Article pubs.acs.org/est
Emissions and Air Quality Impacts of Truck-to-Rail Freight Modal Shifts in the Midwestern United States Erica Bickford,†,‡ Tracey Holloway,*,†,‡ Alexandra Karambelas,† Matt Johnston,†,‡,§ Teresa Adams,‡ Mark Janssen,⊥ and Claus Moberg† †
Center for Sustainability and the Global Environment (SAGE), Nelson Institute for Environmental Studies and ‡National Center for Freight and Infrastructure Research and Education (CFIRE), University of WisconsinMadison, Madison, Wisconsin 53726, United States § Institute on the Environment (IonE), University of Minnesota, St. Paul, Minnesota 55108, United States ⊥ Lake Michigan Air Directors Consortium (LADCO), Rosemont, Illinois 60018, United States S Supporting Information *
ABSTRACT: We present an examination of the potential emissions and air quality benefits of shifting freight from truck to rail in the upper Midwestern United States. Using a novel, freight-specific emissions inventory (the Wisconsin Inventory of Freight Emissions, WIFE) and a three-dimensional Eulerian photochemical transport model (the Community Multiscale Air Quality Model, CMAQ), we quantify how specific freight mode choices impact ambient air pollution concentrations. Using WIFE, we developed two modal shift scenarios: one focusing on intraregional freight movements within the Midwest and a second on through-freight movements through the region. Freight truck and rail emissions inventories for each scenario were gridded to a 12 km × 12 km horizontal resolution as input to CMAQ, along with emissions from all other major sectors, and three-dimensional time-varying meteorology from the Weather Research and Forecasting model (WRF). The through-freight scenario reduced monthly mean (January and July) localized concentrations of nitrogen dioxide (NO2) by 28% (−2.33 ppbV) in highway grid cells, and reduced elemental carbon (EC) by 16% (−0.05 μg/m3) in highway grid cells. There were corresponding localized increases in railway grid cells of 25% (+0.83 ppbV) for NO2, and 22% (+0.05 μg/m3) for EC. The through-freight scenario reduced CO2 emissions 31% compared to baseline trucking. The through-freight scenario yields a July mean change in ground-level ambient PM2.5 and O3 over the central and eastern part of the domain (up to −3%).
■
INTRODUCTION
and carbon emissions. However, the net impact of truck-to-rail modal shift must be carefully evaluated to inform policy-making relevant to freight activity patterns. Modal-shift choices are made by individual shipping companies, but federal policy can incentivize a freight modal shift. A 2013 report by the Department of Energy 2 highlighted particular policy strategies to provide disincentives for trucking in favor of rail. These include motor fuel taxes, tolls, carbon taxes (or other emission-based fees), further restrictions on permissible hours for truck drivers, reducing truck size and weight limits, and public investment in freight rail infrastructure.2 Whether or not such policies and associated investments are worthwhile fundamentally depends on the public benefits achieved by moving freight by train versus truck.
Freight transport is vital to many aspects of the U.S. economy. However, the amount of goods transported, and the modes by which they are transported contribute to local and regional air quality problems. The majority of freight (73% 1) in the U.S. is transported by freight trucks, which are classified as heavy duty diesel vehicles (HDDVs), and are a substantial source of smogproducing nitrogen oxides (NOx) and primary particulates, especially elemental carbon (EC). The important role of trucking emissions has been addressed in recent emission control policies from the U.S. Environmental Protection Agency (EPA), especially the mandate for ultralow sulfur diesel (ULSD), which allows for after-treatment technologies to significantly reduce particulate and NOx emissions from both trucks and trains over the coming decades. However, regulating emissions does not address underlying activity choices, which will be increasingly important as projected domestic freight tonnage doubles by 2050.1 Relatively fuel-efficient rail transport offers an existing technology to potentially address both air quality © 2013 American Chemical Society
Received: Revised: Accepted: Published: 446
April 15, 2013 August 26, 2013 September 4, 2013 September 4, 2013 dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
Article
Table 1. Comparison between Truck-to-Rail Modal Shift Studies in the Literature and This Study, and Associated Emissions Changes for NOx and PMa ΔNOx
ΔPM
Park et al, 2007
−1.7%
Comer et al, 2010
+58% to +283% +12% to +30% (PM10) −59%
1
You et al, 2010
−2% to −3%
−0.1% to −1% (PM2.5)
1
Lee, 2011
−8% to −25%
−8% to −28% (PM10)
1
Bickford et al, 2013 −26%
−2.2% (PM10)
−13% (PM2.5) −16% (PM10)
air quality modeling
ΔCO2 O−D Pairs 1
−31%
study mode-shift summary
no
1,238
20% of port traffic from ports of L.A. and Long Beach traveling 20 miles to downtown L.A. is shifted from truck to rail. no 100% of containerized truck freight shifted to rail on 550 mile corridor between Montreal, QC to Cleveland, OH with emissions factors for a year 2007 truck, and year 2004−2006 truck. no 3% of freight trucks traveling from ports of L.A. and Long Beach to downtown L.A. replaced by trains. Calpuff (no 6% (2005) to 38% (2030) of truck volume at the Ports of L.A. chemistry) and Long Beach are replaced by on-dock rail operations. CMAQ 40% of long haul (>400 miles) freight truck VMT traveling through the upper Midwestern U.S. is shifted to rail.
a
All modal shift analyses in the literature conducted localized studies of a single corridor and none included air quality modeling with a regional photochemical model. Lee (2011) does some near road air quality analysis with a dispersion model (Calpuff); however, it does not include chemical formation of secondary pollutants like O3 or secondary PM2.5.
model, which provides time-varying ambient concentrations of major pollutants and chemical precursors regulated by the EPA. These data can be used directly to evaluate air management strategies, including State Implementation Plan (SIP) development to ensure NAAQS compliance. Air pollution model results also provide input to health and economic analyses, such as thorough analyses using the EPA Benefits Mapping (BenMap) tool.
Several studies have proposed shifting away from trucks toward more fuel-efficient, nonhighway modes.3−10 In an economic analysis, Gorman3 concluded that 25% of freight could be shifted to rail at a lower cost if the infrastructure existed, thus yielding an 80% reduction in social costs of emissions (without air quality modeling), congestion, and safety. Bryan et al.4 found freight modal shift from truck to rail could significantly reduce roadway congestion and, citing public benefit, advocated for public investment in private rail infrastructure. In a report for the Federal Railroad Administration, ICF International5 analyzed truck and rail movements on competitive corridors in the U.S., determining that rail was more fuel efficient on all 23 corridors. Table 1 summarizes other studies that have considered truckto-rail modal shift and quantified net emissions changes for NOx and PM in a comparable manner.6−9 Most studies report that shifting freight from truck to rail reduces net freight emissions. However, results vary widely due to differing scopes, study years, and assumptions about vehicle speeds, activity, pollutant control technologies, and emissions factors. All existing studies in the literature restrict their analysis to modal shift impacts on a single corridor; three of which focus on the same 20-mile corridor in Los Angeles, California. Figure 1 outlines our study structure in a manner comparable to Park et al., Figure 2,6 and You et al., Figure 2,8 with each step discussed in detail in the Supporting Information, section S1. Only Lee9 addresses air quality through the use of a dispersion model, a methodology that does not include chemical formation or destruction of pollutants in the atmosphere, nor considers region-wide air impacts. This study is the first to evaluate truck-to-rail modal shifts over a wide area of nearly a million square miles (vs a single corridor), and the first to evaluate modal shift impacts on ambient air quality, including chemistry and complex meteorology. As such, our analysis directly builds on commodity flow and activity data in a manner relevant to state air quality management and federal policies affecting mode choice. Using an original, commodity-based emissions inventory, the Wisconsin Inventory of Freight Emissions (WIFE), we quantify modal shift scenarios based on commodities, freight corridors, and speed classifications. Truck and rail emissions scenarios are combined with other anthropogenic and natural emissions sources as input into the most advanced air quality model from U.S. EPA, the Community Multiscale Air Quality (CMAQ)
■
DATA AND METHODS The overall structure of our study design is presented in Figure 1, with each component discussed in detail in Supporting Information, section S1. We have developed a policy-relevant emissions inventory (WIFE version 1.5; earlier version discussed in Johnston et al., 201211) to allow for multicorridor, multicommodity, multipollutant emissions inventory calculations. WIFE incorporates HDDV freight activity from the Federal Highway Administration’s (FHWA) Freight Analysis Framework (FAF) version 2.21,12 database and HDDV emissions factors from the EPA’s MOBILE6.2 emissions model.1,13 Freight rail emissions were calculated on the basis of published rail emissions factors (see Table S3 and section S3 in the Supporting Information). Emissions are allocated to the Lake Michigan Air Directors Consortium’s (LADCO) 12 km × 12 km upper Midwest air quality model grid and speciated to chemical constituents appropriate for the Carbon Bond 5 (CB05) mechanism in CMAQ.3,14,15 Gridded and speciated emissions from WIFE and LADCO are input to CMAQ version 4.6, run with 27 vertical layers and horizontal resolution of 12 km × 12 km, along with threedimensional, time-varying meteorology calculated by the Weather Research and Forecasting (WRF) model (see section S1, i−l for details on CMAQ model and inputs; Figure S5 in the Supporting Information for an illustration of the model domains). To capture seasonal effects on both vehicle emissions rates and atmospheric chemistry, all scenarios were modeled for the full months of both January and July 2005, along with a 10day spin-up period. The reliability of results depends on the uncertainties in data inputs and models. It is standard practice to evaluate an air quality modeling system against surface air quality measurements. Although this approach suffers from the potential for compensating errors, it offers a more practical and representative estimate of the combined error introduced by emission inputs (here, WIFE and LADCO), meteorological 447
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
Article
Figure 1. Flowchart outlining study structure, in a style adapted from Park et al., Figure 2,6 and You et al., Figure 2.8 Each step is discussed in detail in the Supporting Information, section S1.
standard truck emissions with WIFE. This baseline ensures that differences in emissions calculation methodology do not affect study results. The Midwest intraregional (I-R) scenario, considers only freight on rail-competitive corridors originating and terminating within the Midwest study region. The Midwest through-freight (T-F) scenario incorporates the I-R scenario, as well as all other freight passing through the Midwest region. Details on this process are provided in Supporting Information, section S1(a−f). Commodities selected for modal shift were limited to goods suitable for transport by either truck or rail, determined by a comparison of the commodities and the modes that transport them in the FAF database, and a literature review of commodity mode-choice for truck−rail competitive corridors.4,5,17−20 The most significant commodities (by tonnage) with potential to shift to rail in both scenarios were base metals, other foodstuffs, nonmetal mineral products, and motorized vehicles (see section S1.d and Table S1 in the Supporting Information). For this analysis, we set 400 miles as the minimum distance to consider modal shift to rail to be economically feasible for both shippers and rail carriers, and in the range of prior studies,5−9,17,19 and found 28 origin−destination (O−D) city pairs. (see section S1.c and Table S2 in the Supporting Information).
inputs (here, WRF), and the parameters and structural assumptions of the air quality model (here, CMAQ). This analysis is discussed in Supporting Information, section S4, and compared with the literature in Table S4. Average errors for NO2 and EC are about 50%, with low mean biases (−3% and +11%, respectively), which represents performance in line or better than comparable prior studies, referenced in Table S4. Fundamental to our estimates of modal shift is the data uncertainties in the FAF database. However, a literature review of studies employing FAF or documents on FAF development found no discussion of associated uncertainties. A 2011 report from the U.S. Government Accountability Office that focused explicitly on uncertainty characterization used FAF as an input to a costbased decision model to evaluate freight modal choices but noted that data limitations preclude assigning confidence of results.16 We note that the current study omits changes to emissions from “nodes” of a freight transport system (e.g., rail yards, intermodal facilities, etc.), which can be a major source in populated urban areas. Three scenarios were modeled for this analysis: (1) a baseline scenario, (2) the Midwest intraregional (I-R) scenario, and (3) the Midwest through-freight (T-F) scenario. The baseline scenario assumes no modal shift but replaces LADCO 448
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
Article
HDDV VMT per day. This scenario amounts to a 40% reduction of Midwest HDDV VMT. For freight rail, the T-F scenario adds 34 854 railcars per day, or 745 231 120 ton-miles per day to the Midwest freight rail network, doubling Midwest freight rail tonnage (see section S2 in Supporting Information for more details on the T-F scenario design). Although theoretically, these estimates represent unrealized cost savings to shippers, there are a number of variables that influence modal choice. These include rail capacity, transport time, cost (which here we only approximate as benefitting rail for trips >400 miles), commodity characteristics (which here we estimate based on generalized categories), shipment size, production process, size of shipper and receiver, operating hours, and other factors.21 Modal Shift Emissions Change. The regional average, net changes in emissions for each scenario are shown in Table 3, given in percent changes relative to baseline (B-L) trucking emissions. Regional average emission reductions for the T-F scenario range from 3% for EC to 40% for particulate sulfate (SO42−). In addition to health-relevant air pollutant reductions, emissions of carbon dioxide (CO2) are reduced 31% in the T-F scenario. Emission changes are much smaller in the I-R scenario, consistent with smaller change in VMT. In both scenarios, sulfur dioxide (SO2) emissions increase, up to 12% in the T-F scenario. In 2005, fuel standards for highway diesel contained 500 ppm sulfur,9,22 while railroad diesel contained 2600 ppm,11,23 so increased freight rail yields increased SO2 emissions. New EPA regulations are stepping down both highway and rail diesel sulfur levels to 15 ppm by 2015. To evaluate how this new rail fuel standard will affect results, we repeated the emissions analysis of the T-F scenario using EPA’s Motor Vehicle Emissions Simulator (MOVES) model for 2035 conditions (in lieu of MOBILE6.2) and assumed a range of technology advancements and reduced sulfur content for both rail and trucks. With these assumptions about future technology, SO2 emissions decrease by 33% under the T-F scenario, with comparable NOx reductions as calculated for 2005 with MOBILE6.2. Within the regional net changes reported in Table 3, there is considerable local heterogeneity. Figure 2 shows the spatial pattern of highways (blue) and railways (red) considered for modal shift in each scenario with net emissions changes for July nitrogen dioxide (NO2) and primary PM2.5. Emissions reductions occur on highways (negative values shown in blue), while emissions increases occur on railways (positive values shown in yellow and red).
The Midwest T-F scenario includes all routes considered in the I-R scenario, as well as freight traveling into, out of, and passing through the Midwest region. For both scenarios, we made a number of assumptions. Through the 400-mile minimum distance threshold, we assume modal shifts to be economically feasible, but not necessarily logistically feasible. Our analysis aims to evaluate the potential benefits of long-term infrastructure investments, so the assumption of adequate rail capacity is part of the study design. Other scenario design assumptions are discussed in section S2 of the Supporting Information.
■
RESULTS Modal Shift Potential. Calculated changes in daily trucks/ railcars and ton-miles shipped, as well as truck VMT, are reported in Table 2. For the Midwest I-R scenario, truck routes Table 2. Summary of Activity Changes Associated with Freight Modal Shift Scenarios for the Midwest Region ΔI-R scenario
ΔT-F scenario
−12,174 −2,534 −1,337,602 +876 +18,182,724
−526,789 −103,450 −52,744,923 +34,854 +745,231,120
truck kilotons per year trucks per day truck VMT per day railcars per day rail ton-miles per day
longer than 400 miles transported 12 602 KT of freight, 4% of the intraregional tonnage in the study domain. Of that, 12 174 KT (97%) were commodities eligible for mode shift, which corresponds to 2534 removed HDDVs per day, or 1 337 602 removed HDDV vehicle miles traveled (VMT) per day. This scenario, while significant for some routes, only amounts to a 1% reduction in total daily Midwest HDDV VMT, because most (80% by tonnage) intraregional freight movements did not meet the 400-mile minimum. The scenario translates to adding 876 railcars per day, or 18 182 724 ton-miles to Midwest rail routes per day, a 5% increase in total Midwest rail freight tonnage (see section S2 in Supporting Information for more details on the I-R scenario design). For the Midwest T-F scenario, we found that 526 789 KT per year could be shifted from truck to rail, 15% of all freight tonnage moving in the Midwest study region. This removes 103 450 trucks per day from Midwest highways, and 52 744 923
Table 3. Midwest Modal Shift Emissions Change by Scenario and Season. Percentage Changes Are Relative to Baseline (B-L) Trucking July B-L
January
I-R
T-F
B-L
I-R
T-F
pollutant
abbrev.
tons/day
Δtons/ day
%
Δtons/day
%
tons/day
Δtons/ day
%
Δtons/day
%
carbon monoxide carbon dioxide ammonia nitrogen oxides particulate elemental carbon particulate matter, coarse particulate matter, fine particulate organic carbon particulate sulfate sulfur dioxide volatile organic compounds
CO CO2 NH3 NOx PEC PMC PMFINE POC PSO4 SO2 VOC
540.46 202 992.73 3.84 2441.03 25.49 7.12 1.69 12.16 4.41 63.05 67.46
−4.62 −1610.75 −0.03 −17.01 −0.03 −0.06 0 −0.07 −0.04 0.16 −0.23
−0.86 −0.79 −0.89 −0.7 −0.12 −0.9 −0.16 −0.6 −1.02 0.25 −0.34
−168.09 −62 752.75 −1.34 −631.81 −0.86 −2.52 −0.08 −2.79 −1.77 7.39 −7.52
−31.1 −30.91 −34.85 −25.88 −3.39 −35.46 −4.97 −22.96 −40.02 11.72 −11.15
547.14 202 992.73 3.84 2178.89 26.41 7.22 1.69 12.57 4.41 63.09 70.61
−4.71 −1 610.75 −0.03 −14.25 −0.04 −0.07 0 −0.08 −0.04 0.16 −0.27
−0.86 −0.79 −0.89 −0.65 −0.15 −0.91 −0.16 −0.61 −1.02 0.25 −0.38
−168.02 −62 752.72 −1.34 −305 −1.24 −2.57 −0.08 −2.96 −1.77 7.37 −9.4
−30.71 −30.91 −34.85 −14 −4.7 −35.53 −4.97 −23.55 −40.02 11.69 −13.31
449
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
Article
Figure 2. Modal shift highway (truck) and railway routes for (a) the intraregional (I-R) scenario and (b) the through-freight (T-F) scenario, with accompanying daily July emissions changes in primary NO2 (b,c) and primary PM2.5 (e,f).
Modal Shift Air Quality Change. Maps depicting percent and absolute pollutant changes for July and January for the T-F scenario are shown in Figure 3 (for surface NO2) and Figure 4 (for surface EC). The T-F scenario yielded significant reductions in ambient NO2 in roadway grid cells of −28% (−2.33 ppbV) in July and −18% (−1.58 ppbV) in January. Added rail freight activity increased ambient NO2 along rail lines +25% (+0.83 ppbV) in July, and +21% (+0.80 ppbV) in January. Because NO2 concentrations in railway areas are lower than highway areas, the absolute concentration change (in ppbV) in railway gridcells is
smaller than the percentage change. The balance of reduced highway emissions and increased rail emissions yields a net regional monthly average change in ambient surface of NO2 of −3.7% in July, and −1.7% in January. Figure 4 shows changes in monthly mean surface EC concentrations for July and January. We see roadway-area concentrations of EC changed −16% (−0.05 μg/m3) in July, and −15% (−0.05 μg/m3) in January. Added rail activity increased railway EC emissions +21% (+0.05 μg/m3) in July and +29% (+0.05 μg/m3) in January. As a result, the region-wide average 450
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
Article
Figure 3. Through-Freight scenario CMAQ modeled (a,b) percent and (c,d) absolute change in July and January 2005 surface nitrogen dioxide (NO2) concentration as a result of shifting freight from truck to rail.
region per year, could be economically shifted off of truck and onto train, if adequate rail infrastructure existed, and policy incentives were structured to favor freight rail selection by shippers. Although a range of transportation policies could be structured to promote modal shifts from truck to rail,2 such policy development would depend on the value of benefits projected to society. Because diesel trucks contribute over 25% to NOx emissions in many urban areas, and impact near-road PM exposure, the air quality and human health impacts of freight modal shift are likely to be the most valuable of modal shift social benefits. Here, we quantify the expected air quality benefits of realistic modal shift scenarios, over a wide geographic region central to U.S. freight transport. The most significant emissions reductions of truck-to-rail modal shift in the Upper Midwestern U.S. are for CO2 under the T-F scenario, which results in a 31% emissions reduction relative to the baseline. These CO2 emissions reductions represent 23 million tons per year or the equivalent of taking four million passenger cars off the road. Comer et al.7 found a
change shows a small net increase in regional monthly mean EC emissions of +0.2% for both July and January. These findings for EC bear relevance for air quality as well as climate, as black carbon comprises 61%−77% of diesel exhaust PM2.5 emissions by mass,13,15 and plays an important role in scattering and absorbing incoming solar radiation.24−26 Modal shift air quality impacts on monthly mean, total PM2.5 (both primary and secondary) and monthly mean, 8-h maximum O3 are shown in Supporting Information, Figures S1 and S2. T-F changes in PM2.5 range from −3% (−0.34 μg/m3) near roadways in July to +1% (+.08 μg/m3) near railways in July. The net change in monthly mean 24-h ambient PM2.5 across the region was −0.7% in July and −0.1% in January. July O3 concentrations show a similar range of percentage change, and region-wide change of −0.6% in monthly mean 8-h maximum O3.
■
DISCUSSION We found 12 million tons of intraregional freight, and 530 million tons of all truck freight moving through the upper Midwest 451
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
Article
Figure 4. Through-Freight scenario CMAQ modeled (a,b) percent and (c,d) absolute change in July and January 2005 surface elemental carbon (EC) concentration as a result of shifting freight from truck to rail.
Estimated NOx reductions under the T-F scenario would contribute to compliance with the NAAQS for NO2, O3, and PM2.5. Regional air quality modeling quantified the impact of these emissions on ground-level pollutant concentrations. The T-F scenario showed reductions up to −28% (−2.33 ppbV) for NO2 in near-roadway gridcells and increases up to +25% (+0.83 ppbV) in near-railway gridcells. Results for NO2 bear relevance for recent EPA regulations setting new 1-h maximum concentration limits for NO2, although near-roadway and near-railway monitors would be expected to show a larger response to NO2 emission reductions than those simulated in on the 12 km × 12 km model grid. The new NO2 standard is likely to increase scrutiny of major NOx emitters, especially in highly populated areas (see Figure S4 in the Supporting Information for maps comparing freight activity corridors and Midwest population centers). The T-F NOx reductions also reduced ambient levels of both O3 and PM2.5, with changes of up to −3% (−1.81 ppbV and −0.34 μg/m3, respectively). While these changes are not as
59% reduction in CO2 emissions associated with a 100% removal of truck VMT on a long-haul (550 mile) route, which agrees well with our estimate based on a 40% VMT removal. As the U.S. considers cost-effective measures to reduce carbon emissions, policies supporting freight modal shift would be a relatively low-cost, and potentially even no-cost, option. The T-F scenario also yields a −26% change in freight-related NOx emissions. In contrast to our findings, Comer et al.7 find that NOx (and PM) emissions increase as a result of truck-to-rail modal shift. The difference is due to assumptions about emissions factors: whereas MOBILE6.2 (used here) accounts for fleet age, turnover rates, operating conditions, activity, efficiency, and pollutant control technologies, the Comer et al. study used maximum allowable emissions standards for a new 2007 or later freight truck (adhering to EPA’s stricter 2007 HDDV emissions standards), and maximum allowable emissions standards for 2005 Tier-2 locomotives (predating EPA’s stricter 2008 locomotive standards). This comparison highlights the sensitivity of criteria pollutant emissions to assumptions about fleet engine characteristics. 452
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
■
great as those seen for primary pollutants like NO2 and EC, even low percentage changes can make a difference between counties passing and failing the NAAQS limits. For comparison, the EPA applies a threshold of 1% (of the NAAQS) in the design of the Cross State Air Pollution Rule (CSAPR), to determine whether a state is significantly contributing to a downwind state, suggesting thatfrom the perspective of the EPA impacts in excess of 1% represent a meaningful contribution. It is worth noting that a 3% (or 1%) change is well below even a qualitative estimate of reliability associated with our modeling system (or that of the EPA in designing CSPAR). Despite this model uncertainty relative to system change, this type of modeling result is routinely used for air quality management. Overall, our modeling system performs as well or better than past air quality studies, as reported in Supporting Information, section S4. Freight transport in the U.S. is increasing, just as health and environmental interests push for reductions in CO2, NOx, PM, and other pollutants. Thus, both transportation and air quality policy communities have a vested interest in reducing the environmental impacts of freight transport at low costs. Calculating the costs and air quality impacts of large-scale modal shifts will always be uncertain, given limitations in data and tools, as well as a changing technological and economic environment. There remains untapped potential to integrate well-developed data sources from the transportation research community with advanced modeling tools for atmospheric analysis to better support multidimensional transportation decision-making.
■
REFERENCES
(1) Transportation Reboot: Restarting America’s Most Essential Operating SystemUnlocking Freight; UGFR-1; American Association of State Highway and Transportation Officials (AASHTO): Washington, DC, 2010; http://nfl.transportation.org/Documents/ UGFR-1-OL.pdf. (2) Brogan, J. J.; Aeppli, A. E.; Beagan, D. F.; Brown, A.; Fischer, M. J.; Grenzeback, L. R.; McKenzie, E.; Vimmerstedt, L.; Vyas, A. D.; Witzke, E. Transportation Energy Futures Series: Freight Transportation Modal Shares: Scenarios for a Low-Carbon Future; Report No. DOE/ GO-102013-3705; National Renewable Energy Laboratory: Golden, CO, 2013; http://www.nrel.gov/docs/fy13osti/55636.pdf. (3) Gorman, M. F. Evaluating the public investment mix in US freight transportation infrastructure. Transport. Res. Part A 2008, 42, 1−14. (4) Bryan, J.; Weisbrod, G.; Martland, C. D. Rail freight as a means of reducing roadway congestionFeasibility considerations for transportation planning. Transp. Res. Rec. 2007, 75−83. (5) Comparative Evaluation of Rail and Truck Fuel Efficiency on Competitive Corridors; ICF International: Fairfax, VA, 2009; http://ntl. bts.gov/lib/31000/31800/31897/Comparative_Evaluation_Rail_ Truck_Fuel_Efficiency.pdf. (6) Park, M.; Regan, A.; Yang, C. H. Emissions impacts of a modal shift: a case study of the Southern California ports region. J. Int. Logistics Trade 2007, 5 (2), 67−81. (7) Comer, B.; Corbett, J. J.; Hawker, J. S.; Korfmacher, K.; Lee, E. E.; Prokop, C.; Winebrake, J. J. Marine vessels as substitutes for heavyduty trucks in great lakes freight transportation. J. Air Waste Manage. Assoc. 2010, 60, 884−890. (8) You, S.; Lee, G.; Ritchie, S.; Saphores, J. Air Pollution Impacts of Shif ting San Pedro Bay Ports Freight f rom Truck to Rail in Southern California; UCTC Research Paper No. UCTC-2010-07 (renumbered as UCTC-FR-2010-07); University of California Transportation Center: Irvine, CA, 2010. (9) Lee, G. Integrated Modeling of Air Quality and Health Impacts of a Freight Transportation Corridor. Ph.D. Thesis, University of California, Irvine, CA, 2011. (10) Farzaneh, M.; Lee, J. S.; Villa, J.; and Zietsman, J. Corridor-Level Air Quality Analysis of Freight Movement: North American Case Study. Transp. Res. Rec.: J. Transport. Res. Board, 2011, 19−26 DOI: 10.3141/2233-03. (11) Johnston, M.; Bickford, E.; Holloway, T.; Dresser, C.; Adams, T. M. Impacts of biodiesel blending on freight emissions in the Midwestern United States. Transp. Res. Part D 2012, 1−11. (12) Freight Analysis Framework (FAF) User Guide, version 2.2; United States Department of Transportation−Federal Highway Administration: Washington, DC, 2006; http://www.ops.fhwa.dot. gov/freight/freight_analysis/faf/faf2userguide/faf2userguide.pdf. (13) User’s Guide to MOBILE6.1 and MOBILE6.2; Report EPA420-R03−010; United States Environmental Protection Agency: Ann Arbor, MI, 2003; http://www.epa.gov/otaq/models/mobile6/420r03010.pdf. (14) Updates to the Carbon Bond Chemical Mechanism: CB05; Final Report No. RT-04−0067;ENVIRON International Corp.: Novato, CA, 2005; http://www.camx.com/publ/pdfs/cb05_final_report_ 120805.pdf. (15) Sarwar, G.; Luecken, D.; Yarwood, G.; Whitten, G. Z.; Carter, W. P. L. Impact of an updated carbon bond mechanism on predictions from the CMAQ modeling system: Preliminary assessment. J. Appl. Meteor. Climatol. 2008, 47, 3−14. (16) Intercity Passenger and Freight Rail: Better Data and Communication of Uncertainties Can Help Decision Makers Understand Benef its and Trade-of fs of Programs and Policies; Report GAO-11-290; United States Government Accountability Office: Washington, DC, 2011; http://www.gao.gov/assets/320/315893.pdf. (17) Analysis of Freight Movement Mode Choice Factors; Award No. BD238; The Center for Urban Transportation Research at the University of South Florida: 2003;http://www.dot.state.fl.us/rail/ Publications/Studies/Planning/ModeChoiceFactors.pdf.
ASSOCIATED CONTENT
S Supporting Information *
Description in more detail of the overall study methodology, the design of the modal shift scenarios, with tables for commodities considered for both scenarios, and the city pairs in the I-R scenario; freight truck and rail emissions factors are presented in common units of grams per ton-mile, and CMAQ model validation methods are described with statistical results compared to literature values; additional maps of modal shift air quality concentration changes for PM2.5, O3, and SO2, maps of highway and railway activity and population density for the Midwest study domain to illustrate how modal shift may impact air quality near population centers, and a map illustrating the model domains. This material is available free of charge via the Internet at http://pubs.acs.org.
■
Article
AUTHOR INFORMATION
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was funded by CFIRE and Wisconsin DOT, with data and support provided by the Lake Michigan Air Directors Consortium. The authors would like to thank Prof. Scott Spak (University of Iowa) for WRF and CMAQ model setup assistance and CFIRE staff for assistance and discussion. We also express deep gratitude to our CFIRE advisory committee for lending their freight and air quality expertise to this project. Finally, we appreciate the time and comments of anonymous reviewers, whose input has benefited this manuscript. 453
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
Environmental Science & Technology
Article
(18) Climate and Transportation Solutions: Findings f rom the 2009 Asilomar Conference on Transportation and Energy Policy; Sperling, D., Cannon, J. S., Eds.; Institute of Transportation Studies: University of California, Davis, CA, 2012. (19) Multimodal Freight Forecasts for Wisconsin; Bureau of Planning, Division of Transportation Investment Management, Wisconsin Dept. of Transportation: Madison, WI, 1996. (20) Quick Response Freight Manual II; Publication No. FHWA-HOP08-010; Cambridge Systematics: Cambridge, MA, 2007; http://ops. fhwa.dot.gov/freight/publications/qrfm2/qrfm.pdf. (21) Freight-Demand Modeling to Support Public-Sector Decision Making; NCFRP Report 8; Transportation Research Board: Washington, DC, 2010; http://onlinepubs.trb.org/onlinepubs/ncfrp/ ncfrp_rpt_008.pdf (22) Technical Guidance on the Use of MOBILE6.2 for Emission Inventory Preparation; United States Environmental Protection Agency Office of Transportation and Air Quality: Ann Arber, MI, 2004; http://www.epa.gov/otaq/models/mobile6/420r04013.pdf. (23) LADCO 2005 Locomotive Emissions; Draft Report for the Lake Michigan Air Director’s Consortium; Environ: Novato, CA, 2007;http://www.ladco.org/reports/technical_support_document/ references/ladco_2005_locomotive_emissions.021406.pdf. (24) Highwood, E.; Kinnersley, R. When smoke gets in our eyes: The multiple impacts of atmospheric black carbon on climate, air quality and health. Environ. Int. 2006, 32, 560−566. (25) Unger, N.; Shindell, D. T.; Wang, J. S. Climate forcing by the on-road transportation and power generation sectors. Atmos. Environ. 2009, 43, 3077−3085. (26) Jansen, K. L.; Larson, T. V.; Koenig, J. Q.; Mar, T. F.; Fields, C.; Stewart, J.; Lippmann, M. Associations between health effects and particulate matter and black carbon in subjects with respiratory disease. Environ. Health Persp. 2005, 113, 1741−1746.
454
dx.doi.org/10.1021/es4016102 | Environ. Sci. Technol. 2014, 48, 446−454
