

oil prices (Table 1) to test sensitivity to these assumptions. We also ran six diagnostic cases,

188

listed in Table S3, to identify which changes made between the BAU and OPT scenarios most

189

influence the results. One case includes all modifications implemented in the OPT scenario

190

except those specific to the LDV sector while the others isolate separate modifications to the

191

LDV sector.

192

Well-To-Wheel Emissions Calculation

193

We use the results from MARKAL to estimate WTW emissions rates for GHG, NOx, and SO2,

194

in grams per mile, for ICEV and BEV technologies. WTW estimates focus on the fuel cycle

195

supporting vehicle operation and neglect emissions from vehicle manufacturing and recycling.

196

We calculate WTW emissions using vehicle efficiencies, upstream emission factors, and the

197

electricity mix corresponding to a certain scenario’s results in each year. We use a 100-year

198

GWP for methane for calculating WTW GHG emissions. Regional calculations for BEV

199

emissions are estimated using the electricity mix for each particular region in each scenario. For

200

ICEV, regional differences are much less significant, so only national average emissions rates

201

are shown.

202

Comparison of LDV Efficiencies and Costs

203

Table 3 illustrates how the optimistic efficiencies from the NRC study compare to projections

204

from the original EPA US9R database and other studies, for gasoline ICEV, HEV and BEV.2,11,25

205

For simplicity, comparisons are shown only for the year 2030 and average or full-size vehicles.

206

Comparisons for other years and fuel-technology combinations are included in Figure S5. The

207

NRC optimistic case projects higher fuel economies across all vehicle types and sizes than the

208

comparison studies. In particular, the recent Argonne National Laboratory (ANL) study of


11


Page 12 of 33

209

prospects for LDVs11 projects fuel economies for BEVs and HEVs that are higher than those in

210

the EPA US9R BAU case but lower than those in the NRC optimistic case.

211

Table 4 compares costs for full-size vehicles in 2030 and 2050 between the NRC optimistic

212

case and the original EPA US9R database. The NRC and EPA costs are similar for gasoline

213

ICEV and HEV technologies in 2030 and 2050. However, for BEV100, the NRC costs are lower

214

than those projected by EPA. For 2030, the NRC’s incremental costs for HEV and BEV100 are

215

also relatively low compared to those for similar vehicles in the recent ANL study.11 Additional

216

comparisons are provided in Figure S4. Table 3. Comparison of projected efficiencies (mpge) for average* or full-sizea LDVs for

217 218

model year 2030. NRC mid.

EPA

ANL

ANL

DOT

DOT

ANL

US9R

avg.25

high25

min.2*

max.2*

201611

NRC opt.5 b,5

ICEV

61

70

53

37

44

31

40

49

HEV

76

88

71

66

83

38

61

76

BEVc

180

207

96

144

163

-

-

172

219

a

220

approach for the other studies to be able to compare the results. ANL 2016 values are for a midsize car.

221

b

222

average (avg.) and high; and from DOT (2010) as minimum and maximum values.

223

c

224

ANL (2016).

225

*: This study focuses on the average vehicle size.

226 227

Midsize and large cars are aggregated to the full-size category in the EPA database. We have taken the same

Estimates from NRC (2013) are characterized as mid-range (mid.) and optimistic (opt.); from ANL (2009) as

BEV range is 130 miles in NRC (2013); 100 miles in the EPA database, 150 miles in ANL (2009), and 90 miles in

Table 4. Comparison of projected cost (thousands 2010 $) for full-size vehicles for model years 2030 and 2050. 2030

2050


12

Page 13 of 33


Technology

EPA US9R

NRC opt.5

EPA US9R

NRC opt.5

ICEV

27.72

28.57

27.74

29.86

HEV

30.85

29.18

30.66

30.45

BEV100

34.95

28.40

32.73

28.01

228 229

Transition Policies

230

For contrast with the BAU scenario, our OPT scenario is constructed based on the assumption

231

that customer and market barriers for penetration of EVs have been lowered to make EVs

232

competitive with gasoline vehicles, in particular after 2025. To get to this point, subsidies or

233

regulations that will lower initial market barriers and encourage increased production volumes

234

are needed. Except for CAFE, these policies are not explicitly represented in our scenarios, but

235

rather they are approximated by the relaxed investment constraints and equalized hurdle rates in

236

the OPT scenario.

237

RESULTS AND DISCUSSION

238

LDV Penetration

239

Figure 1a shows national results for LDV penetration in terms of vehicle stocks for the BAU

240

and OPT scenarios. While gasoline vehicles dominate in the BAU scenario for the entire time

241

horizon, in the OPT scenario BEVs gain a LDV market share of about 15% (all from BEV100)

242

by 2030 and 47% (with 20% share from BEV100 and 27% from BEV200) by 2050. In 2050,

243

these BEVs are mainly in the compact (6% from BEV100 and 21% from BEV200), full (4%

244

from BEV100 and 6% from BEV200), and small SUV (10% from BEV100), size classes, which

245

have relatively high share and lower upfront cost compared to other size classes. HEVs play a

246

negligible role in the BAU scenario, but are adopted to a moderate extent in the midterm in the

247

OPT scenario. Ethanol vehicles account for about 3% of VMT in both scenarios in 2030, but are


13


Page 14 of 33

248

eliminated by 2050 in the OPT scenario. Diesel vehicles account for a steady 4% in both

249

scenarios.

250

As noted above, approaching the OPT scenario for EV penetration would require transition

251

policies to lower market barriers. These are not explicitly modeled in this study. That is, our

252

study assumes market barriers for deploying EVs have been conquered, such as charging stations

253

have become widely available. However, the NRC assessment5 illustrates the level of subsidy

254

that might be required to lower market barriers for EVs. The NRC Committee used the Light-

255

duty Alternative Vehicle Energy Transitions (LAVE-Trans) model of consumer demand in the

256

transportation sector to explore several transition policy scenarios, including one that assumed

257

optimistic EV technology advances together with EV subsidies. The current $7500 federal tax

258

credit for EVs was continued through 2020, briefly increased to $15000 in 2021 and then phased

259

out by ~2030. This scenario achieved market shares of 8% PHEV and 33% BEV in 2050, with

260

no subsidies provided at that point in time.5 Our OPT case is somewhat more aggressive,

261

achieving about a 15% higher EV market share in 2050.

262

Regional results (not shown) generally follow the same pattern as the national scale results,

263

except that the ethanol vehicle share varies across the regions. In 2030, it ranges from 0% to

264

21% for the BAU scenario and from 0% to 16% for the OPT scenario. In 2050, the ethanol share

265

ranges from 0% to 37% for the BAU scenario, but ethanol is not used in 2050 in the OPT

266

scenario. The highest ethanol penetration is in the East North Central region, where corn is

267

relatively abundant and ethanol fuel costs are relatively low. For the OPT scenario, the BEV

268

share is about 15% in 2030 in all regions, but ranges from 40% to 52% across regions in 2050.

269

Six additional diagnostic cases (see Table S3) were run to understand which changes between

270

BAU and OPT were most influential. The results (see Figure S7) show that CAFE is the main


14

Page 15 of 33


271

driving force for HEV penetration in the mid-term (case D1). In case D1, HEVs account for 14%

272

of total VMT in 2030, but are not used in 2050. The reason for this is that by 2025-2030, ICEVs

273

are not sufficiently efficient to satisfy CAFE requirements. HEVs are more efficient than ICEVs,

274

and are cheaper than other alternative technologies that could be used to satisfy the standards.

275

However, the database assumes ICEV efficiency will continue to improve and by 2050 they are

276

sufficiently efficient to meet CAFE requirements. The model selects ICEVs over HEVs at this

277

point because ICEVs are less expensive. In isolation, incorporating optimistic LDV cost and

278

efficiency assumptions (case D2) or lowering hurdle rates and investment constraints (D3) do not

279

significantly alter BEV penetration compared to the BAU scenario. However, significant BEV

280

penetration occurs when these changes are combined (cases D4 and D5). Combining these

281

optimistic assumptions reflects a more realistic and consistent scenario than separating them,

282

since if cost and technological barriers of EVs have been addressed, it is expected that hurdle

283

rates would also be lowered. Lastly, modifications outside the LDV sector have negligible

284

impact on the LDV mix (D6).

285

We also consider results from a variety of scenarios designed to explore the sensitivity of our

286

results to assumptions regarding VMT demand, GHG fees, and oil prices (see Table 1). In the

287

low (high) LDV demand cases, input LDV demand was reduced (increased) by 12% (5%) in

288

2030, and by 20% (4%) in 2050 in the BAU scenario, and by 19% (6%) in 2050 in the OPT

289

scenario. With lower demand, the VMT of gasoline cars was reduced by 14% in the BAU

290

scenario and 24% in the OPT scenario in 2030; and by 20% in both scenarios in 2050. In the

291

OPTLODMD case, HEV VMT is 36% higher in 2030 than in the OPT scenario. Increasing LDV

292

demand results in approximately proportional increases in the VMT of gasoline vehicles in both

293

the BAU and OPT scenarios. BEV VMT in the OPTHIDMD scenario also increases in


15


Page 16 of 33

294

proportion to overall VMT demand. Sensitivity to oil prices was tested using AEO2015

295

projections.34 These changes had negligible impact on the technology mix in the LDV sector in

296

either the BAU or OPT scenario. This is because, as shown in SI Section 8, upfront vehicle costs

297

have a much greater effect than fuel costs on the annualized unit cost of vehicle ownership.

298

To examine how optimistic assumptions about LDV technology advances would alter the

299

response of the energy system to GHG fees, moderate system-wide fees were applied starting in

300

the year 2020 (BAUFEE and OPTFEE). The LDV segment of the transportation sector is able to

301

respond by shifting to more efficient vehicle technologies and/or to fuels with lower in-use and

302

upstream GHG emissions. As shown in Figure 1a, with BAU assumptions for LDV efficiency

303

and costs, applying moderate economy-wide GHG fees has little effect on the LDV technology

304

and fuel mix. With OPT assumptions, application of moderate fees increases the share of BEVs

305

by 4.4% in 2050. Some ethanol also enters the vehicle/fuel mix in 2050 in the OPTFEE scenario.

306

Note that there are uncertainties associated with upstream emissions for ethanol that could offset

307

the uptake of CO2 that is assumed in the model.35-36 Changing the fees as indicated in Table 1

308

produced negligible change in the LDV technology mix in the BAU scenario (results not shown).

309

In the OPT scenario, increasing the fees does not change the share of BEVs in 2050; decreasing

310

the fees reduces the BEV share by 3%. Overall, the influence of GHG emission fees on the LDV

311

technology and fuel mix is limited due to their modest impact on the cost of vehicle ownership.

312

For all vehicle technologies, the impact of emissions fees in future decades is lower than it would

313

be in the near term due to improvements in vehicle efficiency.

314

Electricity Mix

315

Use of an integrated energy system model allows for examination of how changes in the LDV

316

sector might affect energy choices in other sectors, including electricity generation. As shown in


16

Page 17 of 33


317

Figure 1b, while the total electricity generation is equal in the BAU and OPT scenarios in 2030,

318

it is 4% higher in 2050 in the OPT scenario, mainly due to the increased use of BEVs. Electricity

319

generation from natural gas increases over time in both scenarios, whereas generation from

320

existing coal plants declines. Our diagnostic case D6 shows that the extra natural gas generation

321

in the OPT scenario compared to BAU is mainly due to the change we made to the EPA US9R

322

database in limiting the lifetime of existing coal plants to 75 years. In the original database,

323

existing coal plants could be used to the end of the modeled time horizon. Electricity generation

324

from other technologies including wind and solar are similar between the BAU and OPT

325

scenarios, with these two renewable technologies contributing about 9% of generation in 2050.

326

Altering the LDV demand changes the total electricity demand by less than 1% in both low and

327

high demand cases, with either BAU or OPT scenarios.

328

When moderate GHG fees are applied, total electricity demand in both scenarios decreases by

329

3% in 2030 and 5% in 2050. Both BAUFEE and OPTFEE utilize more natural gas (49% and

330

41% in 2030, and 23% and 11% in 2050) and wind (44% and 42% in 2030, and 50% and 61% in

331

2050) and less coal (46% and 48% in 2030, and 50% and 56% in 2050) than the respective

332

scenarios without fees. Electricity consumption is 5% higher in the OPTFEE scenario in 2050

333

compared to the BAUFEE scenario, due to greater BEV penetration. For both BAU and OPT,

334

total electricity generation is 1-3% higher in the low fee cases and 1-2% lower in the high fee

335

cases, compared to the corresponding moderate fee scenarios. Carbon capture and sequestration

336

is applied to 4% of generation in 2050 in the BAUHIFEE case and 3% of generation in the

337

OPTHIFEE case.

338

Crude Oil Consumption


17


Page 18 of 33

339

Changes in crude oil consumption in the BAU and OPT scenarios and corresponding fee and

340

sensitivity cases are shown in Figure S8. Total crude oil consumption in the BAU scenario

341

decreases by 4% from 2010 to 2030, driven by reductions in gasoline, which are somewhat offset

342

by increases in diesel. The reductions in gasoline consumption correspond to average efficiency

343

improvements of 65% for gasoline ICEV, which offset the 18% increase in LDV VMT demand.

344

The increased diesel consumption is mainly for heavy-duty vehicles (HDV). In 2030, gasoline

345

consumption in OPT is 3% lower than in the BAU scenario, largely because 30% of VMT is met

346

with HEVs and BEVs. From 2030 to 2050, gasoline consumption remains steady in the BAU

347

scenario despite an increase in LDV VMT demand. By 2050, gasoline consumption in the OPT

348

scenario is 62% lower than in the BAU scenario, because BEV provide almost half of LDV

349

VMT. The decrease in gasoline consumption in OPT provides more capacity for other refined

350

products, such as jet fuel and petrochemical feedstock, so the total reduction in crude oil

351

consumption from BAU to OPT is not as large as the reduction in gasoline use. Consumption of

352

imported refined products (including gasoline) is almost five times lower than that of

353

domestically refined products in both scenarios throughout the time horizon.

354

In the sensitivity tests, increasing the LDV demand results in less than a 1% increase in crude

355

oil consumption in 2030 and 2050 in both the BAUHIDMD and OPTHIDMD cases. However,

356

reducing LDV demand decreases total crude oil consumption by 3% in 2030 in both cases, and

357

by 6% (BAULODMD) and 3% (OPTLODMD) in 2050, mainly from reduced gasoline

358

consumption. With moderate GHG fees, the largest change is seen in 2050, where the total crude

359

oil consumption in the OPTFEE case is about 3.5% lower than in the OPT scenario.

360

GHG Emissions


18

Page 19 of 33


361

Figure 2a shows GHG emissions for the BAU and OPT scenarios as CO2-eq using the 100-

362

year GWP for methane. Results calculated using a 20-year GWP are shown in Figure S9 and

363

demonstrate similar patterns. Total GHG emissions decrease from 2010 to 2030 and then

364

increase from 2030 to 2050 in both scenarios. In the BAU scenario, GHG emissions from the

365

LDV sector decline by 53% from 2010 to 2030 and by an additional 21% from 2030 to 2050,

366

due largely to existing CAFE regulations. Correspondingly, in the BAU scenario the LDV share

367

of total emissions is reduced from 21% in 2010 to 10% in 2030 and 7% in 2050. Direct

368

emissions from the LDV sector contribute 9% and 5% of GHG emissions in the OPT scenario in

369

2030 and 2050, respectively. That is, our OPT scenario results in GHG emissions from LDVs

370

that are 36% lower than in the BAU scenario in 2050.

371

Total GHG emissions in the OPT scenario are 5% lower than those in the BAU scenario in

372

2030 and 9% lower in 2050. This is a smaller percentage reduction than that from the LDV

373

sector, because LDV emissions represent a declining share of total emissions and because of

374

inter-sectoral shifts in emissions. The differences between the OPT and BAU scenarios come

375

from greater reductions from the LDV segment (16% lower in OPT than BAU in 2030 and 36%

376

lower in 2050); other transportation (16% lower in OPT in 2030 and 25% lower in 2050); and

377

from the electric power sector (5% lower in OPT in 2030, and 7% lower in 2050). The Pacific

378

region shows the largest reductions in GHG emissions in the OPT scenario compared to BAU

379

(23% in 2030 and 22% in 2050), and the West South Central region shows the least reductions

380

(2% in 2030, and 4% in 2050).

381

In diagnostic case D5, with all the OPT improvements in the LDV sector but without

382

modifications in the electric sector, emissions from the electric sector are 2% higher in 2050 than

383

in the BAU scenario, due to greater LDV demand for electricity. The reduction in GHG


19


Page 20 of 33

384

emissions from the non-LDV segment of the transportation sector in OPT versus BAU is mainly

385

due to a shift to more efficient technologies in freight shipping. Consistent with the results for

386

the LDV mix, LDV emissions are reduced or increased approximately in proportion to LDV

387

demand, for both BAU and OPT scenarios.

388

significantly in sensitivity tests with modified oil prices applied in either the BAU or OPT

389

scenarios.

Similarly, CO2-eq emissions do not change

390

For both the BAU and OPT scenarios, application of moderate GHG fees reduces total CO2-eq

391

emissions by 11% in 2030 and by 12% in 2050. In both comparisons, the main reductions are

392

from the electric power sector, based on increased use of natural gas and some renewables in

393

place of coal. GHG emissions from the electric sector are reduced by about 40% in the BAUFEE

394

and OPTFEE scenarios in 2030 compared to 2005. For BAU and OPT, total GHG emissions are

395

9-10% higher with low GHG fees and 7-9% lower with high fees compared to the moderate fee

396

results (Figure S10), again mainly due to changes in electric sector emissions. For the LDV

397

segment, GHG emissions in 2050 are only 9% lower in OPTFEE than in OPT. Thus, despite

398

using more optimistic assumptions for EV cost and efficiency than used by Rudokas et al.16, our

399

results agree with their conclusion that GHG fees would have little effect on LDV emissions.

400

Moreover, over time LDV emissions represent a sharply declining share of total GHG emissions,

401

even without more optimistic technology improvements or fees. On the other hand, the OPT

402

scenario shows that applying energy system-wide GHG fees could help curtail the industrial

403

sector emissions increases that might otherwise occur if widespread use of BEVs induced

404

industrial sector fuel switching.

405

NOx and SO2 Emissions


20

Page 21 of 33


406

Figures 2b and c show how NOx and SO2 emissions compare for the BAU, OPT and

407

corresponding fee scenarios. (Results for VOC emissions are presented in Figure S11.) In the

408

BAU scenario, NOx emissions from LDV decline from about 2 million tonnes in 2010 to 0.4

409

million tonnes in 2030 and to 0.3 million tonnes in 2050, due to tailpipe emissions limits that

410

have already been promulgated. In the OPT scenario, direct NOx emissions from LDV are about

411

0.3 million tonnes in 2030 and 0.2 million tonnes in 2050, which is 35% lower than in the BAU

412

scenario. However, total energy system NOx emissions are about the same in 2050 in the OPT

413

and BAU scenarios, because industrial sector emissions are 27% higher in the former. In the

414

OPT scenario, the industrial sector shifts from use of electricity to use of combustion-based

415

technologies for heat and power. However, the increase in industrial NOx emissions in the OPT

416

scenario is partially offset by a decrease in HDV emissions that results from use of more efficient

417

and less polluting technologies in freight shipping, as more diesel is used in the industrial sector.

418

Loughlin et al.10 similarly saw a shift in NOx emissions from the transportation sector to other

419

sectors in their study of NOx emissions reduction pathways, including vehicle electrification.

420

Application of GHG emissions fees has little impact on LDV NOx emissions.

421

The transportation sector makes a negligible contribution to SO2 emissions (Figure 2c).

422

However, SO2 is of interest with EVs due to the possibility of increased emissions from the

423

electric sector. In the BAU scenario, SO2 emissions from electricity generation fall from 5

424

million tonnes in 2010 to 1.4 million tonnes in 2050, due to existing control requirements and the

425

shift away from coal-fired generation. This reduction is modestly countered in the BAU scenario

426

by an increase in SO2 emissions from the industrial sector. In the OPT scenario, SO2 emissions

427

from the industrial sector in 2050 are 20% higher than in the BAU scenario in 2050, due to

428

increased direct fuel use in place of electricity. This change offsets the reductions from the


21


Page 22 of 33

429

electric sector that result from less use of coal-fired power generation in the OPT scenario. GHG

430

fees reduce total SO2 emissions in the BAUFEE and OPTFEE scenarios, respectively, by 17%

431

and 22% in 2030, and by 21% and 28% in 2050. Most of the reductions occur in the electric

432

sector in the BAU scenario; in the OPT scenario emissions from the industrial sector are reduced

433

as well.

434

Well-To-Wheel Emissions

435

We highlight the impact of potential improvements in LDV efficiency by using our OPT

436

scenario results to calculate WTW GHG emissions from gasoline ICEV and BEV technologies.

437

WTW results for NOx, and SO2 were also calculated and are presented in the Supporting

438

Information (Figures S12 and S13). On average for 2010, we estimate WTW GHG emissions of

439

186 and 450 gCO2-eq mi-1 for full-size BEV100 and gasoline-fueled ICEVs in the US,

440

respectively. For BEV100, 89% of the emissions are from electricity generation and 11% from

441

upstream fuel sectors. For ICEVs, 88% of emissions are from the use phase, 7% from the

442

refinery, and 5% from fuel production. WTW GHG emissions for full-size BEV100 drop to 94

443

gCO2-eq mi-1 in 2030 and 62 gCO2-eq mi-1 in 2050. For full-size ICEVs, emissions are 181

444

gCO2-eq mi-1 in 2030 and 121 gCO2-eq mi-1 in 2050. Thus by 2050, the WTW GHG emissions

445

estimated for both technologies are about one-third of those estimated for 2010. The percentage

446

contributions across WTW stages are relatively constant over time, for both BEVs and ICEVs.

447

ANL (2016)11 estimates higher WTW GHG emissions for midsize ICEV, and BEV90 in 2015

448

and 2030, than our estimations for full-size (combination of midsize and large) cars in those

449

years. Our range for BEV is also slightly higher.

450

In addition to the WTW GHG emissions estimated based on the average national electricity

451

mix, we also calculated average WTW emissions for each MARKAL region (See Figure S14).


22

Page 23 of 33


452

Regional WTW GHG emissions for full-size BEV100 range from 96 to 219 gCO2-eq mi-1 in

453

2010, 42 to 121 gCO2-eq mi-1 in 2030, and 35 to 83 gCO2-eq mi-1 in 2050. Thus the highest

454

regional WTW emissions in 2050 are less than the lowest regional WTW emissions in 2010.

455

Across the full time horizon, the lowest BEV WTW GHG emissions correspond to the Pacific

456

region, which has a low share of electricity from coal power plants and high hydropower

457

resources. We find the highest emissions for the West North Central region, which has the

458

highest share of generation from coal (59% in 2030 and 54% in 2050 in the OPT scenario). Our

459

regional rankings for EV WTW emissions in 2015 match those presented by the Union of

460

Concerned Scientists (2015);12 however, their study did not consider future scenarios.

461

Although not directly examined in this study, vehicle cycle emissions are also important in

462

comparing across technologies. ANL (2016) estimated emissions from vehicle manufacturing of

463

about 41 gCO2-eq mi-1 for current midsize ICEV, and about 64 gCO2-eq mi-1 for BEV90

464

vehicles.11 In the future, these emissions are expected to decline as electricity sector emissions

465

are reduced and manufacturing processes become more efficient.

466

vehicles generally become more efficient, with lower WTW emissions, the vehicle production

467

cycle will likely comprise an increasing fraction of total life cycle emissions. Thus, in future

468

assessments greater attention needs to be focused on vehicle manufacturing and recycling

469

impacts, as well as on upstream emissions and on the potential for inter-sectoral shifts.

However, as light duty

470 471

ASSOCIATED CONTENT

472

Supporting Information


23


Page 24 of 33

473

Additional details describing modifications made to the EPA US9R database, scenario

474

development and additional results. This material is available free of charge via the Internet at

475

http://pubs.acs.org.

476

AUTHOR INFORMATION

477

Corresponding Author

478

*Tel.: (303) 492-5542. Fax: (303) 492-3498. Email: [email protected].

479

ACKNOWLEDGMENT

480

This work was supported by a Sustainable Energy Pathways grant from the National Science

481

Foundation, award number CHE-1231048. We thank Kristen E. Brown, Jeffrey McLeod, Dan

482

Loughlin and Carol Shay for contributions to and assistance with the MARKAL/US9R model.

483

Any opinions, findings, and conclusions or recommendations expressed in this material are those

484

of the authors and do not necessarily reflect the views of the National Science Foundation.

485

REFERENCES

486

(1) Heywood, J.; Baptista, P.; Berry, I.; Bhatt, K.; Cheah, L.; Sisternes, F.; Karplus, V.; Keith,

487

D.; Khusid, M.; MacKenzie, D.; McAulay, J. An action plan for cars, the policies needed to

488

reduce US petroleum consumption and greenhouse gas emissions; An MIT Energy Initiative

489

Report; Massachusetts Institute of Technology, 2009; ISBN 978-0-615-34325-9.

490

(2) Transportation’s role in reducing US greenhouse gas emissions, volume 1: synthesis

491

report;

Report

to

Congress;

US

Department

of

Transportation,

492

http://ntl.bts.gov/lib/32000/32700/32779/DOT_Climate_Change_Report_-_April_2010_-

493

_Volume_1_and_2.pdf.

2010;


24

Page 25 of 33


494

(3) Greene, D. L.; Baker, Jr. H. H.; Plotkin, S.E. Solutions: reducing greenhouse gas emissions

495

from US transportation; Prepared for the PEW Center on Global Climate Change; PEW Center

496

on

497

http://www.c2es.org/publications/reducing-ghg-emissions-from-transportation.

Global

Climate

Change

and

Argonne

National

Laboratory,

2011;

498

(4) Policy options for reducing energy use and greenhouse gas emissions from US

499

transportation; Special Report 307; Committee for a Study of Potential Energy Savings and

500

Greenhouse Gas Reductions from Transportation; Transportation Research Board of the National

501

Academies, 2011; ISBN 978-0-309-16742-0.

502

(5) Transition to alternative vehicles and fuels; National Research Council report; Committee

503

on Transitions to Alternative Vehicles and Fuels; Board on Energy and Environmental Systems;

504

Division on Engineering and Physical Sciences, 2013; ISBN 978-0-309-26852-3.

505

(6) Choi, D. G.; Kreikebaum, F.; Thomas, V. M.; Divan. D. Coordinated EV adaption: double-

506

digit reductions in emissions and fuel use for $40/vehicle-year. Environ. Sci. Technol. 2013, 47

507

(18), 1070310707; DOI 10.1021/es4016926.

508

(7) Peterson, S. B.; Whitacre, J.F.; Apt, J. Net air emissions from electric vehicles: the effect of

509

carbon price and charging strategies. Environ. Sci. Technol. 2011, 45 (5), 17921797; DOI

510

10.1021/es102464y.

511

(8) Kromer, M. A.; Banivadekar, A.; Evans, C. Long- term greenhouse gas emission and

512

petroleum reduction goals: evolutionary pathways for the light-duty vehicle sector. Energy

513

(Oxford, U. K.) 2009, 35(1), 387397; DOI 10.1016/j.energy.2009.10.006.


25


514

(9) Yeh, S.; Farrell, A.; Plevin, R.; Sanstad. A; Weyant, J. Optimizing US

Page 26 of 33

mitigation

515

strategies for the light-duty transportation sector: what we learn from a bottom-up model.

516

Environ. Sci. Technol. 2008, 42 (22), 82028210; DOI 10.1021/es8005805.

517

(10) Loughlin, D. H.; Kaufman, K.R.; Lenox, C. S.; Hubbell, B. J. Analysis of alternative

518

pathways for reducing nitrogen oxide emissions. J. Air Waste Manage. Assoc. 2015, 65(9),

519

10831093; DOI 10.1080/10962247.2015.1062440.

520

(11) Elgowainy, A.; Han, J.; Ward, J.; Joseck, F.; Gohlke, D.; Lindauer, A.; Ramsden, T.;

521

Biddy, M.; Alexander, M.; Barnhart, S.; Sutherland, I.; Verduzco, L.; Wallington, T. Cradle-to-

522

grave lifecycle analysis of U.S. light-duty vehicle-fuel pathways: a greenhouse gas emissions and

523

economic assessment of current (2015) and future (2025-2030) technologies; Argonne National

524

Laboratory, 2016; ANL/ESD-16/7.

525

(12) Nealer, R.; Reichmuth, D.; Anair, D. Cleaner cars from cradle to grave; how electric cars

526

beat gasoline cars on lifetime global warming emissions; Union of Concerned Scientists, 2015;

527

http://www.ucsusa.org/sites/default/files/attach/2015/11/Cleaner-Cars-from-Cradle-to-Grave-

528

full-report.pdf.

529

(13) Bandivadekar, A.; Bodek, K.; Cheah, L.; Evans, C.; Groode, T.; Heywood, J.; Kasseris,

530

E.; Kromer, M.; Weiss, M. On the road in 2035: reducing transportation’s petroleum

531

consumption and GHG emissions; Report No. LFEE 2008-05 RP; Massachusetts Institute of

532

Technology; Laboratory for Energy and Environment, 2008; ISBN 978-0-615-23649-0.

533

(14) Meier, P.J.; Cornin, K. R.; Frost, E. A.; Runge, T. M.; Dale, B. E.; Reinemann, D. J.;

534

Detlor, J. Potential for electrified vehicles to contribute to US petroleum and climate goals and


26

Page 27 of 33


535

implications for advanced biofuels. Environ. Sci. Technol. 2015, 49 (14), 82778286; DOI

536

10.1021/acs.est.5b01691.

537

(15) Elgowainy, A.; Rousseau, A.; Wang, M.; Ruth, M.; Andress, M. D.; Ward, J.; Joseck, F.;

538

Nguyen, T.; Das, S. Cost of ownership and well-to-wheels carbon emissions/oil use of

539

alternative fuels and advanced light-duty vehicle technologies. Energy Sustainable Dev. 2013, 17

540

(6), 626641; DOI 10.1016/j.esd.2013.09.001.

541

(16) Rudokas, J.; Miller, P.J.; Trail, M. A.; Russell, A. G. Regional air quality aspects of

542

climate change: impact of climate mitigation options on regional air emissions. Environ. Sci.

543

Technol. 2015, 49 (8), 51705177; DOI 10.1021/es505159z.

544 545 546

(17) Babaee, S.; Nagpure, A. S.; DeCarolis, J. F. How much do electric drive vehicles matter to future US emissions? Environ. Sci. Technol. 2014, 48 (3), 13821390; DOI 10.1021/es4045677. (18) Loulou, R.; Goldstein, G.; Noble, K. Documentation for MARKAL family of models;

547

Energy

548

etsap.org/web/mrkldoc-i_stdmarkal.pdf.

549

Technology

Systems

Analysis

Programme

(ETSAP),

2004;

http://www.

(19) Light-duty vehicle greenhouse gas emission standards and Corporate Average Fuel

550

Economy

551

http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-+Fuel+Economy/Model+Years+2012-

552

2016:+Final+Rule.

standards;

Final

Rule;

EPA

and

DOT;

NHTSA,

2010;

553

(20) Light-duty automotive technology, carbon dioxide emissions, and fuel economy trends:

554

1975 through 2011; Executive Summary; EPA; Transportation and Climate Division; OTAQ,

555

2012; EPA-420-S-12-001 a;


27


Page 28 of 33

556

http://nepis.epa.gov/Exe/ZyPDF.cgi/P100DYX6.PDF?Dockey=P100DYX6.PDF.

557

(21) Clean Power Plan Final Rule; Federal Registrar, 40 CFR part 60; Environmental

558

Protection Agency (EPA); RIN 2060–AR33, Vol. 80, No.205, 2015; EPA–HQ–OAR–2013–

559

0602; FRL–9930–65– OAR;

560

http://www.epa.gov/cleanpowerplan/clean-power-plan-existing-power-plants.

561

(22) EPA US nine-region MARKAL database, database documentation; US Environmental

562

Protection Agency (EPA), 2013; EPA 600/B-13/203;

563

http://nepis.epa.gov/Adobe/PDF/P100I4RX.pdf.

564

(23) Annual Energy Outlook 2014 with projections to 2040; US Energy Information

565

Administration (EIA), 2014; DOE/EIA-0383(2014);

566

http://www.eia.gov/forecasts/aeo/pdf/0383(2014).pdf.

567

(24) Rick Baker, ERG (Eastern Research Group Inc., Austin Texas), personal communication

568 569 570

with EPA staff, 2011. (25) Plotkin, S.; Singh, M. Multi-path transportation futures study: vehicle characterization and scenario analyses; Argonne National Lab.; Energy System division, 2009; ANL/ESD/09-5.

571

(26) Brown. K.E. Internalizing air quality and greenhouse gas externalities in the US energy

572

system and the effect on future air quality. Ph.D. Comprehensive Exam Report, Department of

573

Mechanical Engineering, University of Colorado Boulder, CO, 2014.


28

Page 29 of 33

574


(27)

Brown,

K.E.;

Henze,

D.

K.;

Milford.

J.

B.

Internalizing

life

cycle

575

externalities in the US energy system. Proceedings from the LCA XIV International Conference,

576

San Francisco, CA, United States 2014, 149160; http://www.lcacenter.org/lca-xiv.aspx.

577 578

(28) Brown, K. E.; Henze, D. K.; Milford. J.B. The effect of criteria pollutant and greenhouse gas damage based fees on emissions from the US energy system. CMAX Conference, 2014;

579

www.cmascenter.org/conference/2014/agenda.cfm.

580

(29) McLeod, J. D. Characterizing the emissions implications of future natural gas production

581

and use in the US and Rocky Mountain region: a scenario-based energy system modeling

582

approach. Master Dissertation, Department of Mechanical Engineering, University of Colorado

583

Boulder, CO, 2014.

584

(30) McLeod, J. D.; Brinkman, G. L.; Milford, J. B. Emissions implications of future natural

585

gas production and use in the US and in the Rocky Mountain region. Environ. Sci. Technol.

586

2014, 48 (22), 1303613044; DOI 10.1021/es5029537.

587 588

(31) Marten, A. L.; Newbold, S. C. Estimating the social cost of non-CO2 GHG emissions: methane and nitrous oxide. Energy Policy 2012, 51, 957972; DOI 10.1016/j.enpol.2012.09.073.

589

(32) Technical support document: technical update of the social cost of carbon for regulatory

590

impact analysis- under Executive Order 12866; US Government, Interagency Working Group on

591

Social Cost of Carbon, 2013;

592 593

https://www.whitehouse.gov/sites/default/files/omb/inforeg/social_cost_of_carbon_for_ria_20 13_update.pdf.


29


594 595

Page 30 of 33

(33) Climate Change 2013, the physical science basis; Intergovernmental panel on climate change (IPCC); WGI, 2013;

596

www.ipcc.ch/report/ar5/wg1/.

597

(34) Annual Energy Outlook 2015 with projections to 2040; US Energy Information

598

Administration (EIA), 2015; DOE/EIA-0383(2015);

599

http://www.eia.gov/forecasts/aeo/pdf/0383(2015).pdf.

600

(35) Crutzen, P. J.; Mosier, A. R.; Smith, K. A.; Winiwarter, W. N2O release from agro-biofuel

601

production negates global warming reduction by replacing fossil fuels. Atmos. Chem. Phys.

602

2008, 8, 389395;

603

http://www.atmos-chem-phys.net/8/389/2008/acp-8-389-2008.pdf.

604

(36) Hill, J.; Nelson, E.; Tilman, D.; Polasky, S.; Tiffany, D. Environmental, economic, and

605

energetic costs and benefits of biodiesel and ethanol biofuels. Proc. Natl. Acad. Sci. U. S. A.

606

2006, 103 (30), 1120611210; DOI 10.1073/pnas.0604600103.


30

Page 31 of 33


Figure 1. US technology mix for (a) LDV and (b) electricity generation by scenario and year. PHEV, ICEV and HEV are gasoline fueled. 114x165mm (300 x 300 DPI)



Figure 2. US emissions in 2050 compared to BAU2010 emissions (a) GHG (b) NOx (c) SO2 by sector and scenario; HDV sector includes off-road vehicles. 152x294mm (300 x 300 DPI)


Page 32 of 33

Page 33 of 33


43x25mm (300 x 300 DPI)


Emission Impacts of Electric Vehicles in the US Transportation Sector

Recommend Documents