Technology Limits for Reducing EU Transport Sector CO2 Emissions

Policy Analysis pubs.acs.org/est

Technology Limits for Reducing EU Transport Sector CO2 Emissions Lynnette M. Dray,*,† Andreas Schaf̈ er,†,‡ and Moshe E. Ben-Akiva§ †

University of Cambridge, Institute for Aviation and the Environment, 1-5 Scroope Terrace, Cambridge CB2 1PX, U.K. Stanford University, Precourt Energy Efficiency Center, Yang & Yamazaki Environment & Energy Building, 473 Via Ortega, Stanford, California 94305-4205, United States § Massachusetts Institute of Technology, Department of Civil and Environmental Engineering, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States ‡

S Supporting Information *

ABSTRACT: Using a new data set describing the techno-economic characteristics of current and projected future transport technologies and a synthesis of existing transport demand models, lifecycle CO2 emissions from 27 EU countries (EU27) were estimated in the absence and presence of new policy interventions to 2050. Future CO2 emissions are strongly dependent on geographical scope and economic growth assumptions, and to a lesser extent on uncertainties in technology characteristics, but in the absence of new policy intervention they continue to rise from present-day values in all three scenarios examined. Consequently, EU27 emissions goals, which may require a 60% decrease in transport sector greenhouse gas emissions from year-1990 values by 2050, will be difficult to meet. This is even the case under widespread adoption of the most promising technologies for all modes, due primarily to limitations in biofuel production capacity and a lack of technologies that would drastically reduce CO2 emissions from heavy trucks and intercontinental aviation.

■

INTRODUCTION EU27 greenhouse gas (GHG) emissions from all sectors but transportation have decreased from year-1990 values. Between 1990 and 2007, transportation GHG emissions increased by around 25%, with emissions from international aviation more than doubling;1 in 2007, transport accounted for around 23− 30% of EU-27 GHG emissions, depending on whether intercontinental travel is included.2 Over the same 18-year period, modest reductions in average emissions per passenger kilometer (pkm) traveled and tonne kilometer (tkm) generated have been countered by strongly rising demand. For example, average tailpipe CO2 emissions per km from new passenger cars have decreased by 15% since 1995,3 but intra-European passenger car pkm traveled over the same time period rose 21%.2,4 EU climate policy aims to limit the global mean temperature increase from anthropogenic climate change to below 2 °C. EC analysis9 indicates that achieving this goal would require cuts of 80−95% in EU GHG emissions with respect to year-1990 levels by 2050. As deeper cuts are likely to be made in other sectors, this requires a cut of at least 60% in transportation GHG emissions, most notably CO2, by midcentury.9 The EU has also made a commitment to reduce emissions in sectors outside the EU ETS, including transportation, by 10% on year-2005 levels by 2020. Past research concludes that achieving these goals through technological measures alone is difficult. For example, iTREN20305 projects intra-EU27 transport CO2 emissions increasing modestly to 2030 without new policy, but decreasing 31% © 2012 American Chemical Society

below year-1990 levels with ambitious policy action. The EU Transport GHG project6 estimates transport CO2 emissions to rise by 74% above year-1990 levels by 2050 with no new policies, or decline by 36% below year-1990 levels with full adoption of low-carbon technologies and fuels. More conservative mitigation potentials are estimated for the United States. The Moving Cooler7 project concluded year-2050 surface transport emissions would likely remain above year1990 levels even with aggressive policy action. The U.S. Department of Transportation’s 2010 Report to Congress8 estimates emissions in 2030 of around 27% above year-1990 levels, which could be reduced by 26−30% to 2050 with aggressive deployment of battery-electric road technology. As demand and hence emissions depend strongly on future assumptions about socioeconomic variables, these values will likely be altered in particularly high- or low-growth futures. Many transportation technologies currently in development show promise in significantly reducing energy use and CO2 emissions before 2050. A detailed survey of these technologies, their likely future techno-economic characteristics, and the uncertainty involved was carried out by the Technology Opportunities and Strategies toward Climate-friendly Transport (TOSCA) project.10 The TOSCA data covering all major transportation modes and its explicit treatment of uncertainty Received: Revised: Accepted: Published: 4734

December 1, 2011 March 9, 2012 March 21, 2012 March 21, 2012 dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741

Environmental Science & Technology

Policy Analysis

typically only average values are quoted. To better quantify uncertainty, the TOSCA project carried out expert surveys on technology and fuel characteristics, and results were used to estimate most likely values and uncertainty ranges in each case (summarized in Tables 1 and 2; in some cases uncertainty ranges have been estimated from related variables). References 15−20 give detailed further information on the technology assumptions and surveys. Scenarios and Demand. Transport sector CO2 emissions strongly depend on socioeconomic variables (e.g., GDP), oil price, the carbon intensity of transportation fuels, and policy measures. Because the future development of socioeconomic variables, the oil price, and the carbon intensity of electricity are outside the transportation sector and uncertain, we specify plausible ranges for their future development via a set of three scenarios.21 Table 3 gives key assumptions about these variables and references for comparable scenario developments. The “Baseline” scenario continues past trends. The “Challenging” and “Favorable” scenarios respectively represent futures in which transportation demand and emissions will grow particularly rapidly or slowly. Demand is also affected by changes in the average journey time and cost from adopting new technologies or policies. Because demand and fleet composition are mutually interdependent, modeling is carried out in two steps. In the first, we project demand by scenario in the absence of new technologies or policies. In the second, we calculate how that projected demand may change in response to changes in fleet composition or policy. The first step was carried out using Transtools,26 a networkbased model which projects passenger and freight demand by mode and trip purpose on a NUTS3 geographic level. Transtools is not suitable to carry out technology modeling, as it does not account for vehicle fleets in detail. However, it is useful to generate detailed initial demand projections. In this step, it is assumed that existing policies remain unchanged (e.g., fuel taxes) and planned near-future interventions will be put in place (e.g., aviation entering the EU ETS in 2012). No new policies are applied and fuel use trends assume only incremental improvements of those technologies currently in use. Intercontinental aviation demand was obtained using the global aviation systems model AIM,27 with the same set of scenarios and input assumptions. A summary of outcomes is given in the Supporting Information. Vehicle Fleet and Emissions. To estimate technology uptake and emissions from these demand projections, a fivestep framework was followed: 1. Calculate existing fleet size and vehicle retirements by country and year, based on base year fleet data, vehicle age, and policies. 2. Calculate how many new vehicles need to enter the fleet to satisfy scenario demand projections. 3. Using scenario and TOSCA data on fuel prices and vehicle characteristics, estimate the type and technology class of these vehicles. 4. Estimate whether the new fleet equilibrium would change passenger or freight demand (e.g., by reducing journey cost). Steps 2−4 are iterated if necessary. 5. Estimate the resulting emissions. The existing fleet and vehicle withdrawal were estimated using fleet data by vehicle location and age, and vehicle agedependent retirement curves, from Eurostat, TREMOVE, and

suggests a new study is justified which looks at what range of reductions in emissions is possible over different scenarios and uncertain input values. This is the subject of this paper.

■

METHODOLOGY To estimate the emissions mitigation potential of alternative technologies, we combine TOSCA’s techno-economic characteristics of transportation technology with existing European transport modeling capabilities. Starting with projections of transportation demand to 2050, the number of new vehicles required is estimated. Adoption of new technologies and fuels is estimated based on their availability and cost effectiveness under projected scenario variables such as fuel price. As new technologies may impact journey time, cost, and hence demand, this step is iterated with an estimation of demand changes. Emissions are then estimated based on fleet composition. Figure 1 gives an overview. Although this process

Figure 1. Outline of the model structure used in this paper. Transtools, SUMMA, TREMOVE, and AIM are existing models.

is simpler than that used in some comparable projects,5,11 it allows for a rapid run time so that multiple scenarios and input variable uncertainty can be assessed. Each step is described below. Technology Characteristics. Technology and fuel characteristics needed include R&D requirements, market entry year, investment and operating costs, fuel use, and lifecycle CO2 emissions. These values can be uncertain. Although there are existing estimates of technology characteristics, particularly for passenger cars (e.g., Schäfer et al., 2009;4 GHG-transpord 2012;11 Krail and Schade 2011;12 McKinsey 2009,13 201014), 4735

dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741

Environmental Science & Technology

Policy Analysis

Table 1. Projected Values for Some Characteristics of Major Technologies and Fuels from TOSCA: Summary passenger cars R&D requirementsa reference midsize gasoline ICEe−bioethanol blend (E85) plug-in hybrid electric vehicle battery electric vehicle fuel cell hybrid electric vehicle

reference diesel hybrid electric vehicle fuel cell hybrid electric vehicle

reference diesel resistance reduction idling reduction

reference narrowbody narrowbody replacement fast open rotor reduced-speed open rotor

reference electric combination technologyh

reference electric combination technologyh

technology readiness year

retail pricec (€ (2009))

direct (lifecycle) CO2 emissionsc (gCO2/vkm)

learning rateb

emission reduction rated (%/year)

none

2009

16,500

0.1

145.9 (170.9)

0.8−1.0

insignificant

2015−2020

16,700−19,230

0.1

25.6−33.4 (58.6−76.7)

0.8−1.0

substantial

2012−2017

19,800−23,100

0.2

42.8−59.3 (99.3−115.0)

0.7

substantial substantial

2010−2020 2010−2015

19,800−26,400 19,800−33,000

0.4 0.4

0 (63.8−115.0) 0 (71.4−122.4)

0.5 0.5

technology readiness year 2009 2015−2020 2015

light trucks retail pricec (€ learning (2009)) rateb 28,880 0.1 37,300 0.1 35,765 0.4

R&D requirementsa none significant substantial

R&D requirementsa none insignificant significant R&D requirementsa none

technology readiness year 2009 2010 2010−2015 technology readiness year 2009

heavy trucks retail pricec (€ learning (2009)) rateb 66,000 0.1 71,980 0.1

direct (lifecycle) CO2 emissionsc (gCO2/tkm) 256.2 (304.9) 234.2 (278.7) 0 (156.9)

direct (lifecycle) CO2 emissionsc (gCO2/tkm) 39.3 (46.7) 36.7 (43.9)

75,990 0.1 aircraft (stage length = 983 km)f retail pricec (million cost reduction € (2009)) rateg, %/year 55 −0.41−0.93

37.5 (44.5) direct (lifecycle) CO2 emissionsc (gCO2/pkm) 76.0 (92.0)

emission reduction rated (%/year) 0.9−1.1 0.9−1.1 0.5

emission reduction rated (%/year) 0.3−0.5 0.3−0.5 0.3−0.5 emission reduction rated (%/year) 0.13−0.29

significant

2020−2030

60−73

−0.41−0.93

59.2−62.9 (71.6−76.1)

0.13−0.29

substantial substantial

2020−2030 2020−2030

66−93 66−93

−0.41−0.93 −0.41−0.93

48.8−51.8 (59.0−62.7) 41.9−44.5 (50.7−53.8)

0.13−0.29 0.13−0.29

R&D requirementsa existing substantial

technology readiness year 2009 2020−2030

passenger trains retail pricec (million cost reduction € (2009)) rateg, %/year 13.1−17.1 0.5 13.7−19.3 0.5

direct (lifecycle) CO2 emissionsc (gCO2/pkm) 0 (44.0) 0 (20.0−28.0)

R&D requirementsa existing substantial

technology readiness year 2009 2020−2030

freight trains retail pricec (million cost reduction € (2009)) rateg, %/year 3.4−4.4 0.5 4.2−5.6 0.5

direct (lifecycle) CO2 emissionsc (gCO2/tkm) 0 (18.0) 0 (10.0−13.0)

emission reduction rated (%/year) 0.30 0.92

emission reduction rated (%/year) 0.18 0.18

“Insignificant” indicates a technology that is essentially already developed, “Significant” indicates a company-level R&D effort would be required to bring the technology to market with these characteristics, and “Substantial” indicates that an EU-wide R&D effort would be needed. bTechnology learning rates are defined as the percent reduction in vehicle production costs for a doubling of the cumulative production. A learning rate of 0.1 represents a 10% reduction in production costs for a doubling of the cumulative number of vehicles produced. cFor the year of market entry. Prices exclude VAT. dDefined as the percentage reduction in emissions per year through technology improvements for new vehicles after the year of market entry. eInternal combustion engine. fAlthough a long-haul blended wing body aircraft was considered as part of the TOSCA project, expert opinion was divided as to whether it could make an impact on emissions before 2050. Therefore it is excluded here. gFor aircraft and trains, a percentage rate for reduction of production costs per year is used instead of learning rates. hThis is a combination of low drag, low mass, energy recovery, eco-driving, energy efficiency, and space efficiency technologies. For the values in this table, average EU27 year-2009 values for the carbon intensity of electricity generation are assumed; for model runs, scenario values are ulilized. a

AIM.3,28,29 The available pkm and tkm by country from the existing fleet were estimated using TOSCA technology characteristics (including vehicle utilization). The number of new vehicles required was then estimated based on the shortfall between these amounts and projected scenario pkm and tkm. Purchaser technology choice for these new vehicles was assessed on a cost-effectiveness basis using net present value

(NPV) as a decision criterion, with parameters chosen to take account of factors such as consumer myopia with regard to fuel cost savings. Parameters and references are given in the Supporting Information. For large commercial vehicles such as aircraft, NPV is often used for real-life purchase decisions. However, it represents a significant approximation for passenger cars, where many other attributes affect purchase 4736

dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741

Environmental Science & Technology

Policy Analysis

Table 2. Projected Values for Some Characteristics of Major Fuels from TOSCA: Summary fuel (feedstock)

R&D requirements

technology readiness year

production + distribution costsa (€ (2009)/ GJfuel)

direct (lifecycle) CO2 emissions (gCO2/ MJfuel)

gasoline (fossil) diesel (fossil) jet A1 (fossil) bioethanol (wood) BTL (wood) CNG (fossil) HVO hydrogen (fossil) hydrogen (wood)

existing existing existing significant substantial existing existing existing substantial

2009 2009 2009 2015−2020 2025 2009 2009 2009 2030

13.2 + 0.4 13.2 + 0.4 13.2 + 0.4 18.0−35.0 + 0.8−7.5 30.6 + 0.4−3.6 7.5−19.0 + 0.2−12.6 19.0−30.0 + 0.9−4.7 14.0−20.3 + 3.0−40.0 14.5−20.8 + 3.0−40.0

73.0 (85.5) 74.8 (89.0) 74.3 (88.5) 0 (17.5−43.7) 0 (6.9−39.0) 56.2 (64.9−78.2) 0 (24.9−60.3) 0 (64.0−112.5) 0 (7.5−14.8)

a

Excludes excise duty and VAT. The change in these values over time depends on economies of scale and on feedstock price; see Perimenis et al. (2011)20 for more details.

Table 3. Key Relevant and Uncertain Exogenous Scenario Values Used in TOSCA: Summary scenario assumptions scenario 0: “baseline” comparable projections scenario 1: “challenging” comparable projections scenario 2: “favorable” comparable projections

change in oil price, %/yr (€ (2009)/bbl)

change in EU27 GDP, %/yr

change in CO2 intensity of electricity generation, %/yr

+1.8 (54 − 113)

−1.7

continues past trends; Eurostat (2010); iTren-2030 (2009)5 +2.5

continues past trends; Eurostat (2010)3 ±0 (54 − 54)

continues long-term trend 1970−2010; IEA (2010)22 −0.5

IEA (2008)23

iTren-2030 (2009)5

continues recent trend 1995−2010; IEA (2010a)22

+0.7

+2.5 (54 − 144)

−3.0

close to Eurostat projected average for 2007−2012 period3

IEA (2010b)24 “current policies” scenario

assumes approximate 80% reduction by 2050 on year1990 levels; PRIMES (2010)25

+1.7 3

decisions in addition to cost-effectiveness.30 Given the wide range of combinations of vehicle technologies in TOSCA, the data available within the scope of the project did not allow estimating a more complex consumer choice model. Therefore, although the broad outcomes are consistent with those of other models,4,31,32 these results are intended only to generate a firstorder understanding of fleet behavior. The costs associated with owning and operating each vehicle were taken from the TOSCA techno-economic studies,10 including cost trends over time and with increased technology production. New technologies are likely to change journey cost and/or time, resulting in demand changes. Due to high runtime (>5 days per model year), rerunning Transtools here is not feasible. However, as only a limited number of variables change from the base case, it is possible to use a simpler model. To account for the change in demand from Transtools values we use elasticities and cross-elasticities of demand estimated from the outputs of the transport demand meta-model SUMMA33 (a development of EXPEDITE34). SUMMA is designed to straightforwardly calculate aggregate changes in demand from a base projection, given policy-induced changes in journey time and cost. Therefore, although it is not suitable for carrying out baseline demand projections, it can be used here to estimate changes in demand. Fuel use and emissions are calculated using the TOSCA vehicle characteristics set (Table 1) based on estimated pkm and tkm by vehicle type. These include trends over time in the fuel use of new vehicles due to incremental improvements in vehicle technology, and also to account for the higher emissions of older vehicles remaining in the fleet. To account for uncertainties in technology parameters, we run Monte Carlo simulations. Following the process used in Allaire (2010)35 we assume the maximum entropy probability distributions

corresponding to the type of information provided by the survey outputs; as most values are estimated in terms of most likely value and lower and upper bounds, this corresponds to a beta distribution (here approximated by triangular distributions). Simulations are run until output parameters (e.g., interquartile range of emissions totals) remain stable to two significant figures beyond the accuracy reported in this paper typically 3000−6000 runs per scenario/policy case combination. Policies. This model framework offers several locations for policy levers. Policies can affect whether a technology with a “substantial” (EU-wide) R&D requirement will be produced (Table 1) or an older technology will be retired. Vehicle subsidies, penalties, or feebate systems affect the purchase price of a technology, and carbon taxes that affect the price of travel of all vehicles in the fleet. Sample policies are discussed further in the results section below.

■

RESULTS AND DISCUSSION

No New Policies Case. As a reference development, we run all three scenarios with no new policies applied. Major nearfuture confirmed policies, such as the inclusion of aviation in the EU ETS, are included, but it is not assumed that emissions targets will necessarily be met. Year-2009 levels of excise duty, VAT, and subsidies are assumed unchanged for all fuels through 2050. Technologies and fuels rated as needing “substantial” R&D investments (Table 1) are assumed unfunded, i.e. will not enter the market. This also assumes that technology development is independent of scenario variables such as oil price. CO2 emissions are shown in Figure 2. For totals which include intercontinental transport, we also 4737

dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741

Environmental Science & Technology

Policy Analysis

Figure 2. Emissions by mode and scenario for the No New Policies case. The horizontal black lines indicate a 60% reduction from year-1990 levels. Shaded areas (typically small for this case) indicate the range of uncertainty in emissions totals due to uncertainty in technology characteristics. Historical lifecycle emissions assume year-2009 ratios of direct to lifecycle emissions by fuel type.

contrast, CO2 emissions decrease through 2050 by about 10% relative to 2010 in the “Favorable” scenario with only intraEU27 traffic. Sample Policies. The following policies were chosen to demonstrate a range of possible fleet outcomes: • R&D onlythis option is assumed to make available all technologies with “substantial” (EU-wide) R&D requirements. • R&D plus carbon taxas above, plus a carbon tax is applied from 2015, and is increased over 10 years to a maximum value of €100/t CO2. • R&D plus electric vehicle subsidyR&D as above, plus a €3000/vehicle purchase subsidy is available for plug-in hybrid and battery electric vehicles. • R&D plus fuel cell electric vehicle subsidyR&D as above, plus a €3000/vehicle purchase subsidy for fuel cell electric vehicles. Midrange parameters are used for all uncertain technology characteristics. Although emissions stringencies are not modeled as a separate policy, the rates of improvement in existing technology (Table 1) assume existing stringency

include estimates of shipping emissions to 2050 based on the simple GDP-sensitive model in Eyring et al.36 Emissions trajectories vary by mode and geographical scope, but in nearly all cases are projected to increase from presentday values by 2050. The biggest increase comes from intercontinental aviation emissions. Although growth rates are high, they are in line with aviation industry demand growth projections of 5%/year37,38 for scenarios similar to the TOSCA baseline. Adoption of new technologies is small in the No New Policies case, because of the R&D limitations discussed above and the high costs associated with some available technologies. Consequently, uncertainty ranges are small, mainly reflecting uncertainties in the development of existing technology. Technologies widely adopted in the fleet in all scenarios include heavy truck driving resistance reduction, the compositematerial intensive evolutionary narrowbody replacement aircraft, and space-efficient trains. Carbon accounting practices also have a strong impact on total emissions. Figure 2 shows totals for intra-EU27 transport only and for intra-EU27 plus half of intercontinental CO2 emissions. In the “Challenging” scenario including intercontinental transport, emissions more than double by 2050. In 4738

dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741

Environmental Science & Technology

Policy Analysis

Figure 3. Lifecycle emissions by sample policy case and scenario. The horizontal black lines indicate a 60% reduction from year-1990 levels. Historical lifecycle emissions assume year-2009 ratios of direct to lifecycle emissions by fuel type.

requirements, most notably for passenger cars, will continue to be tightened. More extreme stringency scenarios forcing the use of electric or fuel cell car technology will behave similarly to the subsidy scenarios studied. Lifecycle CO2 emissions are shown in Figure 3 by scenario. In the R&D only case, year 2050 emissions are reduced by 8−10% (depending on scenario) from “No New Policy” values, primarily due to the use of alternative fuels from wood feedstocks (these runs assume that second generation biofuels will be made available primarily to aircraft and trucks, as supply is limited; these vehicles have fewer other emissions mitigation measures available and are thus willing to pay the highest price). Although alternative passenger car technologies are available, their adoption is very limited. This conclusion is supported by other studies4,5,31 and reflects the high costs of these technologies compared to current technology (Table 1). If R&D investments are complemented by a €100/t CO2 carbon tax, lifecycle CO2 emissions reduction double to 16−19% compared to the “No New Policies” case. This results mainly from increased use of alternative fuels and demand reduction from increased journey costs; passenger car alternative technology uptake remains low. The two subsidy cases result in reductions of year-2050 emissions by slightly lower levels, i.e. 14−17% (plug-in hybrid electric vehicles) and 12−17% (fuel cell electric vehicles) and alternative technology fleet penetration of 40−55%. While the carbon tax reduces transportation demand and induces extra biofuel use, the vehicle subsidies cause much greater technology penetration for passenger cars, and a demand increase for passenger cars due to reduced journey costs. Subsidy support would require up to 0.8% of EU27 GDP, whereas the carbon taxation policy would increase government revenues by a similar amount. Although these sample policies would reduce

absolute emission levels from present-day values for the “Favorable” scenario by up to 10% even if including emissions from intercontinental traffic, none of these sample policy cases approaches either a 10% reduction by 2020 from year-2005 levels, or a 60% reduction by 2050 from year-1990 levels. “Lowest-Emissions” Case. Finally, we examine the maximum emissions reduction achievable through transportation technologies alone. In this case, sufficient subsidy for widespread adoption is applied to the lowest-emission vehicle, fuel, and capacity technology combination in each category. This involves radical changes from current fleets (e.g., fuel cell vehicles powered by hydrogen from biomass). As biofuels are widely used, production capacity is important. Based on a literature review of future biomass production capabilities (e.g., OFID (2009),39 IEA (2009),40 REFUEL (2007)41), we limit biomass availability to projected global production by year. This assumption represents an extreme lower limit on emissions. In addition to R&D and infrastructure support, government spending of at least 2% of EU27 GDP would be required in subsidies and to compensate for lost fuel tax revenue. Some technologies used also have unresolved challenges (e.g., open rotor aircraft engine noise) which may further decrease governments’ willingness to provide support. Figure 4 shows the resulting CO2 emissions. Emissions again differ strongly depending on scenario and geographical scope. As more new technologies are in operation, the outcomes are also more uncertain. An emissions decrease of 60% or greater is achievable by 2050 only for direct intra-EU27 transport CO2 emissions in the “Baseline” and “Favorable” scenarios, and is never achieved in the “Challenging” scenario, which reaches the biofuel production limit. The goal of reducing emissions by 10% on year-2005 levels by 2020 is also missed. These results 4739

dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741

Environmental Science & Technology

Policy Analysis

Figure 4. Comparison between emissions in the “No New Policies” case (NNP) and the “Lowest” case (L) by scenario and geographical scope. Shaded areas indicate uncertainty ranges in the outputs due to uncertainty in technology characteristics; the interquartile range is indicated by the darker shaded areas in each case. The horizontal black lines indicate a 60% reduction from year-1990 levels. Historical lifecycle emissions assume year-2009 ratios of direct to lifecycle emissions by fuel type.

■

build on those of other studies5−7 by confirming that using currently projected technologies and fuels supported by policy interventions to achieve the 60% goal is extremely challenging, even under optimistic scenario conditions, absent any behavioral change that would lead to lower levels of transportation demand. These results also apply when explicitly taking into account the uncertainty underlying the technoeconomic characteristics of future technologies. Future Work. To carry out this study, a number of simplifications were made which could be expanded on in future work. These include vehicle choice assessment accounting for factors other than cost and including risk aversion, a more detailed treatment of maritime transport, and an assessment of the impact of behavioral measures in conjunction with technological measures to reduce emissions. Concluding Remarks. Exploiting the full potential of currently projected technology and fuel opportunities requires policy intervention. Many critical technologies and fuels will need EU-wide R&D investments in order to be produced at commercial scale. High carbon prices, stringent regulation, or other significant policy intervention will likely be needed to induce market penetration of breakthrough passenger car and aircraft technologies. Realizing these opportunities requires society to prioritize climate change mitigation, as such interventions may lead to additional public expenditures, higher prices, and decreased mobility. Even then, technological measures alone cannot produce large enough emissions reductions to meet EU climate goals. The question then is better understanding the potential for behavioral measures to mitigate transport sector GHG emissions.

ASSOCIATED CONTENT

S Supporting Information *

(1) Technologies modeled, (2) demand results, (3) model parameters, (4) emissions tables, (5) comparison with literature scenarios. This information is available free of charge via the Internet at http://pubs.acs.org/.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; phone: +44 1223 760124; fax: +44 1223 332960. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This paper is based on the EU FP7-funded TOSCA project, carried out by Cambridge University, ETHZ, DBFZ, the Paul Scherrer Institute, ECORYS, KTH, and NTUA. Adnan Rahman, Robert Kok, and Konstantina Laparidou provided scenario data and transtools modeling; their support is gratefully acknowledged, as is that of the TOSCA advisory board, especially Meinrad Eberle, John Green, and Remy Prud’homme.

■

REFERENCES

(1) The European EnvironmentState and Outlook 2010: Data and Maps; European Environment Agency, 2010; http://www.eea.europa. eu/data-and-maps/.

4740

dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741

Environmental Science & Technology

Policy Analysis

(2) EU Energy and Transport in FiguresStatistical Pocketbook 2010; European Commission; Publications Office of the European Union: Luxembourg, 2010. (3) Eurostat online database: transport statistics; Eurostat, 2010; http://epp.eurostat.ec.europa.eu/portal/page/portal/transport/data/ database. (4) Schäfer, A.; Heywood, J. B.; Jacoby, H. D.; Waitz, I. A. Transportation in a Climate-Constrained World; MIT Press: Cambridge, MA, 2009. (5) Integrated transport and energy baseline until 2030, Deliverables 1−4, 2009; iTREN-2030 project, 2009; http://www.tmleuven.be/ project/itren2030/home.htm. (6) Skinner, I.; Van Essen, H.; Smokers, R.; Hill, N. Towards the Decarbonisation of the EU’s Transport Sector by 2050. Final Report of the EU Transport GHG: Routes to 2050 Project; 2010; http:// www.eutransportghg2050.eu/cms/eu-transport-ghg-routes-to-2050project-reports/. (7) Moving Cooler: An Analysis of Transportation Strategies for Reducing Greenhouse Gas Emissions. Executive Summary; 2009; http://www.movingcooler.info/Library/Documents/ Moving%20Cooler%20Executive%20Summary.pdf. (8) Transportation’s Role in Reducing U.S. Greenhouse Gas Emissions; U.S. Department of Transportation Report to Congress; 2010; http://ntl.bts.gov/lib/32000/32700/32779/DOT_Climate_ Change_Report_-_April_2010_-_Volume_1_and_2.pdf. (9) Roadmap to a Single European Transport AreaTowards a competitive and resource efficient transport system; European Commission White Paper COM(2011) 144, Brussels, 2011. (10) Schäfer, A. et al. TOSCA Project Final Report; 2011; http:// www.toscaproject.org/FinalReports/TOSCA_FinalReport.pdf. (11) GHG-transpord project website; http://www.ghg-transpord.eu/ ghg-transpord/index.php. (12) Krail, M.; Schade, W. Technological Potential for CO2 Emission Reductions of Passenger Cars. In Transport Moving to Climate Intelligence: New Chances for Controlling Climate Impacts of Transport after the Economic Crisis; Rothengatter, W., Ed.; Springer: New York, 2011; pp 271−290. (13) Pathways to a Low-Carbon Economy; McKinsey, 2009; http:// www.mckinsey.com/Client_Service/Sustainability/Latest_thinking/ Pathways_to_a_low_carbon_economy. (14) A Portfolio of Power-trains for Europe: A fact-based analysis; McKinsey, 2010; http://www.europeanclimate.org/documents/ Power_trains_for_Europe.pdf. (15) Safarianova, S.; Noembrini, F.; Boulouchos, K.; Dietrich, P. Techno-Economic Analysis of Low-GHG Emission Passenger Cars; TOSCA Project WP1 Final Report; 2011. http://www.toscaproject. org/FinalReports/TOSCA_WP1_Car.pdf. (16) Safarianova, S.; Noembrini, F.; Boulouchos, K.; Dietrich, P. Techno-Economic Analysis of Low-GHG Emission Light, Medium and Heavy Duty Vehicles; TOSCA Project WP1 Final Report; 2011. http:// www.toscaproject.org/FinalReports/TOSCA_WP1_Truck.pdf. (17) Andersson, E.; Berg, M.; Nelldal, B.-L.; Fröidh, O. Rail Passenger Transport: Techno-economic analysis of energy and greenhouse gas reductions; TOSCA Project WP3 Final Report; 2011. http://www. toscaproject.org/FinalReports/TOSCA_WP3_RailPassenger.pdf (18) Andersson, E.; Berg, M.; Nelldal, B.-L.; Fröidh, O. Rail Freight Transport: Techno-economic analysis of energy and greenhouse gas reductions; TOSCA Project WP3 Final Report; 2011. http://www. toscaproject.org/FinalReports/TOSCA_WP3_RailFreight.pdf. (19) Vera-Morales, M.; Graham, W.; Hall, C.; Schäfer, A. TechnoEconomic Analysis of Aircraft; TOSCA Project WP2 Final Report; 2011. http://www.toscaproject.org/FinalReports/TOSCA_WP2_Aircraft. pdf. (20) Perimenis, A.; Majer, S.; Zech, K.; Holland, M.; Müller-Langer, F. Lifecycle Assessment of Transportation Fuels; TOSCA Project WP4 Final Report. http://www.toscaproject.org/FinalReports/TOSCA_ WP4_Fuels.pdf. (21) Kok, R.; Laparidou, K.; Rahman, A.; Dray, L. M. Scenarios of European Transport Futures in a Global Context; TOSCA Project WP6

Final Report; 2011; http://www.toscaproject.org/FinalReports/ TOSCA_WP6_Scenarios.pdf. (22) World Energy Statistics and Balances; International Energy Agency; OECD iLibrary; 2010. http://www.oecd-ilibrary.org/energy/ data/iea-world-energy-statistics-and-balances_enestats-data-en. (23) World Energy Outlook 2008; International Energy Agency, 2008. http://www.iea.org/weo/2008.asp. (24) World Energy Outlook 2010; International Energy Agency, 2010. http://www.iea.org/weo/2010.asp. (25) E3M-Lab. PRIMES Model, E3Mlab of ICCS/NTUA, Version used for the 2010 Scenarios for the European Commission including new sub-models; 2010. (26) TRANS-TOOLS version 2; Model and Data Improvements; TENconnect Final Report; DG TREN: Brussels, 2010. (27) Dray, L.; Evans, A.; Reynolds, T.; Schafer, A. Mitigation of Aviation Emissions of Carbon Dioxide. Transp. Res. Rec. 2010, 2177, 17−26. (28) TREMOVE Project Final Report; European Commission DG Environment: Brussels; http://www.tremove.org/documentation/ Final_Report_TREMOVE_9July2007c.pdf. (29) Morrell, P.; Dray, L. M. Environmental Aspects of Fleet Turnover, Retirement and Life Cycle; Final Report for the Omega Consortium, 2009. (30) de Jong, G.; Fox, J.; Pieters, M.; Daly, A. J.; Smith, R. A Comparison of Car Ownership Models. Transp. Rev. 2004, 24 (4), 379−408. (31) Christidis, P.; Hidalgo, I.; Soria, A. Trends in Vehicle and Fuel Technologies: Scenarios for Future Trends; IPTS Technical Report Series, EUR 20748 EN; European Commission: Seville, 2003. (32) Greene, D. L.; Patterson, P. D.; Singh, M.; Li, J. Feebates, Rebates and Gas-guzzler Taxes: A Study of Incentives for Increased Fuel Economy. Energy Policy 2005, 33, 757−775. (33) van Grol, R.; Walker, W.; de Jong, G.; Rahman, A. Using a Meta-Model to Analyse Sustainable Transport Policies for Europe: The SUMMA Project’s Fast Simple Model; Final Report, 2001. http://www.summa-eu.org. (34) de Jong, G.; Gunn, H.; Ben-Akiva, M. A meta-model for passenger and freight transport in Europe. Transport Policy 2004, 11, 329−344. (35) Allaire, D. L. Uncertainty Assessment of Complex Models with Application to Aviation Environmental Systems. PhD Dissertation, Massachusetts Institute of Technology, MA, 2009. (36) Eyring, V.; Köhler, H. W.; Lauer, A.; Lemper, B. Emissions from international shipping: 1. Impact of future technologies on scenarios until 2050. J. Geophys. Res. 2005, 110, D17306. (37) Global Market Forecast 2009−2028; Airbus Industrie:Toulouse, 2009; http://www.airbus.com/en/myairbus/global_market_forecast. html. (38) Current Market Outlook 2010−2029; The Boeing Company: Seattle, WA, 2009; http://www.boeing.com/commercial/cmo/. (39) Biofuels and Food Security; OFID, 2009; http://www.ofid.org/ publications/PDF/biofuels_book.pdf. (40) Transport, Energy and CO2, Moving Towards Sustainability; IEA: Paris, 2009. (41) Assessment of Biomass Potentials for Biofuel Feedstock in Europe: Methodology; REFUEL WP2 Final Report, 2007; http://www.refuel. eu/fleadmin/refuel/user/docs/Refuel-D6-Jul2007-final6.pdf.

4741

dx.doi.org/10.1021/es204301z | Environ. Sci. Technol. 2012, 46, 4734−4741