Evaluation of Life-Cycle Air Emission Factors of Freight Transportation

Sep 19, 2007 - Life Cycle Assessment of Creosote-Treated Wooden Railroad Crossties in the US with Comparisons to Concrete and Plastic Composite Railro...
0 downloads 11 Views 272KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 7138-7144

Evaluation of Life-Cycle Air Emission Factors of Freight Transportation CRISTIANO FACANHA ICF International, 394 Pacific Avenue, 2nd Floor, San Francisco, California 94111 ARPAD HORVATH* Department of Civil and Environmental Engineering, University of California, 215 McLaughlin Hall, Berkeley, California 94720-1712

Life-cycle air emission factors associated with road, rail, and air transportation of freight in the United States are analyzed. All life-cycle phases of vehicles, infrastructure, and fuels are accounted for in a hybrid life-cycle assessment (LCA). It includes not only fuel combustion, but also emissions from vehicle manufacturing, maintenance, and end of life, infrastructure construction, operation, maintenance, and end of life, and petroleum exploration, refining, and fuel distribution. Results indicate that total lifecycle emissions of freight transportation modes are underestimated if only tailpipe emissions are accounted for. In the case of CO2 and NOx, tailpipe emissions underestimate total emissions by up to 38%, depending on the mode. Total life-cycle emissions of CO and SO2 are up to seven times higher than tailpipe emissions. Sensitivity analysis considers the effects of vehicle type, geography, and mode efficiency on the final results. Policy implications of this analysis are also discussed. For example, while it is widely assumed that currently proposed regulations will result in substantial reductions in emissions, we find that this is true for NOx emissions, because fuel combustion is the main cause, and to a lesser extent for SO2, but not for PM10 emissions, which are significantly affected by the other life-cycle phases.

Introduction While transportation brings significant economic benefits to society, the environmental and social drawbacks (e.g., air emissions, energy use, noise, accidents, congestion, water runoff, land use) also need to be addressed. The Clean Air Act of 1970 and its subsequent 1977 and 1990 amendments enabled regulations to promote cleaner fuels, cleaner vehicles (e.g., cars, trucks, buses, locomotives, and aircraft), inspection and maintenance programs, and transportation policies favoring carpooling and transit alternatives (1). As a result, motor vehicle emissions of criteria pollutants were reduced by 77% between 1970 and 1999 (2), despite the increasing number of vehicle miles traveled. Emission standards for heavy-duty diesel trucks have not followed the same pace. Although significant reductions in truck diesel emission factors are yet to be realized, the prospects are positive. In 2003, the U.S. Environmental Protection Agency (EPA) issued new emission standards for * Corresponding author tel: 510-642-7300; fax: 510-643-8919; e-mail: [email protected]. 7138

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 20, 2007

model year 2007 and later heavy-duty trucks (3) which are expected to reduce emissions of NOx, PM10, HC, and SOx by about 90% compared to 2000 emission standards. The introduction of ultralow-sulfur fuels will also reduce SOx emissions dramatically, by about 2 orders of magnitude. Fuel efficiency of heavy-duty trucks has also been improving: between 1992 and 2002, improvements between 9 and 17% have been achieved. Despite these substantial reductions, fuel efficiency of the heaviest truck classes, which account for two-thirds of the fuel consumed by trucks, has remained relatively constant between 1997 and 2002 (4). The EPA has established three tiers of locomotive emission standards, dependent on the year the locomotive was first manufactured. These standards were enforced anytime the locomotive was remanufactured. Tier 0 applied to locomotives manufactured between 1973 and 2001, tier 1 applied to years 2002-2004, and tier 2 applied to locomotives originally manufactured in 2005 and later (5). Tier 2 emission standards of NOx and PM10 are 40 and 65% lower than the original tier 0 standards. The EPA is currently proposing more stringent locomotive emission standards for existing and new locomotives (6). Based on these standards, emissions of NOx and PM10 could be reduced by 80 and 90% compared to current standards. The International Civil Aviation Organization (ICAO) sets emission standards for jet engines, based on EPA regulations (7). NOx standards that went into effect for engines entering service in 2004 reflect a 16% reduction over the 1996 standards, and 33% over the original standards from 1981. The ICAO has also approved a further 12% reduction for 2008 engines.

Scope This study quantifies life-cycle emission factors (measured in grams of pollutant per ton-mile of freight activity) associated with road, rail, and air transportation of freight in the continental United States. All life-cycle phases of vehicles, transportation infrastructure, and fuels are considered, including the production, use, maintenance, and end of life of vehicles and infrastructure, as well as the life cycle of diesel and aviation fuels (petroleum extraction and refining, fuel distribution). The main goal of this analysis is to determine modespecific emission factors that account for all elements that are part of the provision of freight transportation (i.e., vehicles, infrastructure, fuels). These emission factors can then be used to quantify emissions from transportation systems. Intermodal operations, as well as first and last legs of intermodal moves, are not part of this study. Rail yard operations are included since rail cars are rearranged during shifting operations. The use of airport cargo terminals is also included because many flights are not direct and require handling operations at cargo terminals. Since trucking dominates the short- and medium-distance (less than 500 miles) transportation markets, this study focuses on long-distance transportation where road, rail, and air transportation compete. Inland shipping is not included as it is generally limited geographically (e.g., along the Mississippi River). This paper summarizes the findings from Facanha (8), whose preliminary results have been published in reference 9. The latter describes the assumptions, modeling methods, parameters, and data sources. The current paper builds upon the modeling details of Facanha (9), but expands the modeled variables to account for differences in vehicle type, geography, and mode efficiency. Policy implications of life-cycle con10.1021/es070989q CCC: $37.00

 2007 American Chemical Society Published on Web 09/19/2007

TABLE 1. Input Parameters for Baseline Scenario road

rail

air

vehicle type

variable

model 2000 Class 8b heavy-duty diesel truck

Boeing 747-400 dedicated to freight operations (conforming to 1981 standards)

equipment capacity equipment utilization empty miles average speed infrastructure

22.3 tons 75% 25% 50 mph leveled four-lane asphalt highway

4,000 hp diesel-electric locomotives (conforming to tier 0 standards) and 70 rail cars (double-stack) conforming to tier 0 standards 3,721 tons 75% 25% 30 mph leveled rails and wood crossties built on top of 3 subbase layers

siderations applied to freight transportation are also discussed. Life-cycle emission factors are presented for a variety of scenarios that depend on vehicle capacity and utilization. Input parameters associated with a baseline scenario are included in Table 1. The remaining assumptions were reported previously (8, 9). The focus on air emissions is justified as they are at the forefront of policy debates regarding environmental effects of transportation (e.g., fuel efficiency standards, hydrogen economy, fuel cells, oil dependency, global warming). The analysis accounts for CO2, NOx, PM10, CO, and SO2 emissions. Attempts were made to include additional pollutants such as PM2.5, but data were not available for all the processes included in the scope of this study. In order to combine components (e.g., vehicle manufacturing, infrastructure maintenance) into a single metric (grams of pollutant/ton-mile), emissions from different components are divided by the freight activity incurred in the typical lifetime of a component. For example, locomotives are commonly assumed to have a lifetime of 22 years. By spreading emissions associated with locomotive manufacturing over the freight activity of 22 years, an emission factor in grams of pollutant/ton-mile can be calculated. Allocation rules are set in place to account for the fact that infrastructure is shared between passenger and freight transportation. The allocation of road infrastructure to freight is calculated according to the share of the damage that trucks impose on roads (10). Since only Class I rail is considered, rail infrastructure is 100% dedicated to freight. Air infrastructure is allocated according to landed weight at airports. Only a few publications provide environmental comparisons of different freight transportation modes through more than just tailpipe emissions. Stodolsky (11) concluded that in the United States freight trains have three times less associated energy and emissions than trucks on a ton-mile basis. The analysis included vehicle manufacturing, use, and precombustion processes, but did not account for the transportation infrastructure. To date the most comprehensive European life-cycle inventory of air emissions from road, rail, and water transportation of goods has been published by Spielmann (12). The study is based on aggregate national data, and uses life-cycle assessment (LCA) to quantify air emissions from all life-cycle phases of vehicles, infrastructure, and fuels. The analysis focuses on the percent share of each life-cycle phase in the total rather than on a comparison of modes based on absolute numbers. Also in Europe, Marheineke (13) concentrated his analysis on trucks and roads in Germany. A comprehensive list of relevant studies has been compiled by Facanha (8). This paper refrains from supporting any particular mode of freight transportation for several reasons. First, this research concentrated on long-distance transportation, which may not be relevant for transportation corridors that are less than 500 miles. Second, intermodal considerations were not taken into account. Since rail and air transportation depend on drayage movements to handle the first and last

124 tons 75% 25% 546 mph based on average airport

legs, a direct comparison of modes is not meaningful. Third, not all modes compete for the same routes and the same markets due to factors such as economics. For example, air and rail transportation are generally not in direct competition. Finally, attention should also be given to local impacts of specific transportation modes (e.g., noise), which are beyond the scope of this analysis.

Methodology LCA is a process that evaluates and quantifies environmental impacts associated with a product, process, or service throughout its lifetime, from design to end of life (14). In a conventional process-based LCA a product is decomposed into processes and activities, which are associated with inputs (e.g., energy, water, raw materials) and outputs (e.g., environmental releases, byproducts). By normalizing all results to a common functional unit, the total environmental impact associated with a product can be determined. A processbased LCA enables very detailed analysis, but can be expensive, time-consuming, and might prove infeasible due to data quality. The issue of system boundaries is a common problem as it is practically very difficult to include all upstream suppliers. Since studies usually include different boundaries, it is difficult to compare their results. Some of these issues have been overcome with the introduction of the economic input-output analysis-based LCA (EIO-LCA) (15). It combines environmental data with the economic input-output tables of an economic system (e.g., the United States), and quantifies environmental releases associated with the production of one commodity. Since this model accounts for the interdependencies among all economic sectors, upstream suppliers are included in the results. For example, the analysis of the production of a vehicle accounts for all environmental impacts associated with the supply chain of this activity. Results from EIO-LCA models can be comprehensive, but they might not be very accurate if a commodity falls within a sector that is broadly defined or if the data are of poor quality. Since the provision of freight transportation relies on vehicles, transportation infrastructure, and fuels, it is very difficult to do a purely process-based LCA analysis. At the same time, the aggregation of commodities in EIO-LCA hinders an analysis based solely on the economic inputoutput model. A hybrid LCA, combining the two methods, is chosen for this study as it overcomes most drawbacks from both methods while leveraging their advantages (16). Process-based LCA is used whenever there are accurate data available. One example is the use phase of vehicles for all three transportation modes. Whenever process data are not available, EIO-LCA is used in order to account for upstream processes and suppliers. As values associated with an economic sector are averages across all enterprises falling under that sector, such values might not be representative if there is significant variation of production technologies and/or business practices within the sector. This is, for VOL. 41, NO. 20, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7139

TABLE 3. Commodity Density Correction Factors

road rail air

commodity density (ton/m3)a

CO2

NOx

PM10

CO

SO2

0.15 0.10 0.20 0.32 0.20 0.40 0.12 0.08 0.16

0% 28% -14% 0% 44% -15% 0% 37% -18%

0% 28% -14% 0% 44% -16% 0% 36% -18%

0% 12% -6% 0% 41% -14% 0% 4% -2%

0% 43% -22% 0% 42% -15% 0% 9% -4%

0% 38% -19% 0% 44% -16% 0% 11% -6%

a Density of selected commodities: wood products/furniture, 0.11; mixed freight, 0.34; gasoline, 0.42; coal, 0.85; gravel, 1.28.

on the emission factors in Table 2) to account for different commodity densities.

Phase Comparison FIGURE 1. Mode comparison.

TABLE 2. Emission Factors (grams/ton-mile) payload (tons) road Class 8b

12.5

Class 5

3.1

Class 2b

1.6

rail

Intermodal Rail

2,093

air

Boeing 747-400

70

DC 10

52

Boeing 757

22

CO2 187 76% 230 68% 289 62% 40 89% 1,358 70% 1,473 72% 1,577 73%

NOx PM10 2.57 93% 1.63 84% 1.77 80% 0.74 98% 6.18 69% 7.02 72% 5.74 65%

0.35 18% 0.47 33% 0.59 42% 0.05 89% 0.81 4% 0.78 0% 0.81 3%

CO

SO2

0.60 35% 1.20 34% 1.84 29% 0.42 92% 5.87 14% 6.70 24% 6.21 16%

0.15 29% 0.30 36% 0.45 35% 0.12 89% 2.21 14% 2.27 15% 2.34 16%

example, the case for the maintenance of rail tracks and airports. In those cases, hybrid LCA is used by incorporating process data in the analysis.

Emission Factors Figure 1 shows a comparison of mode-specific emission factors for a baseline scenario that aims to represent standard industry practices. Due to the focus on long-distance transportation, larger vehicles tend to be the norm. The variation bars account for intramodal variations due to different vehicle types (e.g., small trucks versus large trucks), as well as parameter uncertainty. Rail scores better than road for all five emissions by a factor of 1.2-7. Air transportation is 2-15 times worse than road transportation. The explanation is twofold: not only is air transportation very energy intensive due to technological matters, but all emissions are spread across a smaller base since the cargo handled by air is light relative to its volume. Table 2 provides emission factors for a selected group of vehicle types based on 2005 models. The tailpipe share of total life-cycle emissions is presented immediately below the emission factors. Because increases in vehicle payload can lead to lower fuel consumption per ton-mile, there is a wide range of variation in emission factors depending on vehicle capacity and utilization. There are large differences in transportation-related energy intensity and emissions depending on the commodity shipped. Table 3 includes correction factors (to be applied 7140

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 20, 2007

For all three modes, the burning of transportation fuel is the dominant life-cycle phase for CO2 and NOx. This is also the case for CO, PM10, and SO2 in rail transportation, which changes previous notions that the rail infrastructure might have higher relative impacts than the road and air infrastructure. Figure 2 shows a phase comparison for all pollutants and modes, according to the baseline scenario. Tailpipe emissions are responsible for 70-90% of total CO2 emissions for all three modes. Aircraft manufacturing, maintenance, and end of life have twice the share of road and rail vehicle phases due to the much higher costs associated with aircraft on a ton-mile basis. Pre-combustion processes are similar for all three modes, representing about 5% of total emissions. Fuel combustion accounts for 70-98% of total NOx emissions. The share of vehicle-related NOx emissions is smaller than for CO2 as manufacturing processes have relatively low NOx emissions. Emissions associated with the air transportation infrastructure are proportionally higher than for road and rail due to the use of ground support equipment, which are mostly small trucks that emit large amounts of NOx during the operations phase. This also explains why fuel combustion of cargo aircraft has a lower share than the operation of road and rail equipment. The use of alternative fuels to reduce emissions from ground support equipment was not taken into account in this analysis. Infrastructure is the dominant phase for PM10 emissions in both road and air transportation due to construction processes, which emit large amounts of particulate matter, as well as ground support equipment at airports. The infrastructure share of PM10 in rail transportation is lower because only the subbase portion of its infrastructure is intensive in terms of PM10 emissions. Vehicle manufacturing processes are also contributors of particulate matter emissions. Overall, there is substantial inaccuracy regarding PM10 emissions associated with the fuel combustion from air transportation since PM10 emission factors for cruise mode were not found in the literature. Due to reductions in CO emissions from road transportation, the use phase of vehicles has a less than 50% share of the total emissions. Air transportation follows the same trend, which makes the remaining life-cycle phases more important. Due to less strict regulations, road and air vehicle manufacturing have higher CO emissions than fuel combustion. The case is similar for air infrastructure and ground support equipment at airports as well as for tailpipe emissions from rail transportation.

FIGURE 2. Share of life-cycle phases in transportation air emissions. The air transportation infrastructure and road vehicle manufacturing have dominant roles in SO2 emissions as road diesel and jet fuel tend to have low levels of sulfur. This is not the case for rail diesel, which has had higher levels of sulfur than road diesel. Vehicle end of life (less than 0.1% of the total for most pollutants) has no significant impact on vehicle phase emissions for any of the modes. The ratio between manufacturing and maintenance is very different depending on the mode, and is determined by a similar trend in life-cycle costs. Maintenance of rail rolling stock has a higher share of the total than for road transportation due to the much higher mileage (and longer lifetime) achieved by locomotives than tractors. While manufacturing costs dominate the life-cycle of a truck, the opposite is true for an aircraft with high upfront costs, but even higher maintenance costs in order to extend the aircraft’s lifetime through regular and preventive maintenance. The rail infrastructure scores better than the road and air infrastructure. It is not uncommon to hear arguments in favor of rail due to the assumption that rail requires a higher upfront investment than road, but demands lower levels of subsequent maintenance. While this might be true financially, the environmental results suggest differently. Emissions associated with the construction of rail infrastructure are lower than those associated with the road infrastructure (on a ton-mile basis), while the maintenance of rail infrastructure does not necessarily have a lower share of total infrastructure emissions than the maintenance of other transportation modes. For example, the relative share of rail infrastructure maintenance is higher than the share of road infrastructure maintenance for NOx, PM10, and SO2. Even though processes in a petroleum refinery are different for diesel and jet fuels, they do not have a strong influence on the results. The contribution of fuel refining emissions to total pre-combustion emissions is very similar for all modes. Fuel refining is the dominant phase in precombustion for all pollutants except NOx. One of the most important contributions of this study is to demonstrate the importance of accounting for total lifecycle emissions. Fuel combustion is not the dominant phase for several pollutants. In terms of total CO2 emissions, there is an increase between 12 to 43% over tailpipe emissions, depending on the mode. With respect to NOx emissions, total rail transportation emissions are underestimated by less than 2%, but for road and air transportation they are underes-

timated by 7 to 31%, respectively, if only fuel combustion emissions are considered. While the difference between fuel combustion and total emissions for rail transportation is approximately 10% for CO and PM10, respectively, the gap is much wider for road and air transportation. Due to strict regulations of CO for fuel combustion of tractors, truck manufacturing has substantial emissions of CO, representing more than 50% of total emissions. Thus total emissions of CO for road transportation are almost three times higher than fuel combustion emissions. In the case of air transportation, total emissions of CO are almost 7 times higher than fuel combustion emissions (due to the operations of ground support equipment). In terms of particulate matter, road infrastructure is the dominant phase, explaining the almost 6-fold difference over tailpipe emissions. In air transportation the wide difference between total and tailpipe emissions of particulate matter is due to airport infrastructure (especially runways) and to the fact that there are no available PM10 emission factors for air cruise mode.

Sensitivity Analysis Although the results presented herein are based on a scenario that represents standard industry practices, it is worthwhile to examine the influences of different parameters on the results. This analysis focuses on three key parameters: vehicle type, geography, and mode efficiency. Vehicle Type. Vehicle type can have an important impact on the environmental performance of a transportation mode due to differences in fuel consumption and capacity. For example, the operational requirements associated with the movement of an intermodal container are different from those of a coal gondola. This is due to differences in fuel consumption coming from different aerodynamic coefficients and vehicle weights. Additionally, rail yard operations might vary due to load prioritization and train sequencing. Three scenarios are created to represent a range of intramodal variation. Baseline scenarios are not based on an average vehicle size, but on a vehicle that is the most representative in the industry. As the focus of this study is long-distance transportation, larger vehicles are more common and are selected to represent the baseline case. Road and air transportation scenarios are differentiated based on vehicle capacity. Heavy-duty trucks of class 2b and class 5 are considered in addition to the class 8b truck that was taken as the baseline. A Boeing 747-400 (baseline), a Boeing VOL. 41, NO. 20, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7141

TABLE 4. Variations from Vehicle Type rail transportation road transportation air transportation

CO2

NOx

PM10

CO

SO2

2% 55% 16%

2% -37% 14%

2% 69% -4%

2% 200% 14%

2% 200% 6%

757, and a DC 10 characterize the air transportation scenarios. Because rail transportation has more flexibility to adjust train capacity, the scenarios are differentiated by rail car type. While the baseline case considers intermodal rail cars to provide some level of competition with trucking, the two additional scenarios include coal gondolas and general cargo on flat cars. Vehicle production, vehicle use (fuel combustion), and pre-combustion processes are differentiated among scenarios due to differences in vehicle configuration and fuel consumption. It is assumed that the use of different vehicle types does not have an influence on infrastructure. Table 4 presents the variations among scenarios. Intramodal variations were the highest for road transportation, ranging from 37% for NOx to 200% for CO emissions. Emissions associated with air transportation also vary considerably, but generally less than in road transportation. This is the case because fuel combustion and vehicle production, which are the phases that are more sensitive to the scenario differences, represent a smaller share in air transportation than in road transportation. Rail shows the least variations due to the fact that it is not the size of the train that varies, but its rail car types. All pollutants varied by less than 3%. This is an indication that the payload has a much larger impact on emission factors than vehicle type. Geography. Although this model has a national focus, it includes parameters that enable its application to a regional environment. While emissions associated with the use phase are generally independent of location, upstream emissions are only weakly dependent on where the vehicle use occurs. Evaporative emissions, which depend on ambient temperature, are not part of this model. Geography affects results in two ways: infrastructure maintenance requirements, which are significantly different depending on the location, as well as road and rail track grades, which have a strong influence on fuel consumption. Monte Carlo simulation was called upon to evaluate the influence of input variables on the final results. This enables the calculation of pollutant elasticities, which determine how final results will vary with a 1% increase in a given input variable. For example, when truck emission factors are raised by 10%, final CO2 emissions for road transportation increase by 7.9% (elasticity: 0.79). The CO2 elasticity of infrastructure maintenance is around 0.01, therefore regional variations do not have a strong impact on results. Road and rail track grades, on the other hand, have a strong influence on fuel consumption, which in turn has high CO2 elasticities (0.550.75). However, the interstate system is aligned in such a way that steep grades are not typical (grades above 5% are uncommon). The same is generally true for railways. Based on the above, geography does not play a substantial role in differentiating emissions. Geography does play a crucial role in air quality and health effects, and local impacts from freight transportation can be significantly different depending on the exact location. Cities and neighborhoods that have a strong presence of vehicle manufacturing plants suffer a disproportionate burden due to local emissions. The same is true for areas around freight terminals and adjacent to heavy-traffic highways. When evaluating support for specific transportation policies, it is important to understand its local impacts. Even if a specific 7142

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 20, 2007

mode results in lower emissions on an aggregate level, local impacts can still be quite large (e.g., areas around rail yards). Mode Efficiency. Equipment utilization and the share of empty miles have an important influence on emissions. The CO2 elasticity for equipment utilization is almost 0.6 for all modes, while the share of empty miles has CO2 elasticities of 0.09 for rail and about 0.3 for road and air transportation. In other words, improvement in equipment utilization and a reduction in empty backhaul can lead to substantial reductions in emissions. Transportation policy has traditionally not addressed these factors since the private industry has always been motivated to improve equipment utilization and reduce empty backhaul for economic considerations.

Policy Implications The research analyzed herein quantifies the differences between total life-cycle emissions and fuel combustion emissions from freight transportation. These differences are very important given that transportation policies targeting environmental performance have been focused on fuel combustion emissions. The examination of a simple example can shed light on the importance of accounting for total life-cycle emissions when developing transportation policies to address emission reduction strategies for freight transportation. An increase in truck capacity can lead to lower fuel consumption per ton-mile, and an increase in road capacity (17). If only fuel combustion emissions are accounted for, this measure can result in lower emissions. However, it is also important to consider infrastructure-related emissions since larger trucks and higher weight limits impose a heavier load on the infrastructure, requiring thicker pavements, larger-load bearing bridges and overpasses, and more frequent maintenance. Assuming that pavement wear and tear is proportional to weight per axle to the fourth power (18), it is possible to determine how emissions vary with an increase in equipment size. For example, if a 5-axle truck capacity is increased by 20%, CO2 and CO total emissions are reduced by about 3%, NOx emissions are reduced by approximately 11%, but SO2 and PM10 emissions are increased by 7% and almost 80%, respectively. If such increase in capacity is implemented jointly with the addition of one more axle, emissions from all pollutants are reduced by 4-16%. This points to the importance of accounting for total emissions when evaluating different measures for emissions reduction. Results confirm the reasonable assumption that reducing emission factors from fuel combustion emissions and increasing vehicle fuel efficiency are two of the most effective strategies to reduce total emissions. This is especially true for CO2 and NOx emissions since fuel combustion emissions have a higher share of total emissions than the remaining life-cycle phases. It is important to emphasize that for pollutants whose emissions are due significantly to the infrastructure, vehicle manufacturing and maintenance, and pre-combustion phases (PM10 and SO2), focusing on fuel combustion emission factors can be insufficient. Embracing a Life-Cycle Perspective. There are several benefits from shifting transportation policy toward embracing a life-cycle perspective. First, it is necessary to address emissions associated with other components such as vehicle manufacturing and infrastructure. Depending on the pollutant, the use of vehicles is not the most dominant phase, therefore, emissions might be largely underestimated if only fuel combustion is considered. As use-phase emissions associated with vehicles are reduced due to stricter standards, other phases will become more prominent if not assessed and properly managed (as has been shown in the case of buildings in refs 19 and 20). Second, equity considerations are different depending on the emission source and location. For example, emissions associated with vehicle manufactur-

It is widely assumed that currently proposed regulations will result in substantial reductions in total emissions. This is true for NOx emissions since fuel combustion is the main source, and to a lesser extent for SO2, but not for PM10 emissions, which are significantly affected by other life-cycle phases. Increases in freight activity and the need for more infrastructure, vehicles, and fuels will likely outpace the reduction of PM10 emissions due to stricter tailpipe standards.

Outlook While this paper shows the environmental benefits of railways in freight transportation, it should be noted that the analysis concentrated on the long-distance market, and that neither drayage movements nor intermodal operations were accounted for. A time-, cost- and environmentally efficient transportation system should rely on a combination of all modes. Future work needs to closely monitor the implementation of new emission standards because substantial improvements are planned for both locomotive and truck engines. The same is true for the introduction of alternative and ultralow-sulfur fuels. Future research can also complement this study by assessing additional emissions and other environmental effects, as well as other modes and intermodal operations. It is also necessary to evaluate the impacts of the emissions inventoried in this study. This task is complicated enough for impact categories such as global warming potential and acidification for which there are inventories of emissions. For other categories, such as human toxicity potential, even the inventories of most of the toxic chemicals are missing (23). Given the importance of freight transportation in our society, these difficult but important tasks need to be taken on as soon as possible.

Acknowledgments

FIGURE 3. Impact of emission standards on U.S. national inventories. ing are geographically concentrated, imposing a burden on nearby communities which suffer disproportional health and other environmental impacts. Similarly, infrastructure construction is typically temporally “concentrated” (such as the construction of the Interstate highway system over about three decades in the latter half of the last century), affecting different generations in different ways. Third, equal treatment should be given to different transportation modes. By limiting regulations to tailpipe emissions, modes whose emissions fall under categories other than fuel combustion benefit unfairly due to lack of attention to those other categories. Finally, transportation policy should provide the public with more accurate information on the total environmental impacts associated with different transportation modes as well as alternatives (such as telework (21, 22)) so as to promote more optimal choices for society’s benefit. To evaluate the benefits of newly informed transportation policies, we quantify the impacts of new emission standards for locomotive and truck diesel engines on total emissions of NOx, PM10, and SO2. Two scenarios are compared. Scenario A assumes the traditional focus on fuel combustion emissions according to proposed emission standards for locomotives and trucks (3, 5). In Scenario B, transportation policy assumes a life-cycle perspective, and applies the same improvements to total emissions. Figure 3 illustrates how nationwide emissions of NOx, PM10, and SO2 can be reduced (compared to year 2000 levels) due to proposed emission standards for new locomotives and truck engines. The analysis assumes a 2% annual increase in freight activity, and that the share of new locomotives and trucks increases by 5% annually.

C.F. gratefully acknowledges the financial support of the National Science Foundation’s Graduate Student Fellowship. A.H. is grateful for the support of the Technologies for Sustainable Enterprise grant (R-82959701-0) from the U.S. EPA. In addition, this material is based upon work supported by the National Science Foundation under Grant 0094022.

Literature Cited (1) U.S. Environmental Protection Agency. The plain English guide to the Clean Air Act; EPA-456/K-07-001; U.S. EPA: Washington, DC, 2007. (2) Jensen, G. Air quality and transportation. Public Roads 2003, 67 (1); http://www.tfhrc.gov/pubrds/03jul/10.htm. (3) U.S. Environmental Protection Agency. Proposal for cleaner heavy-duty trucks and buses and cleaner diesel fuel - Fact sheet; U.S. EPA: Washington, DC, 2000; http://www.epa.gov/otaq/ regs/fuels/diesel/factshet.pdf. (4) U.S. Census Bureau. Vehicle inventory and use survey - 2002 data releases; http://www.census.gov/svsd/www/vius/2002. html. (5) U.S. Environmental Protection Agency. Regulatory announcement: Final emissions standards for locomotives; EPA-420-F97-048; U.S. EPA: Washington, DC, 1999. (6) U.S. Environmental Protection Agency. Regulatory announcement: EPA proposal for more stringent emissions standards for locomotives and marine compression-ignition engines; EPA-420F-07-015; U.S. EPA: Washington, DC, 2007. (7) FAA. Aviation & emissions - A primer; U.S. Department of Transportation, Federal Aviation Administration, Office of Environment and Energy: Washington, DC, 2005; http:// www.faa.gov/regulations_policies/policy_guidance/envir_policy/ media/AEPRIMER.pdf. (8) Facanha, C. Life-cycle inventory of freight transportation in the United States. Ph.D. dissertation, Department of Civil and Environmental Engineering, University of California, Berkeley, 2006. (9) Facanha, C.; Horvath, A. Environmental assessment of freight transportation in the United States. Int. J. LCA 2006, 11 (4), 229-239. VOL. 41, NO. 20, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7143

(10) March, J. Federal highway cost allocation study. Public Roads 1998, 61 (4), 44-50. (11) Stodolsky, F.; Gaines, L.; Cuenca, R.; Eberhardt, J. Lifecycle analysis for freight transport. Paper no. 982206 in Proceedings of the 1998 Total Life-cycle Conference; Society of Automotive Engineers, Inc.: Warrendale, PA, 1998. (12) Spielmann, M.; Scholz, R. Life cycle inventories of transport services - Background data for freight transport. Int. J. LCA 2005, 10 (1), 85-94. (13) Marheineke, T.; Friedrick, R.; Krewitt, W. Application of a hybridapproach to the life-cycle inventory analysis of a freight transport task. Paper no. 982201 in Proceedings of the 1998 Total Lifecycle Conference; Society of Automotive Engineers, Inc.: Warrendale, PA, 1998. (14) Hendrickson, C.; Horvath, A.; Joshi, S.; Lave, L. Economic inputoutput models for environmental life-cycle assessment. Environ. Sci. Technol. 1998, 32 (4), 184-191. (15) Hendrickson, C. T., Lave. L. B., Matthews, H. S., Eds. Environmental Life Cycle Assessment of Goods and Services: An InputOutput Approach; Resources for the Future: Washington, DC, 2006. (16) Suh, S.; Lenzen, M.; Treloar, G. J.; Hondo, H.; Horvath, A.; Huppes, G.; Jolliet, O.; Klann, U.; Krewitt, W.; Moriguchi, Y.; Munksgaard, J.; Norris, G. System boundary selection in lifecycle inventories using hybrid approaches. Environ. Sci. Technol. 2004, 38 (3), 657-664.

7144

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 20, 2007

(17) Facanha, C.; Horvath, A. Environmental assessment of logistics outsourcing. J. Manage. Eng. 2005, 21 (1), 27-37. (18) American Association of State Highway and Transportation Officials. A policy on geometric design of highways and streets; AASHTO: Washington, DC, 1990. (19) Guggemos, A.; Horvath, A. Comparison of environmental effects of steel- and concrete-framed buildings. J. Infrastructure Syst. 2005, 11 (2), 93-101. (20) Junnila, S.; Horvath, A.; Guggemos, A. Life-cycle assessment of office buildings in Europe and the U.S. J. Infrastructure Syst. 2006, 12 (1), 10-17. (21) Kitou, E.; Horvath, A. Transportation choices and air pollution effects of telework. J. Infrastructure Syst. 2006, 12 (2), 121134. (22) Kitou, E.; Horvath, A. External air pollution costs of telework. Int. J. LCA 2007, DOI: http://dx.doi.org/10.1065/lca2007.06.338. (23) Horvath, A.; Hendrickson, C. T.; Lave, L. B.; McMichael, F. C.; Wu, T-S. Toxic emissions indices for green design and inventory. Environ. Sci. Technol. 1995, 29 (2), 86-90.

Received for review April 26, 2007. Revised manuscript received August 8, 2007. Accepted August 10, 2007. ES070989Q