Incorporating Time-Corrected Life Cycle Greenhouse Gas Emissions

Jan 27, 2012 - However, improvements in vehicle technology, lower-carbon fuels, and improvements in GHG accounting practices which account for ...
0 downloads 3 Views 621KB Size
Download PDF
Policy Analysis pubs.acs.org/est

Incorporating Time-Corrected Life Cycle Greenhouse Gas Emissions in Vehicle Regulations Alissa Kendall*,† and Lindsay Price† †

Department of Civil and Environmental Engineering, University of California, One Shields Avenue, Davis, California 95616, United States S Supporting Information *

ABSTRACT: Beginning with model year 2012, light-duty vehicles sold in the U.S. are subject to new rules that regulate tailpipe greenhouse gas (GHG) emissions based on grams of CO2−equivalent per mile (gCO2e/mi). However, improvements in vehicle technology, lower-carbon fuels, and improvements in GHG accounting practices which account for distortions related to emissions timing all contribute to shifting a greater portion of life cycle emissions away from the vehicle use phase and toward the vehicle production phase. This article proposes methods for calculating time-corrected life cycle emissions intensity on a gCO2e/mi basis and explores whether regulating only tailpipe CO2 could lead to an undesirable regulatory outcome, where technologies and vehicle architectures with higher life cycle GHGs are favored over technologies with lower life cycle emissions but with higher tailpipe GHG emissions. Two life cycle GHG assessments for future vehicles are presented in addition to time correction factors for production and end-of-life GHG emissions. Results demonstrate that, based on the vehicle designs considered here, there is a potential for favoring vehicles with higher life cycle emissions if only tailpipe emissions are regulated; moreover, the application of time correction factors amplifies the importance of production emissions and the potential for a perverse outcome.

1. INTRODUCTION Transportation sector emissions contribute nearly one-third of all U.S. greenhouse gas (GHG) emissions, and combustion of motor gasoline alone comprises more than 60% of these emissions or nearly one-fifth of all GHG emissions in the U.S.1 Reducing fuel consumption in the light duty vehicle sector is thus a critical step in achieving any climate change mitigation target for the nation. Fuel economy, and by consequence GHG emissions, in the light-duty sector has historically been managed by the corporate average fuel economy (CAFE) standard. However, beginning with model year 2012, vehicles sold in the U.S. are subject to additional rules that regulate tailpipe GHG emissions based on grams of CO2−equivalent (CO2e) per mile. Fuel combustion during vehicle operation has been, and continues to be, the dominant source of emissions during the vehicle life cycle. Life cycle assessments (LCAs) of standard passenger vehicles have estimated use-phase or operation emissions at approximately 85−95% of total life cycle GHG emissions.2−4 As a result, current regulatory approaches, which fundamentally address fuel consumption, have likely succeeded in limiting the average GHG emissions from new vehicles and may continue to do so for some time. However, two emerging trends require that we re-examine the effectiveness of this approach to light-duty vehicle regulations: (i) Improvements in powertrain technology, lower-carbon fuels, and vehicle light-weighting achieve reduced CO2 © 2012 American Chemical Society

emissions during vehicle operation, while often increasing the material and manufacturing burdens for vehicles. Both trends shift a greater portion of life cycle emissions away from the vehicle use phase and toward the vehicle production phase.5,6 (ii) Improved linking of the science of climate change and GHG accounting practices (including accounting practices within LCA) demonstrate that, all else equal, GHG emissions that occur at the outset of a product life cycle, have a greater global warming impact over typical analytical time horizons, than those occurring in the future.7−10 By the same logic, avoided emissions from recycling that occur at a vehicle’s end-of-life (EOL) have less benefit when their timing is accounted for. In combination these trends could lead to a perverse regulatory outcome, where technologies and vehicle architectures with higher life cycle GHGs are favored over technologies with lower life cycle emissions but with higher use-phase GHG emissions. A life cycle-based GHG emissions intensity standard for vehicles, particularly one that includes the effect of emissions timing, could avert the risk of such an outcome. This article describes the development of research that tests the usefulness and feasibility of a life cycle-based GHG Received: Revised: Accepted: Published: 2557

September 3, 2011 January 27, 2012 January 27, 2012 January 27, 2012 dx.doi.org/10.1021/es203098j | Environ. Sci. Technol. 2012, 46, 2557−2563

Environmental Science & Technology

Policy Analysis

reduced by the fraction of time in the future that they occur. Credits (or debits) in emissions that occur 20 years in the future, for example, would be scaled down by 20%. While both PAS 2050 and ILCD make important contributions by highlighting the importance of emissions timing, both treat the problem with simplified solutions that do not represent the behavior of emissions in the atmosphere and their contribution to global warming. To understand why summing CO2e emissions over time would distort the true global warming effect of product or system, we must dissect how the IPCC generates their GWPs. GWP is based on a ratio of cumulative radiative forcing (CRF) between some GHG and the reference gas, CO2, over a specified time horizon, typically 100 years. CRF is the integral of radiative forcing over a time horizon, and a measure of the capacity of a GHG to trap heat in the atmosphere over time. The trapping of heat in the atmosphere is one of the first steps in the climate change impact chain; thus, CRF is a causal indicator of the likely global warming effect of a GHG over time, not a measure of actual global warming or its impacts on climate. More accurate and complex methods have been proposed by other researchers that use the IPCC’s CRF indicator. For example, Levasseur et al. (2010) propose a method for calculating the actual global warming effect of different GHG emissions in global warming impact assessments.9 Their method is based on a computer model that dynamically assesses the CRF of each emission of GHG when it occurs. While conceptually and scientifically sound, their method cannot necessarily be implemented by other LCA practitioners and does not yield units that can be used in regulatory contexts, which typically rely on units of mass of CO2e. O’Hare et al. (2009) develop a method tailored to the problem of biofuels, proposing a new indicator for climate change effects: the fuel warming potential (FWP).10 This method considers timing in the context of the biofuels, namely corn ethanol, compared to gasoline using CRF, and is not broadly applicable to other LCA applications. Kendall et al. (2009) present a time correction factor (TCF) for emissions occurring at the outset of a life cycle resulting from capital investments or other processes such as land use change for biofuels.8 The TCF only addresses upfront emissions amortized over a prescribed time horizon, unlike the dynamic approach proposed by Levasseur et al. which can assess emissions occurring at any time in a life cycle. In addition Kendall et al. assumed an analytical time horizon equal to the amortization period. The methods proposed in this article enhance previous methods, particularly those presented in ref 8 by developing an approach to calculating time-corrected CO2e emissions intensity for both amortized upfront and EOL emissions for various analytical time horizons in units of CO2e. The approach uses scaling factors to facilitate implementation and conforms to units of CO2e which are mandated in current policy that addresses GHG emissions. 1.3. Emissions Timing in the Vehicle Life Cycle. For vehicles the problem of timing arises in a number of contexts, and recycling credits are one of the more important ones. There are a number of methods for treating recycling and recyclable materials in LCA, and all except the recycled content method (also referred to as the cutoff method) apply emissions credits to recyclable materials as if they were earned at the time of production. For vehicles, this is particularly important because recycling credits are influential in determining the

emissions intensity standard for vehicles. To demonstrate the necessity and feasibility of calculating life cycle GHG emissions for vehicles, GHG LCAs of two projected future vehicles are described. Then, two new global warming potential scaling factors are proposed that correct for the timing of emissions occurring at vehicle production and end-of-life. Combining the LCAs and scaling factors allows for consideration of (i) whether life cycle emissions may be important for avoiding perverse outcomes, and (ii) whether accounting for the timing of emissions over a vehicle’s life cycle is important for estimating the actual global warming potential (GWP) of that vehicle. 1.1. Vehicle Trends and Their Effect on Life Cycle GHGs. A number of advancements in powertrain technology and light-weighting technology have been modeled using LCA. Many studies demonstrate that these advancements will reduce total GHG emissions from the vehicle life cycle but will shift some emissions from the use phase to the production stage. Lower carbon fuels in internal combustion engine vehicles (ICEVs) or lower-carbon electricity grids for battery electric vehicles (BEVs) have the capacity to further reduce use-phase emissions, increasing the proportional contribution of nonuse (i.e., production and EOL) emissions. This phenomena is not discussed here but has been addressed in previous studies. Recent LCAs conducted to compare conventional and advanced powertrain vehicles, assuming current average fuel and electricity grid carbon intensities, found an increased contribution from the production stage, from about 15% to 30%, for advanced powertrain vehicles.5,6 Light-weighting and other strategies for reducing use-phase emissions result in similar outcomes, particularly if materials with high energy intensities and GHG emissions are used in lightweight designs. 1.2. The Effect of GHG Emissions Timing. LCA methods for GHGs typically track net emissions over some time horizon and report total summed emissions in CO2e values. CO2e emissions are usually calculated using the Intergovernmental Panel on Climate Change’s (IPCC’s) widely used 100-year GWPs.11 This method ignores the effect of GHG emissions timing, which when comparing or developing strategies for climate change mitigation can distort the outcome. Some researchers8,10,12 have observed that this omission can significantly distort the true climate change effects of emissions and may distort the benefits and drawbacks of technology and innovation alternatives. Similarly, carbon credits and exchanges may also be distorted due to the timing of avoided or credited emissions.7 The first published carbon footprint standard, PAS 2050, also recognizes the problem of summing GHG emissions over time horizons of a few years or more and calls out the particular problem of EOL processes. The standard addresses emissions timing by using a simplified equation based on the timing of a future emission and an abbreviated expression of CO2’s atmospheric lifetime.13 For continuous or more complex emissions profiles the PAS 2050 standard further simplifies characterization by simply using a weighted average of emissions’ time in the atmosphere over a 100-year time horizon and does not include an expression for CO2’s atmospheric lifetime.14 Similar to PAS 2050, the International Reference Life Cycle Data System Handbook (ILCD) also highlights the importance of considering the timing of EOL processes, specifically recycling, for long-lived products.15 ILCD recommends a simple linear scaling for emissions timing within a 100-year time horizon; in other words, recycling credits would be 2558

dx.doi.org/10.1021/es203098j | Environ. Sci. Technol. 2012, 46, 2557−2563

Environmental Science & Technology

Policy Analysis

emissions intensity of steel, aluminum, and magnesium, all important automotive materials. Figure 1 illustrates the effect on CRF of assigning recycling credits for 1 kg of steel at the time of production versus the

lighter weight magnesium, aluminum, and advanced highstrength steel (AHSS) parts were used in place of traditional materials, such as steel, throughout the vehicle, and a hybrid powertrain was used. The bill of materials (BOM) reported in ref 16 was linked to life cycle inventory data through the UCSB Advanced Powertrain Model (UCSBAPM) and to data sets from the Ecoinvent and PE databases.4,17,18 The vehicle use phase was also modeled based on total fuel cycle emissions from the UCSBAPM. Section S-1 in the Supporting Information provides a more detailed description of the LCA methods used, and Table S-1 in the Supporting Information provides the mass of materials and the life cycle inventory data sources used in the LCA model. Preliminary results for life cycle GHG emissions for the HD 2020 vehicle were calculated assuming 131 g CO2e/km based on an 11-year, 202,616 km life, and a 16-year, 261,358 km life.1 As shown in Table 1, the 11 year service life corresponds to a

Figure 1. The effect of recycling credit timing on CRF for 1 kg steel.

Table 1. Total Life Cycle Emissions by Stage for the HD and HD-LD 2020 Vehicles (kg CO2e)

time of recycling, assuming a 16-year vehicle lifetime. The CRF was calculated based on the model described in the Supporting Information, Section S3. The production emissions and recycling credits are based on flat carbon steel using consequential system expansion for recycling credits, where production emissions are approximately twice the magnitude of the recycling credit.4 Figure 1 demonstrates that crediting recycling at the time of production underestimates the true global warming effect of a material, and for materials where recycling credits are proportionally higher, timing is even more influential on the outcome. Thus the time between production of a material and its recycling should be accounted for in the calculation of a material’s net GWP. The current regulatory approach for vehicles requires the use of emissions-intensity metrics, where emissions associated with a product or system are averaged or amortized over some measure of productivity, in this case on a per-mile basis. Thus both production and EOL emissions must be amortized over the entire service life of the vehicle in order to calculate a vehicle’s life cycle emissions intensity. Accounting for the timing of emission will increase the global warming contribution of production or other upfront emissions and likewise reduce the contribution of recycling credits or disposal emissions. To address the effect of production and EOL timing, two correction factors are developed to be used concurrently that scale amortized upfront and EOL emissions to reflect their true GWP.

HD 2020 vehicle HD-LD 2020 vehicle

emissions from production

emissions from 11-year use phase

emissions from 16-year use phase

emissions from EOL

17,300

26,500

34,200

−6,400

9,900

28,200

36,400

−3,100

71% contribution of life cycle emissions by the use-phase, while the 16-year service life increases the proportion of emissions attributable to the use-phase to 76%. These proportions do not reflect the effect of emissions timing but rather the net flow of emissions. A second vehicle architecture was also modeled to test whether a lower-development vehicle which uses less advanced light-weighting techniques could result in a vehicle design with higher use phase emissions but lower total life cycle emissions. If such a scenario emerged, it would demonstrate the risk that a policy examining only use-phase emissions could result in perverse outcomes, where a vehicle with higher total emissions is preferred over one with lower emissions. The second vehicle architecture was developed by taking the HD 2020 vehicle and substituting a heavier low-development (LD) body structure. This body structure is taken from Lotus Engineering Inc.’s report which develops a LD 2020 Toyota Venza in addition to the HD vehicle already described.16 Here the combined LD body structure with the HD vehicle is referred to as the LD-HD 2020 vehicle. Table S-1 describes the changes to the original HD 2020 BOM caused by this substitution. The LD body structure is composed largely of mild steel and AHSS, adding approximately 100 kg to the weight of the HD 2020 vehicle (from approximately 1200 to 1300 kg). The increased weight reduces the vehicle fuel economy. The decrease in fuel economy, approximately 1.3 L/km (3 mpg), was estimated using the UCSBAPM model.4 As illustrated in Table 1, this leads to an increase in use phase emissions compared to the HD 2020 vehicle, resulting in use phase contributions of 81% for the 11-year service life and 84% for the 16-year service life. While the LD-HD 2020 vehicle design leads to increased use-phase emissions, it results in a net

2. METHODS 2.1. LCA of MY 2020 Light-Duty Vehicles. To demonstrate the importance of life cycle emissions and the application of time correction factors for vehicle GHG emissions standards, an LCA for a future vehicle, the highdevelopment (HD) model year 2020 Toyota Venza (referred to in this article as the HD 2020 vehicle) as described in Lotus Engineering Inc.’s 2010 report, was developed.16 Lotus Engineering Inc.’s analysis redesigns the current Toyota Venza given predefined cost targets to meet equivalent performance while achieving improved fuel economy through light-weighting and powertrain actions. Among other actions, 2559

dx.doi.org/10.1021/es203098j | Environ. Sci. Technol. 2012, 46, 2557−2563

Environmental Science & Technology

Policy Analysis

This production emissions time correction factor (pTCF) is shown in eq 1

reduction in total life cycle emissions of 2,300 and 1,800 kg CO2e for the 11- and 16-year service lives, respectively. Assessing the global warming effect of both vehicles using CRF, so that we may consider the timing of emissions over the vehicle life cycle, shows that the HD 2020 vehicle has 5% greater global warming 100 years in the future for the 16-year vehicle life and 7% greater for the 11-year vehicle life. The CRF model used for this calculation is described in the Supporting Information Section S-3. 2.2. Development of Scaling Factors for Emissions Timing in GHG Emissions-Intensity Calculations. In the context of calculating emissions intensity on an amortized basis (e.g., g CO2/year, or gCO2/km) for a vehicle or other products where EOL emissions may be important, we must use correction factors with equal analytical time horizons for both upfront emissions and EOL recycling8 proposed a time correction factor (TCF) for amortized upfront GHG emissions. The TCF is a unitless scaling factor, much like a GWP, and is applied simply by multiplying the amortized emission by the TCF. The TCF described in ref 8 ensures that at the end of the amortization period the CRF of the actual upfront emission and that of the amortized emission are equal. Implicitly, the TCF assumes that the analytical time horizon is equal to the period of amortization. While equating the analytical time horizon and amortization period may function for some limited applications, it is not reasonable for vehicles or many products for the following reasons: (1) GHG emissions occurring near the end or after the analytical time horizon are assigned little to no global warming effect because CRF is evaluated only between the time an emission occurs and the end of the analytical time horizon. If the analytical time horizon is too short, it could lead to short-sighted decision-making and an incentive for pushing emissions and climate change problems to an arbitrary time in the near future. (2) In the special case where the analytical time horizon and service life are equal, EOL emissions would be assigned no impact at all. This means that recyclability and other EOL processes are entirely omitted from global warming potential calculations. (3) Specified analytical time horizons that are independent of the application may already exist, for example the prescribed use of 100-year GWPs in U.S. and international climate-change related policies and agreements.

AT

∫0

pTCFAT =

RFCO2, t = 0(t )dt

( i AT RFCO2,t=0/AP(t)dt)

∑iAP =0 ∫

(1)

where RFCO2,t=0 refers to the radiative forcing attributable to the GHG flows occurring at the outset of the lifecycle, and RF(CO2,t=0)/AP refers to the radiative forcing for that same flow divided over the amortization period. The pTCF ensures the CRF of the actual emissions and amortized emissions are equal at the selected analytical time horizon. In order to facilitate the use of this correction factor, eqs 2−4 define pTCF as a function of the amortization period for three analytical time horizons: 30, 50, and 100 years. These equations are valid over amortization periods from 2 to 30 years. The equations were developed using least-squares fit for linear and second order polynomial equations. Tables S3 and S5 in the Supporting Information show additional information for the pTCF, including the goodness-of-fit for each equation and a look-up table for pTCF values. For most products, including vehicles, the amortization period will be equal to a product’s service life pTCF100 = 0.0045 × AP + 0.9921

(2)

pTCF50 = 0.0001 × AP2 + 0.0067 × AP + 0.9956

(3)

pTCF30 = 0.0007 × AP2 + 0.0045 × AP + 1.0188

(4)

2.1.2. Development of a Scaling Factor for End-of-Life Emissions. Emissions occurring at a vehicle’s retirement present additional challenges for considering emissions timing. As indicated in Figure 1, the practice of assigning emissions credits from recycling at the time of production will overestimate the benefits of recycling from a GWP perspective. A time correction factor for emissions occurring at the EOL, rTCF, can be defined much as the pTCF described in eq 1 AT

rTCFAT =

∫AP RFCO2, t = EOL(t )dt

( i AT RFCO2,t=EOL/AP(t)dt)

∑iAP =0 ∫

(5)

where RFCO2,t=EOL refers to the radiative forcing attributable to the GHG flows at EOL, including recycling credits or emissions; and RF(CO2,t=EOL)/AP refers to the radiative forcing for that same flow divided over the amortization period. Note, the amortization period is presumed equal to the service life, which in turn is equal to the time between production and EOL. Just as with the pTCF, we have generated a scaling factor that can be approximated by simple linear or polynomial expressions. Here it is defined for the same three analytical time horizons, 30, 50, and 100 years, as a function of the amortization period (in years). Tables S4 and S5 in the Supporting Information show additional information, including the goodness-of-fit for each equation and a look-up table for various amortization periods. Note that the time to EOL must be less than the analytical time horizon for these functions to apply. And, for compatibility with the pTCF, the same

For these reasons this study proposes scaling factors for both upfront and EOL emissions applicable to emissions intensity calculations for which the analytical time horizon and amortization period are variable. 2.2.1. Development of a Scaling Factor for Production Emissions. Recall that CRF is the time integration of a GHG’s radiative forcing (RF) over some analytical time horizon. We can define a time correction factor that corrects for amortized production emissions, similar to that proposed in ref 8 but which treats both the analytical time horizon (denoted by AT in equations, figures, and tables) and amortization period (denoted as AP in equations, figures, and tables) as variables. 2560

dx.doi.org/10.1021/es203098j | Environ. Sci. Technol. 2012, 46, 2557−2563

Environmental Science & Technology

Policy Analysis

Table 2. Life Cycle Emissions Intensity (g CO2e/km, g CO2e/mi Shown in Parentheses) with Time Correction Factors for the HD 2020 Vehicle HD 2020 vehicle description 16-year service life

11-year service life

prod

no time correction 100-year time correction 30-year time correction 30-year time correction, LCF no time correction 100-year time correction 30-year time correction 30-year time correction, LCF

66 70 85 85 85 89 99 99

use

(106) (113) (136) (136) (137) (143) (160) (160)

131 131 131 105 131 131 131 105

total

rec −25 −23 −17 −17 −32 −30 −26 −26

(211) (211) (211) (169) (211) (211) (211) (169)

(−39) (−37) (−28) (−28) (−51) (−49) (−42) (−42)

172 178 198 172 184 189 204 178

(278) (287) (309) (267) (297) (305) (329) (286)

% use 76% 74% 66% 61% 71% 69% 64% 59%

Table 3. Emissions Intensity (g CO2e/mi) for the LD-HD 2020 Vehicle HD 2020 vehicle, LD body structure description 16-year service life

11-year service life

no time correction 100-year time correction 30-year time correction 30-year time correction, LCF no time correction 100-year time correction 30-year time correction 30-year time correction, LCF

prod 38 40 49 49 49 51 57 57

use

(61) (65) (78) (78) (79) (82) (92) (92)

139 139 139 111 139 139 139 111

(6)

rTCF50 = − 0.0002 × AP2 − 0.0064 × AP + 1.0041 (7)

rTCF30 = − 0.001 × AP2 − 0.0018 × AP + 0.979

−12 −11 −11 −11 −15 −15 −13 −13

(−19) (−18) (−18) (−18) (−25) (−24) (−21) (−21)

165 169 177 149 173 176 184 156

(266) (271) (284) (239) (278) (283) (296) (251)

% use

change from HD 2020

84% 83% 79% 75% 81% 79% 76% 71%

−7 (−11) −10 (−15) −22 (−25) −23 (−28) −12 (−19) −14 (−22) −20 (−33) −22 (−35)

VMT are very small; less than 1% for 100-year CRF. The actual annual VMT modeled is based on average annual VMT by vehicle age, derived from ref 1 and is reported in Table S-2 of the Supporting Information. Table 2 reports the life cycle CO2e emissions for the HD 2020 vehicle on an emissions intensity basis (g CO2e/km and g CO2e/mi) for production, use, and EOL emissions and includes time corrections for 30- and 100-year time horizons. Implementing time correction factors leads to an increased contribution of vehicle production emissions to life cycle GWP (as reported in CO2e) and a decrease in the value of recycling credits compared to the same emissions profiles without correction factors. A lower-carbon fuel (LCF) scenario is also included and assumes a 20% reduction in fuel carbon-intensity. Table 2 confirms findings from previous LCAs indicating that for advanced vehicles around 30% of net life cycle emissions attributable to the production and recycling stages, this proportion rises to 40% when time correction factors with 30-year analytical time horizons are used. Uncertainty in the service life of vehicles contributes to variability in the outcomes, and a shorter service life increases the importance of production emissions. Table 3 shows the emissions intensity estimates for the LDHD vehicle. The last column in Table 3 shows that for all scenarios, the LD-HD vehicle has lower total life cycle emissions compared to the HD vehicle but higher use phase emissions. The reduction in emissions intensity for the LD-HD vehicle is as high as 23 g CO2e/km when lower carbon fuels and time correction factors are applied. Figure 2 shows CRF for the HD 2020 vehicle over a 16-year vehicle life and illustrates the effect of applying pTCF and rTCF for 30- and 100-year analytical time horizons. For timecorrected amortized emissions the CRF of the actual life cycle emissions and amortized life cycle emissions become equal at the end of the selected analytical time horizons, while uncorrected amortized emissions result in an underestimation

analytical time horizons must be selected for both factors when used in tandem rTCF100 = − 0.0045 × AP + 1.0074

total

rec

(224) (224) (224) (179) (224) (224) (224) (179)

( 8)

3. RESULTS 3.1. Application of pTCF and rTCF to Life Cycle Emissions for Model Year 2020 Light-Duty Vehicles. The pTCF and rTCF have been developed to address the challenge of creating emissions-intensity estimates for vehicles in units of g CO2e per distance traveled that better reflect their actual global warming potential. Equation 9 defines a method for calculating time-corrected global warming intensity on a permile basis in order to conform to current EPA standards which use these units. Equation 9 simplifies the vehicle use phase by assuming constant annual vehicle miles traveled (VMT) during each year of vehicle use ⎛ Gprod × pTCF + Guse + GEOL × rTCF ⎞ gCO2 e =⎜ ⎟ mile lifetime VMT ⎝ ⎠ (9)

where Gprod = GHG emissions attributable at the time of vehicle production, Guse = GHG attributable to vehicle use, such as fuel production and consumption, assumed to be constant over the service life, and GEOL = GHG emissions attributable at the time of vehicle recycling. The assumption that average annual VMT is constant introduces some uncertainty in CRF calculations. However, CRF modeling comparing actual annual VMT and constant VMT showed that differences introduced by assuming constant 2561

dx.doi.org/10.1021/es203098j | Environ. Sci. Technol. 2012, 46, 2557−2563

Environmental Science & Technology

Policy Analysis

selecting policies or mitigation strategies that do not achieve optimal outcomes because the 100-year time horizon distorts our preferences for how to best mitigate GHG emissions. On the other hand, the effect of long-lived gases on climate change may be underestimated if we focus on shorter time horizons, so any preference for a particular time horizon carries benefits and drawbacks. Despite the potential for distortions, the current use of a 100year time horizon in existing policies suggests that it may persist as the preferable analytical time horizon. To ensure that the choice of analytical time horizon does not lead to significant differences or distortions in the outcome of a study, we recommend applying shorter analytical time horizons for comparison. A more comprehensive solution might be to compare the CRF of systems over time rather than reporting outcomes in terms of CO2e, however. 4.2. Uncertainty Introduced by Using CO2e Emissions. The previous sections describing the development and application of the pTCF and rTCF only address CO2 and essentially assume that CO2e can be used in calculations that account for the timing of emissions over a product lifecycle. One reason for using CO2e is to keep calculations as simple as possible, which could be particularly important for policy implementation. If CO2e emissions cannot be used, each gas would require a separate correction factor or a more dynamic and flexible approach to calculating time-corrected global warming. In an attempt to validate the simpler approach of using CO2e, we undertook CRF modeling to determine whether using CO2e emissions in time correction problems introduces a significant distortion in outcomes. We developed a CRF model for long-lived GHGs in which radiative efficiencies and atmospheric lifetimes are defined along with indirect effects, where applicable. The following gases are included in the CRF model: CO2, CH4, N2O, PFC-14, SF6, CH2Cl2, CFC-11, CFC114, PFC-116, CFC-12, CFC-13, HCFC-22. This model was used to compare CRF for actual GHG emissions associated with automotive materials, fuel, and fuel combustion, with the CRF generated using CO2e emissions based on GWP100. Modeling details are provided in Section S-3 of the Supporting Information. These comparisons showed that using CO2e in these analyses introduces very small errors, in part because CO2 is the dominant emission, and because other important emissions, like SF6 emitted during magnesium production and perfluorocarbons during aluminum production, are longer lived than 100 years. If a system has large contributions of GHGs from shorter-lived gases, like CH4, this simplification could introduce significant error. 4.3. Significance of Life Cycle GHG Emissions and Timing in Light-Duty Vehicle Regulation. Results from GHG LCAs of the HD 2020 and LD-HD 2020 vehicles demonstrated (i) that production emissions will likely comprise a greater proportion of life cycle emissions in future years; (ii) if emissions timing is accounted for, the importance of nonuse phase emissions increases; and (iii) without regulatory approaches that include life cycle emissions there is a risk of favoring vehicles with higher life cycle emissions over those with lower life cycle emissions. The third conclusion is demonstrated by the results shown in Tables 2 and 3, which document emissions-intensity by life cycle stage for both the HD 2020 and LD-HD 2020 vehicles. In each scenario, the HD 2020 vehicle had lower use phase

Figure 2. Application of the pTCF and rTCF at 30- and 100-year analytical time horizons to the HD 2020 vehicle emissions profile with a 16-year service life.

of CRF. The difference between the uncorrected amortized emissions and corrected emissions at year 100 is small and may indicate that if a 100-year analytical time horizon is prescribed in policy, time correction factors are not warranted because they add complexity for policy implementation without having a significant effect on outcomes. At a 30-year analytical time horizon, however, the difference is more than 10% and suggests time correction should be considered in policies targeting GHGs from light duty vehicles. These conclusions are specific to this particular vehicle’s emissions profile and cannot necessarily be generalized to other emissions profiles, in particular, to future vehicles where production emissions may contribute an even greater proportion of life cycle emissions. As additional vehicle LCAs emerge that consider future technologies, fuels or power source, and architectures, time correction factors should be applied to help establish whether they are required for robust GWP calculations.

4. DISCUSSION 4.1. Selection of an Analytical Time Horizon. Selecting an analytical time horizon presents a challenge for policies that rely on CO2e or other global warming indicators based on similar principles, because any analytical time horizon is essentially arbitrary. The debate over analytical time horizons for evaluating climate change effects is not new and has been addressed both implicitly and explicitly in the literature discussing the use and limitations of the IPCC’s GWPs (e.g., refs 19−21 to name just a few). Despite this ongoing debate, regulation and governmental policy almost exclusively use GWP100 in GHG accounting (e.g., the Kyoto Protocol, the U.S. Greenhouse Gas Inventory Guidelines, and California’s Low Carbon Fuel Standard). The widespread use of 100-year time horizons has important ramifications on the outcomes of policies and analyses. For one, the analytical time horizon establishes what future point in time we are concerned about and overlooks all earlier points in time. Because of this, the contributions of GHGs with lifetimes well under 100 years (e.g., methane, methylene chloride, CFC-11, and others) will be underestimated. In addition, because global warming causes irreversible effects on the climate system, ignoring or underestimating the near-term impacts of shorter lived gases could mean that we fall short of climate change mitigation goals or occasionally end up 2562

dx.doi.org/10.1021/es203098j | Environ. Sci. Technol. 2012, 46, 2557−2563

Environmental Science & Technology

Policy Analysis

(6) Samaras, C.; Meisterling, K. Life Cycle Assessent of greenhouse Gas Emissions from Plug-in Hybrid Vehicles: Implications for Policy. Environ. Sci. Technol. 2008, 42 (9), 3170−3176. (7) Costa, P. M.; Wilson, C. An Equivalence Factor Between CO2 Avoided Emissions and Sequestration - Description and Applications in Forestry. Mitig. Adapt. Strat. Glob. Change 2000, 5, 51−60. (8) Kendall, A.; Chang, B.; Sharpe, B. Accounting for TimeDependent Effects in Biofuel Life Cycle Greenhouse Gas Emissions Calculations. Environ. Sci. Technol. 2009, 43 (18), 7142. (9) Levasseur, A.; Lesage, P.; Margni, M.; Deschênes, L.; Samson, R. j. Considering Time in LCA: Dynamic LCA and Its Application to Global Warming Impact Assessments. Environ. Sci. Technol. 2010, 44 (8), 3169−3174.. (10) O’Hare, M.; Plevin, R. J.; Martin, J. I.; Kendall, A.; Hopson, E., Proper Accounting for Time Increases Crop-Based Biofuels’ GHG Deficit versus Petroleum. Environ. Res. Lett. 2009, 4 (024001). (11) IPCC Fourth Assessment Report: Climate Change 2007 - The Physical Science Basis; Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, 2007. (12) Levasseur, A.; Lesage, P.; Margni, M.; Deschenes, L.; Samson, R. Considering Time in LCA: Dynamic LCA and it Application to Global Warming Impact Assessments. Environ. Sci. Technol. 2010, 44, 3169− 3174. (13) PAS 2050:2011 - Specif ication for the assessment of the life cycle greenhouse gas emissions of goods and services; British Standards Institution: London, 2011. (14) PAS2050:2008 Specif ication for the assessment of the life cycle grenhouse as emissions of goods and services; British Standards Institution: London, 2008. (15) International Reference Life Cycle Data System (ILCD) Handbook - General guide for Life Cycle Assessment - Detailed guidance; EUR 24708 EN; European Commission Joint Research Centre Institute for Environment and Sustainability: Luxembourg, 2010. (16) An Assessment of Mass Reduction Opportunities for a 2017−2020 Model Year Vehicle Program; Lotus Engineering Inc.; The International Council on Clean Transportation: San Francisco, 2010. http://www. theicct.org/sites/default/files/publications/Mass_reduction_final_ 2010.pdf (accessed month day, year). (17) ecoinvent Data v2.0; Swiss Centre for Life Cycle Assessment; Ecoinvent Centre: Duebendorf, Switzerland, 2008. (18) Diesel at Ref inery, U.S.; PE International: LeinfeldenEchterdingen, 2006. (19) Fearnside, P. M. Why a 100-Year Time Horizon should be used for GlobalWarming Mitigation Calculations. Mitig. Adapt. Strat. Glob. Change 2002, 7 (1), 19−30. (20) Fuglestvedt, J. S.; Berntsen, T. K.; Godal, O.; Sausen, R.; Shine, K. P.; Skodvin, T. Metrics of Climate Change: Assessing Radiative Forcing and Emission Indices. Clim. Change 2003, 58 (3), 267−331. (21) Sygna, L.; Fuglestvedt, J. S.; Aaheim, H. A. The adequacy of GWPs as indicators of damage costsincurred by global warming. Mitig. Adapt. Strat. Glob. Chang. 2002, 7 (1), 45−62.

emissions, while the LD-HD 2020 had lower life cycle emissions. If life cycle emissions were not accounted for in regulation, the HD vehicle would be favored over the LD-HD design. The GHG LCA was based on a bill of materials generated by computer-aided engineering software. The widespread use of such software in current product development may be an opportunity to streamline GHG LCAs, reducing the compliance burdens of automobile manufactures and their suppliers if life cycle emission standards are implemented. 4.3.1. Emissions Timing in Vehicle GHG Policy. Accounting for emissions timing will unambiguously increase the contribution of nonuse phase emissions to life cycle emissions intensity estimates. However, if a 100-year GWP is used, the differences introduced by scaling CO2e for emissions timing are relatively small, at least for the vehicle architectures modeled here. If shorter analytical time horizons are adopted, and as production-related emissions increase their proportional contributions to life cycle emissions, emissions timing becomes more important and time correction factors should be considered. While the research community has recognized shortcomings in the widespread use of a 100-year analytical time horizon, the mandated use of 100-year GWPs in domestic and international policy suggest that this analytical time horizon will continue to be used. Using multiple analytical time horizons, particularly shorter time horizons, or adopting a dynamic CRF modeling approach could help anticipate and avert unintended outcomes.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional data. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (530) 752-5722. Fax: (530) 752-7872. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by a grant from the American Iron and Steel Institute entitled “Evaluating Life Cycle Emissions for Passenger Vehicles and the Effect of Emissions Timing on the Climate Change Impact of Vehicles” (PI: Alissa Kendall).

■

■

NOTE ADDED AFTER ASAP PUBLICATION This paper published February 22, 2012 with incorrect versions of Figure 1 and 2. The correct version published February 23, 2012.

REFERENCES

(1) Davis, S. C.; Diegel, S. W.; Boundy, R. G. Transportation Energy Data Book, 29th ed.; Oak Ridge National Laboratory: Oak Ridge, TN, 2010. (2) Castro, M. B. G.; Remmerswaal, J. A. M.; Reuter, M. A. Life Cycle Impact Assessment of the Average Passenger Vehicle in the Netherlands. Int. J. Life Cycle Assess. 2003, 8 (5), 297−304. (3) Kim, H. C.; Keoleian, G. A.; Grande, D. E.; Bean, J. C. Life Cycle Optimization of Automobile Replacement: Model and Application. Environ. Sci. Technol. 2003, 37 (23), 5407−5413. (4) Geyer, R. UCSB Advanced Powertrain Model; Santa Barbara, CA, 2010. (5) Notter, D. A.; Gauch, M.; Widmer, R.; Wäger, P.; Stamp, A.; Zah, R.; Althaus, H.-J. r. Contribution of Li-Ion Batteries to the Environmental Impact of Electric Vehicles. Environ. Sci. Technol. 2010, 44 (17), 6550−6556. 2563

dx.doi.org/10.1021/es203098j | Environ. Sci. Technol. 2012, 46, 2557−2563