Regional On-Road Vehicle Running Emissions Modeling and

Sep 22, 2009 - For an urban case study, passenger cars were found to be the largest sources of HC, CO, and CO2 emissions, whereas trucks contributed t...
2 downloads 10 Views 525KB Size
Environ. Sci. Technol. 2009, 43, 8449–8455

Regional On-Road Vehicle Running Emissions Modeling and Evaluation for Conventional and Alternative Vehicle Technologies H . C H R I S T O P H E R F R E Y , * ,† H A I B O Z H A I , ‡ AND NAGUI M. ROUPHAIL§ Department of Civil, Construction and Environmental Engineering, North Carolina State University, Campus Box 7908, Raleigh, North Carolina 27695-7908, Department of Engineering and Public Policy, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, and Institute for Transportation Research and Education, North Carolina State University, Campus Box 8601, Raleigh, North Carolina 27695-8601

Received February 18, 2009. Revised manuscript received August 24, 2009. Accepted September 1, 2009.

This study presents a methodology for estimating highresolution, regional on-road vehicle emissions and the associated reductions in air pollutant emissions from vehicles that utilize alternative fuels or propulsion technologies. The fuels considered are gasoline, diesel, ethanol, biodiesel, compressed natural gas, hydrogen, and electricity. The technologies considered are internal combustion or compression engines, hybrids, fuel cell, and electric. Road link-based emission models are developed using modal fuel use and emission rates applied to facilityand speed-specific driving cycles. For an urban case study, passenger cars were found to be the largest sources of HC, CO, and CO2 emissions, whereas trucks contributed the largest share of NOx emissions. When alternative fuel and propulsion technologies were introduced in the fleet at a modest market penetration level of 27%, their emission reductions were found to be 3-14%. Emissions for all pollutants generally decreased with an increase in the market share of alternative vehicle technologies. Turnover of the light duty fleet to newer Tier 2 vehicles reduced emissions of HC, CO, and NOx substantially. However, modest improvements in fuel economy may be offset by VMT growth and reductions in overall average speed.

Introduction Highway vehicles account for an estimated 54% of carbon monoxide (CO), 36% of nitrogen oxides (NOx), and 22% of volatile organic compounds (VOC) emissions in the U.S. (1). The U.S. Environmental Protection Agency’s (EPA) emission factor model, MOBILE6, calculates average fleet emission factors for hydrocarbons (HC), carbon monoxide (CO), nitrogen oxides (NOx), and carbon dioxide (CO2) (2). MOBILE6 estimates are based on dynamometer tests from standardized * Corresponding author phone: 1-919-5151155; fax: 1-919-5157908; e-mail: [email protected]. † Department of Civil, Construction and Environmental Engineering, North Carolina State University. ‡ Carnegie Mellon University. § Institute for Transportation Research and Education, North Carolina State University. 10.1021/es900535s CCC: $40.75

Published on Web 09/22/2009

 2009 American Chemical Society

driving cycles (3). Multiple cycles represent different levels of service (LOS) and roadway type (4-8). Although MOBILE6 can estimate average emission rates for some combinations of facility type and LOS, it has some limitations. For example, MOBILE6 does not distinguish between emission rates for vehicles at on- and off-ramps (2, 4), which are significantly different because of the predominance of acceleration events at on-ramps versus decelerations at off-ramps (8). In addition, emission factors for CO2 are not adjusted by speed and facility type. Travel demand models (TDMs) typically produce linkbased vehicle activity data (9), whereas emission factors models such as MOBILE6 are cycle-based. Link-based emissions estimates are therefore required to enable compatibility with data generated from TDMs. Fuel-based emission inventories have been developed using location-specific emission factors from remote sensing measurements (10). New measurement and analysis techniques quantify actual emissions in time and space (3). For example, portable emissions monitoring systems (PEMS) quantify vehicle activities and emissions for representative conditions along travel paths at a road link level for links and routes (7, 8, 11, 12). Vehicle emissions can be reduced through the use of alternative fuels and new propulsion technologies. For example, on a life-cycle basis, a blend of gasoline with 85% ethanol (E85) can reduce greenhouse gas (GHG) emissions by 15-20% compared to gasoline (13). Compressed natural gas (CNG) reduces CO2 emissions up to 30% (14). Hybrid gasoline-electric vehicles (HEVs) reduce life-cycle GHG emissions by 30-50% (15). E85 flex-fuel, HEV, advanced diesel, electric and fuel cell vehicles may comprise 27-63% of new LDV sales by 2030 (16, 17). The principal objectives of this study are to (a) estimate and characterize regional on-road vehicle emission factors and inventories; and (b) quantify the impacts of market penetration for alternative vehicle technologies on regional vehicle emissions.

Materials and Methods In order to develop TDM-compatible emissions estimates, microscale emissions factors were developed for various levels of service (LOS) at the roadway link level. PEMS data were available for light-duty gasoline, diesel, E85, CNG, and hybrid vehicles; heavy-duty diesel and biodiesel trucks; and diesel buses. Where PEMS data were sparse, other data sources such as vehicle certification tests were used to characterize vehicle emissions. The methodology includes (a) development of link-based tailpipe emissions models; (b) coupling of tailpipe emissions models and link-based vehicle activity data for estimating an emissions inventory; and (c) multiple scenarios to quantify the impacts of alternative technologies on regional tailpipe emissions. Link-Based Emission Factors. Vehicle emission factors are affected by fuel type, engine technology, link average speed, facility type (e.g., freeway, arterial), emission control standards, inspection and maintenance (I/M), ambient conditions (temperature, humidity, and barometric pressure), and vehicle class and age. Vehicle classes included are lightduty vehicles (LDVs), heavy-duty vehicles (HDVs), and buses. Link-based tailpipe emission factors for each technology class are estimated from basic emission rates (BERs) obtained using the MOBILE6 model, and correction factors. Thus, VOL. 43, NO. 21, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8449

EFY,T,f,V ) BERY,T ′,f,v × TECF × HCF × PCF × CCFT ′ × TCFT ′,T × SCFT,f,V (1) Where BERY,T ’,f,v ) basic emission rate (g/sec) for calendar year Y, conventional technology T ’, facility type f, and average cycle speed v under standard temperature, pressure and humidity; CCFT ’ ) cycle correction factor to convert BER at FTP cycle speed to real world link-based emissions at the same average cycle speed for conventional LDV technology; EFY,T,f,V ) real-world link-based emission factor (g/sec) for calendar year Y, technology class T, facility type f, and linkbased average speed V; HCF ) local relative humidity correction factor (dimensionless); PCF ) local pressure correction factor (dimensionless); SCFT,f,V ) speed- and facility-specific correction factor for a given technology (dimensionless); TCFT ’,T ) correction factor for technology alternative (T) versus conventional technology T ’ (dimensionless); TECF ) temperature correction factor (dimensionless); f ) facility type (freeway, arterial, ramp, local and collector); T ) vehicle technology index class (conventional and advanced alternative technologies); T ’ ) conventional technology index (LDGVs, HDDVs, and diesel buses); v ) standard driving cycle speed (19.6 mph for LDGV and 20.0 mph for HDDV and bus); V ) average speed on roadway link (mph); Y ) calendar year (CY2005, CY2030). MOBILE6’s BERs for LDVs are based on chassis dynamometer tests using the LA4 driving cycle, which is part of federal test procedure (FTP). LA4 has an average speed of 19.6 mph (4). BERs for heavy-duty diesel vehicles (HDDVs) and buses are based on engine dynamometer tests. BERs for HC, CO, and NOx are estimated for light-duty gasoline vehicles (LDGVs) and diesel vehicles (LDDVs), and HDDVs and buses. BERs for direct CO2 emissions for LDVs are estimated based on fuel economy data from the FTP-75 and highway fuel economy test (HFET) driving cycles, whereas BERs for HDDVs and buses are from MOBILE6. BERs are adjusted for ambient conditions using MOBILE6 correction factors. Cycle correction factors (CCFs) are estimated by comparing average emission rates for real-world link-based cycles to those for the baseline LA4 cycle, taking into account differences in speed and acceleration profiles using modal tailpipe emissions models. The modes are defined based on ranges of vehicle specific power (VSP), which is an estimate of engine power demand taking into account aerodynamic drag, tire rolling resistance and road grade (18). Technology correction factors (TCFs) for LDVs, HDVs, and buses account for differences in emissions between

alternative and conventional technologies. TCFs are estimated from comparisons of vehicle certification tests where possible. The speed correction factor (SCF) is the ratio of link average emission rate for a particular speed and facility combination to that at the baseline speed for an arterial. SCFs are derived from linked-based real-world speed profiles for conventional and alternative vehicle technologies, in order to more realistically reflect the effect of driving patterns. Methods for quantifying uncertainty in emission factors as calculated in eq 1 have been demonstrated by Frey and Zheng (19). The absolute uncertainty associated with base emission rates can be approximately (20 to (50%, typically, depending on the sample size and intervehicle variability (19). However, the uncertainty in differences between scenarios is less than absolute differences, since multiple scenarios share common sources of uncertainty, including driving cycles, humidity, ambient pressure, and temperature. Emission Inventory (EI) Estimation. Travel demand models typically generate outputs such as link volume, origin and destination demands, facility type, vehicle class distribution, free flow and actual/congested speeds, and traffic volume and VMT by vehicle class. Such data were obtained from running the Triangle Regional Model (TRM) in North Carolina as an illustrative case study (20). Link-based emissions models are coupled with link-level activity data to estimate emission inventories, as shown in Figure 1. Total emissions for a single link for a given facility type are estimated from eqs 2 and 3 below: tiT )

TEi,f )

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 21, 2009

∑ EF

× 3600

Y,T,f,Vi

· tiT · voliT

(2)

(3)

T

Where i ) link ID; VTi ) mean speed of vehicles in technology class T traveling on link i (mph); Li ) length of link i (miles); tTi ) average travel time for vehicle technology class T on link i(sec/veh); volTi ) vehicle volume in technology class T on link i (veh/h); TEi,f ) total emissions on link i for facility class f(grams). Summing TE across all links on a network produces a high resolution emission inventory estimate. Regional Emissions Estimation Scenarios. Multiple scenarios for alternative fuel and vehicle technology penetration rates were developed to investigate the spatial characteristics and magnitude of regional emissions. Table 1 describes the present (2005) and future (2030) scenarios.

FIGURE 1. Framework for on-road vehicle emission inventory estimation. 8450

Li ViT

TABLE 1. Scenario Description for Conventional Alternative Vehicle Fuels and Technologies

and

market penetration by vehicle class (%) present scenario (2005) future scenario (2030) vehicle type

fuel technology

LDGV E85 HEV car LDDV CNG EV and fuel cell HDDT truck B20 trucks HDDB Bus CNG bus

baseline alternativea,b baseline alternativea,b

100 0 0 0 0 0 100 0 100 0

73 9.9 9.9 5.9 1.2 0.1 73 27 73 27

100 0 0 0 0 0 100 0 100 0

73 9.9 9.9 5.9 1.2 0.1 73 27 73 27

a

The market penetration rate for each LDV technology in the alternative scenarios is estimated based on the EIA for 2030 (16). b Alternative vehicle technologies are assumed to comprise 27% of the vehicle fleet for each vehicle class.

Each period includes two fleet mix scenarios. The baseline fleet mix includes only contemporary conventional technologies, whereas the alternative fleet mix includes both conventional and alternative technologies. The market penetration rate for each alternative LDV technology is based on the U.S. Energy Information Administration (EIA) (16). Emissions differences between alternative and baseline simulations quantify the influence of alternative vehicle technologies on regional emissions. A sensitivity analysis was carried out to evaluate the effect of varying market penetration rates on regional emissions magnitudes. Comparative uncertainties in the difference between technology scenarios are attributed mainly to the TCFs, since multiple scenarios share common uncertainty sources such as those associated with link travel time and volume. Two levels of future travel demand were considered: no VMT growth and 33% VMT growth, the latter reflecting anticipated regional population and employment growth. The TDM is used to estimate how VMT growth affects linkbased travel speeds, assuming no changes were made to the network road capacity.

Results Emission factors are coupled with outputs from the TDMs to estimate an emission inventory. The impacts of advanced vehicle technologies on regional emissions are quantified. Emission Factor Estimates. The modeling month was July, which represents the peak season for high ambient levels of air pollutants such as ozone, (for which vehicle emissions are a precursor). Calendar year 2005 was selected for the present scenario and year 2030 for the future scenario. We accounted for local hourly average temperature and relative humidity. Since counties comprising the Research Triangle region have implemented a vehicle emissions inspection program, onboard diagnostic (OBD) (21), test and repair computerized (TRC) and evaporative I/M programs were used in the modeling. The assumed OBD compliance rate was 95% and the waiver rate was 5%. Basic Emission Rates and Adjustment for HC, CO, and NOx. BERs were adjusted for local characteristics including ambient conditions, I/M program, vehicle class and age distribution, and emissions standards. BERs were weighted by the distribution of vehicle age in the fleet by vehicle class. The local vehicle age distribution was acquired from the NC Division of Air Quality (21). Vehicles complying with Federal

Tier 2 LDV and 2007 HDV emission standards are assumed to be widely deployed by 2030. Distance-based BERs in units of gram per mile were converted to time-based emission rates using the average cycle speed. Basic Emission Rates for CO2. For LDVs, the FTP cycle is often interpreted as representing urban street driving and the HFET to represent freeway driving. These cycles have average speeds of 21 and 48 mph, respectively. CO2 BERs for LDGVs and LDDVs were estimated based on fuel economy data (22, 23). In addition, the EIA estimates that the average fuel economy for new LDVs in 2030 is likely to increase by 16% (16) when compared to the 2005 figures. For each type of LDV, the same relative improvement in fuel economy was assumed. This percentage is used to adjust fuel economy and estimate CO2 emission factors for all LDV technologies in the future scenarios. CO2 emission factors for HDDV and buses were obtained from MOBILE6, since fuel economy data similar to that for LDVs were not available. Cycle Correction Factors. The ratio of link average emission rates to trip-based speed distributions was estimated using VSP modal models (8). Fourteen discrete VSP modes were defined, and modal average emission rates for LDGVs were estimated from PEMS data (24). The modal models were applied to both the FTP and PEMS real world driving cycles at similar average speed. The distribution of time spent in different ranges of VSP was estimated based on second-bysecond speed and acceleration traces, taking into account real-world road grades for the real world cycles. Average emission rates are subsequently estimated from the product of VSP modal average emission rates and the percentage of link travel time spent in each mode. The CCF for each pollutant is estimated as follows: CCFT ′ )

CERT ′,LSP,v CERT ′,SDC,v

(4)

Where CERT ’,LSP,v ) average cycle emission rate for link speed profile LSP at a speed of v(g/s); CERT ’,SDC,v ) average cycle emission rate for standard cycle SDC at a speed of v (g/s); LSP ) real-world link speed profile; SDC ) standard driving cycle. For HDDVs and buses, the empirical data were inadequate to convert BERs from an engine dynamometer test to an equivalent road speed-based driving cycle. Hence, CCFs for these vehicles could not be estimated and are assumed to be equal to unity. Technology Correction Factors. Gasoline fuel was used as the basis for estimating TCFs for LDVs, excluding LDDVs, whereas diesel fuel was used as a basis for HDVs. TCFs are thus 1.0 for LDGVs, HDDVs, and diesel buses. For LDDVs, TCFs are 1.0 since their BERs were also estimated using MOBILE6. TCF is taken to be zero for electric and fuel cell vehicles since they produce no tailpipe emissions. TCFs for lightduty E85, CNG, and hybrid electric vehicles were estimated based on average certification cycle emission rates for vehicle model years 2001 through 2007 versus those of comparable model, model year, and engine size LDGVs (25). These certification data are based on the FTP cycle and are used to calculate TCFs for CO, HC, and NOx: TCFT ′,T )

CERT,SDC,v CERT ′,SDC,v

(5)

Where CERT,SDC,v) average cycle emission rate based on EPA’s certification tests (gram/mile) for alternative (T ) or conventional (T ’) technologies. VOL. 43, NO. 21, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8451

TABLE 2. Light-Duty Vehicle Technology Correction Factors (TCFs) and Speed Correction Factors (SCFs) for Time-based Tailpipe Emission Factors on Arterials pollutant (ratio) correction factor

average technology speeda (km/h) HC

E85 CNG technology correction HEV Factor (TCF)b,c EV fuel cell

gasoline

E85

speed correction factor (SCF)d

CNG

HEV

diesel

34.1

10-20 20-30 30-40 40-50 >50 10-20 20-30 30-40 40-50 >50 10-20 20-30 30-40 40-50 >50 10-20 20-30 30-40 40-50 >50 10-20 20-30 30-40 40-50 >50

1.12 0.11 0.29 0.0 0.0 0.83 0.94 1.00 1.09 1.20 0.74 0.88 1.00 1.15 1.26 0.75 0.89 1.00 1.15 1.25 0.44 0.72 1.00 1.04 0.99 0.81 0.94 1.00 1.06 1.13

CO NOx CO2 0.78 0.75 0.50 0.0 0.0 0.44 0.78 1.00 1.02 1.19 0.40 0.75 1.00 1.09 1.26 0.62 0.83 1.00 1.19 1.34 0.39 0.69 1.00 1.18 0.95 0.67 0.81 1.00 1.19 1.41

0.92 0.89 0.76 0.0 0.0 0.59 0.82 1.00 1.18 1.36 0.59 0.84 1.00 1.16 1.31 0.57 0.81 1.00 1.22 1.39 0.38 0.69 1.00 1.06 1.00 0.61 0.82 1.00 1.18 1.35

0.98 0.84 0.63 0.0 0.0 0.74 0.89 1.00 1.12 1.23 0.74 0.88 1.00 1.15 1.26 0.74 0.89 1.00 1.14 1.26 0.4 0.7 1.00 1.02 0.93 0.65 0.83 1.00 1.21 1.38

a For speed correction factors, average link-based emission rates for a base cycle and a cycle at the desired link-based average speed range were estimated based on average results of multiple real world cycles within a range of speeds. Ranges of speeds were selected in order to obtain a sufficient number of real-world cycles to lead to statistically stable estimates of link average emission rates. b Technology correction factors for E85, CNG, and HEV are estimated based on the ratio of FTP certification cycle emission rates versus those of conventional LDGVs of the same year, model, and engine size. EVs and fuel cell vehicles do not have tailpipe emissions of the selected pollutants and thus have zero TCFs. c Uncertainty estimates for mean TCFs are detailed in the Supporting Information. For NOx, TCFs for E85, CNG, and HEV technologies are not significantly different from 1.0. The E85 TCFs for HC and CO2 are also not significantly different from 1.0. All other TCFs are statistically significantly different from 1.0. d For SCFs, pollutant emission rates were estimated using modal emission models for a given speed profile. The results shown here are the ratios of estimated emission rates for cycles of the indicated average speed versus those of a baseline speed.

TCFs for CO2 for LDVs were based on the ratios of fuel economies for alternative LDV technologies versus gasoline. TCFs for LDVs are given in Table 2. Compared to LDGVs, E85 fueled cars have higher HC emissions, lower CO emissions, and comparable NOx and CO2 emissions. On the other hand CNG and HEVs have consistently lower emissions for all pollutants. For alternative heavy-duty vehicle technologies, TCFs are estimated from reported comparisons between B20 (20% biodiesel and 80% petroleum diesel by volume) versus petroleum diesel, and CNG versus diesel buses (13, 26, 27). The resulting TCFs for B20 versus petroleum diesel are 0.79, 0.89, 1.00, 1.01 for HC, CO, NOx, and CO2, respectively. Both diesel and biodiesel fuels have similar NOx and CO2 emissions. When comparing CNG versus diesel buses with after8452

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 21, 2009

treatment equipment, the TCFs are 7.6, 0.69, 0.57, 0.85 for HC, CO, NOx, and CO2, respectively (28-30). High levels of CNG-related methane emissions account for the large TCF for HC. Speed Correction Factors. SCFs were developed for vehicle technologies including LDVs, HDVs, and buses. Link emission rates were estimated for these technologies using either VSP or speed-acceleration modal approaches (8, 12, 25, 31). Those in turn were developed from facility-specific link speed profiles measured under real-world traffic conditions by PEMS. The SCFs are estimated from the ratio of average emission rates for speed profiles at or near a link average speed of interest versus those at the reference average speed, similar to that for BER and CCF. SCFT,f,V )

CERT,LSPf,V CERT,LSP,v v′

(6)

The speed profiles were obtained from real-world observations and are subject to variability because of variations in traffic conditions. In order to develop a statistically reliable estimate of link-based average emission rates, multiple real world speed profiles within a specified narrow range of average speeds were used. In previous work, link speed profiles were binned by small average speed ranges (8). Here, the speed range closest to the desired average cycle speed was chosen as the base speed for estimating SCFs. Since the average cycle speed for estimating BERs for HC, CO, and NOx in MOBILE6 is 20 mph, the SCFs are estimated with respect to link average emissions at link average speed range of 19-25 mph on arterials. For HDDVs and buses, the baseline average speed in MOBILE6 for estimating trip-based BERs for HC, CO, and NOx is 20 mph. In previous work, link-based average emission rates were estimated based on multiple real world profiles stratified in 10 mph average speed increments (12). Therefore, the average of link-based emission rates for two speed ranges of 10-20 mph and 20-30 mph on arterials were used as the basis for estimating SCFs for both vehicle classes. HDDV speed profiles were available for a variety of link-based average speeds for freeways, arterials, and local roadways, and were used to estimate HDDV SCFs. In the case of transit diesel buses, real world speed profiles were available on major arterials and local roadways, and these were used to estimate SCFs. Since no freeway speed profiles for diesel buses were available, the MOBILE6 speed correction equations for HDVs were applied to diesel buses on freeways (4). For LDGVs, link-based emission rates for a speed range of 19-25 mph on arterials were used as the basis for estimating SCFs for CO2, whereas link-based emission rates for a speed range of 44-50 mph on freeways were used as the basis for estimating SCFs on freeways. In Table 2, SCFs estimated on an arterial link tended to increase with average speed for various vehicle technologies. Although not shown, SCFs were also found to increase with link-based average speed for freeways and local roadways. SCFs were found to be similar for arterials and local roadways. However, SCFs for on-ramps and off- ramps were found to be significantly different (8). Speed- and Facility-Specific Emission Factors for Conventional and Alternative Technologies. Link-based emission factors are shown in Figure 2. These factors are sensitive to speed variation, for all technologies. For a given speed range, LDGVs and E85 LDVs have higher CO2 emission rates than other technologies. LDGVs have higher emission rates for CO than other vehicle technologies. However, NOx emission rates for LDDVs are higher than those for other vehicle technologies.

FIGURE 2. Tailpipe emission factors for various classes of light-duty vehicles on arterials as a function of link-based average speed (calendar year 2005).

TABLE 3. Estimated Total Triangle Regional Emissions for A Weekday Morning Peak Hour and Percentage of Emissions Changes Compared to Benchmark Scenario (2005 Baseline) scenario year

technology mix

2005 2005 2030 2030 2030 2030

baseline alternative baseline alternative baseline alternative

emissions changes relative to 2005 baseline (%)b

total emissions (tons) a

VMT growth

HC

CO

NOx

CO2

HC

no no yes yes

0.85 0.79 0.15 0.15 0.24 0.24

35 30 10 8 14 13

4.6 4.5 0.4 0.4 0.6 0.6

1400 1300 1200 1200 1800 1800

-8 -82 -83 -71 -72

CO

NOx

CO2

benchmark -14 -3 -72 -92 -76 -92 -58 -87 -64 -88

-4 -13 -15 34 29

a The scenario without VMT growth assumes the same socioeconomic data in the design year as in the base year; The total regional VMT for scenarios without VMT growth is 3.25 × 106 miles; The 28% reduction in average speed for the scenario with VMT growth is obtained from the triangle regional model. b The percentage differences between the 2005 alternative versus 2005 baseline scenario are statistically significant for HC, CO, and CO2, based on an analysis reported in Section S-7 of the Supporting Information. Differences for all other scenarios are significant.

Comparison of link- versus cycle- based emission rates on arterials for LDGVs show differences from 15 to 22% for HC, 22 to 24% for NOx, and 38 to 97% for CO, which indicates larger estimates of emissions associated with the link-based model. These differences are attributed to the activity pattern of real-world duty cycles used as the basis for the cycle and speed correction factors. Emissions Inventory. Regional mobile source emissions are estimated and characterized for multiple scenarios. A sensitivity analysis is carried out to investigate the potential impacts of alternative technologies on regional emissions. Scenario Results. The TRM was executed to produce linkbased vehicle activity data for a weekday morning peak hour on that network. The modeled network includes the cities of Raleigh, Durham, and Chapel Hill. Using eqs 2 and 3 and the vehicle technology distributions shown in Table 1, regional on-road vehicle emissions were estimated and are provided in Table 3. Relative emissions changes compared to the 2005 baseline scenario are also shown in Table 3. Assuming a market penetration rate of 27% for all advanced vehicle technologies (16), the resulting emission reductions are estimated at 8, 14, 3, and 4% for HC, CO, NOx, and CO2, respectively. A comparative uncertainty analysis for light-duty vehicle fleet indicates that the estimated HC, CO, and CO2 emission reductions are significantly different from zero; NOx

reductions, on the other hand, were not. These results imply that, in the short term, there may be only modest emissions reductions, especially for NOx and CO2, with only a limited market penetration of alternative vehicle technologies. All future scenarios are compared with the 2005 baseline scenario to evaluate the impacts of new vehicles and alternative vehicle technologies on regional emissions. As shown in Table 3, year 2030 emissions for HC, CO and NOx significantly decreased by 58% or more with or without alternative technologies. This is mainly because of the turnover of the LDGV and HDDV fleets to Tier 2 vehicles. By 2030, Tier 3 and 4 HDDVs will be in the fleet. Thus, the actual emissions reductions may be larger. For the future VMT growth scenario, the overall average trip speed for the network decreased by 28%, whereas CO2 emissions increased by approximately 30%, irrespective of the fleet mix composition. Thus, improvements in fuel economy and CO2 emissions reduction per vehicle may be offset by VMT growth and decrease in level of service. With VMT growth, a larger portion of the network would be operating at a degraded level of service, since no roadway network improvements were modeled. In this case travel times increase for all modes and vehicles, as well their associated emissions. VOL. 43, NO. 21, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8453

TABLE 4. Network Emissions Distribution by Vehicle Type for Present Baseline Scenario (%) vehicle type

VMT distribution (%)

HC

CO

NOx

CO2

light-duty heavy-duty bus

88 11 1

83.0 17.0 1.0

98.0 1.8 0.2

45.0 48.0 7.0

67.0 30.0 3.0

Regional Mobile Source Emissions Characterization. Total link emissions aggregated across all vehicle classes in the peak hour were estimated and normalized by link length and time duration to illustrate the spatial characteristics (or density) of emissions. Normalized link emissions for the baseline scenario are classified into quantiles. Emissions tended to occur on freeways and ramps in greater proportion than the distribution of VMT on those facilities. For example, 51% of the regional NOx emissions occur on freeways and ramps while travel on those links comprised 39% of the regional VMT. Transportation network emissions were also characterized by vehicle type. The results indicate that while 11% of regional VMT is due to HDDVs, as shown in Table 4, this vehicle class contributed 48% of the regional NOx emissions, whereas LDVs dominate HC, CO and CO2 emissions. This is because HDDVs have much larger emission factors for NOx compared to LDVs. Sensitivity Analysis of Regional Mobile Source Emissions. Since the actual values of alternative technology penetration rates are uncertain at this time, regional total emissions were calculated for the full range of market penetrations. Regional emissions linearly decreased as the total fraction of alternative vehicle technologies increased. With full market penetration, emissions reductions would be on the order of 28% for HC, 51% for CO, 11% for NOx, and 13% for CO2. These results imply that increasing the share of alternative technologies may be one way to reduce regional emissions. However, increasing market penetration may not be sufficient to offset the effect of VMT growth on fuel use and CO2 emission rates unless policies are put into place to improve vehicle fuel economy.

Recommendations This paper has presented a methodology for estimating regional emission inventories based on link-based emission factors that are compatible with outputs generated from regional travel demand models. HEVs were found to be effective in reducing emissions for all pollutants. Therefore, increasing the share of HEVs in the market is advisible. Other potentially effective strategies could include retiring older vehicles, replacing these with new conventional and preferably hybrid vehicles, and slowing or decreasing VMT growth and improving vehicle fuel economy. In terms of vehicle technology, the impacts of new propulsion systems (e.g., HEV) or alternative fuels (e.g., CNG, E85, etc.) on regional emissions were found to depend in part on (a) their market penetration in the vehicle fleet; (b) whether the fleet was characterized by more Tier 2 vehicles; and (c) whether their presence coincided with significant increases in VMT. The default assumption, based on EIA projections of a 27% market penetration of such vehicles in the 2030 fleet and no VMT increase, shows only modest reductions in emissions. However, when coupled with a projected 33% VMT increase between 2005 and 2030 (a very conservative growth assumption), regional CO2 emissions increased. However, substantial reductions in emissions of other pollutants were robust to different assumptions regarding VMT growth, mainly because of fleet turnover. 8454

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 21, 2009

The networks and case study reported here could be refined in future work to account for factors such as (a) expected or possible changes in the transportation network capacity under continued growth patterns; (b) implications of policy changes in land use, such as promotion of smart growth or mass transit; (c) improved characterization of vehicle and fleet characteristics, such as implementation of future emission standards for heavy-duty diesel vehicles; (d) characterization of other advanced technologies such as plug-in hybrids; (e) the need for larger samples to enable the quantification of uncertainties on TCFs for heavy-duty vehicles; and (f) more extensive quantification of uncertainty in comparative estimates of differences in emissions. In addition, fuel life cycle emissions changes need to be considered when replacing conventional technologies with alternative technologies. Finally, basic CO2 emission rates for heavy-duty diesel trucks and buses should be updated when PEMS fuel consumption data become available.

Acknowledgments The work was supported by U.S. EPA STAR Grant R831835 via the University of North Carolina-Chapel Hill, and National Science Foundation grants 0230506 and 0756263. H.Z. contributed to this project as a graduate research assistant at NCSU. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the EPA. Joe Huegy, Mei Ingram, and Bing Mei at Institute for Transportation Research and Education provided Triangle Region Model data.

Supporting Information Available Table and figures regarding source of data, data summaries, VSP-based modal models, link definition, comparisons of cycle- versus link-based emission factors, illustrative uncertainty analysis, emissions spatial characterization, and sensitivity analysis for market penetration rates. This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) U.S. Environmental Protection Agency. National Emissions Inventory (NEI) Air Pollutant Emissions Trends Data: 19702006; EPA: Ann Arbor, MI, July 2007; http://www.epa.gov/ttn/ chief/trends/index.html. (2) U.S. Environmental Protection Agency. Technical Guidance on the Use of MOBILE6.2 for Emission Inventory Preparation, EPA420-R-04-013; EPA: Ann Arbor, MI, August 2004. (3) Miller, C. A.; Hidy, G.; Hales, J.; Kolb, C. E.; Werner, A. S.; Haneke, B.; Parrish, D.; Frey, H. C.; Leonora, R.-B.; Deslauriers, M.; Pennell, B.; Mobley, J. D.; Miller, A. Air emission inventories in North America: a critical assessment. J. Air Waste Manage. Assoc. 2006, 56 (8), 1115–1129. (4) U.S. Environmental Protection Agency. Final Facility Specific Speed Correction Factors; EPA420-R-01-060; EPA: Ann Arbor, MI, November 2001. (5) Unal, A.; Rouphail, N. M.; Frey, H. C. Effect of arterial signalization and level of service on measured vehicle emissions. Transp. Res. Rec. 2003, 1842, 47–56. (6) Frey, H. C.; Rouphail, N. M.; Unal, A.; Colyar, J. D. Emissions Reduction through Better Traffic Management: An Empirical Evaluation based upon On-road Measurements, FHWY/NC/ 2002-001; Prepared by North Carolina State University for North Carolina Department of Transportation, North Carolina State University: Raleigh, NC, December 2001. (7) Frey, H. C.; Unal, A.; Rouphail, N. M.; Colyar, J. D. On-road measurement of vehicle tailpipe emissions using a portable instrument. J. Air Waste Manage. Assoc. 2003, 53 (8), 992–1002. (8) Frey, H. C.; Rouphail, N. M.; Zhai, H. Speed- and facility-specific emission estimates for on-road LDVs on the basis of real-world speed profiles. Transp. Res. Rec. 2006, 1987, 28–137. (9) Caliper Corporation. TransCAD User’s Guide; Newton, MA, 2000.

(10) Pokharel, S.; Bishop, G. A.; Stedman, D. H. An on-road motor vehicle emission inventory for Denver: an efficient alternative to modeling. Atmos. Environ. 2002, 36 (33), 5177–5184. (11) Rykowski, R. A.; Nam, E. K.; Hoffman, G. On-road testing and characterization of fuel economy of LDVs. J. Fuels Lubricants 2005, 114 (4), 276–285. (12) Frey, H. C.; Rouphail, N. M.; Zhai, H. Link-based emission factors for heavy-duty diesel trucks based on real-world data. Transp. Res. Rec. 2008, 2058, 23–32. (13) U.S. Environmental Protection Agency. Alternative Fuel: E85 and Flex Fuel Vehicles, EPA 420-F-06-047’ EPA: Ann Arbor, MI, October 2006. (14) Lave, L. B.; MacLean, H. L.; Hendrickson, C.; Lankey, R. Lifecycle analysis of alternative automobile fuel/propulsion technologies. Environ. Sci. Technol. 2000, 34 (7), 3598–3605. (15) Romm, J. The car and fuel of the future. Energy Policy 2006, 34 (17), 2609–2614. (16) Energy Information Administration. Annual Energy Outlook 2007 with Projections to 2030, DOE/EIA-0383; Office of Integrated Analysis and Forecasting, U.S. Department of Energy: Washington, DC, February 2007. (17) Energy Information Administration. Annual Energy Outlook 2009 with Projections to 2030, DOE/EIA-0383; Office of Integrated Analysis and Forecasting, U.S. Department of Energy: Washington, DC, March 2009. (18) Jime´nez-Palacios, J. Understanding and Quantifying Motor Vehicle Emissions with Vehicle Specific Power and TILDAS Remote Sensing; PhD Thesis, Massachusetts Institute of Technology: Cambridge, MA, 1999. (19) Frey, H. C.; Zheng, J. Probabilistic analysis of driving cyclebased highway vehicle emission factors. Environ. Sci. Technol. 2002, 36 (23), 5184–5191. (20) Institute for Transportation Research and Education. Triangle Regional Model; North Carolina State University: Raleigh, NC, Available at http://www.itre.ncsu.edu/TRM/about/reportsFacts. html. (21) Abraczinskas, M.; Jones, P.; Chandler, V.; Bello, P. Emissions and Air Quality Analysis of the California Low Emission Vehicle II (CA LEV-II) Standards in NC; The 16th Annual International Emission Inventory Conference: Raleigh, NC, May 14-17,2007. (22) U.S. Environmental Protection Agency. Light-Duty Automotive Technology and Fuel Economy Trends: 1975 through 2006, EPA420-R-06-011; EPA: Ann Arbor, MI, July 2006.

(23) U.S. Department of Energy and Environmental Protection Agency. Model Year 2008 Fuel Economy Guide; DOE/EE-0321: Washington, DC, 2008. (24) Frey, H. C.; Unal, A.; Chen, J.; Li, S.; Xuan, C. Methodology for Developing Modal Emission Rates for EPA’s Multi-scale Motor Vehicle & Equipment Emission System, EPA420-R-02-027; Prepared by North Carolina State University for U.S. Environmental Protection Agency: Ann Arbor, MI, August 2002. (25) Zhai, H.; Frey, H. C.; Rouphail, N. M.; Farias, T. L.; Gonc¸alves, G. A. Fuel consumption and emissions comparisons between ethanol 85 and gasoline fuels for flexible fuel vehicles, Paper 2007-A-444; Proceeding of the 100th Annual Meeting of the Air & Waste Management Association, Pittsburgh, PA, June 26-29, 2007. (26) U.S. Environmental Protection Agency. A Comprehensive Analysis of Biodiesel Impacts on Exhaust Emissions, EPA-420-P-02001; EPA: Ann Arbor, MI, October 2002. (27) McCormick, R. L.; Williams, A.; Ireland, J.; Brimhall, M.; Hayes, R. R. Effects of Biodiesel Blends on Vehicle Emissions, NREL/ MP-540-40554; National Renewable Energy Laboratory: Ann Arbor, MI, October 2006. (28) Chandler, K.; Malcosky, N.; Motta, R.; Norton, P.; Kelly, K.; Schumacher, L.; Lyons, D. Alternative Fuel Transit Bus Evaluation Program Results; Society for Automotive Engineers, International Spring Fuels and Lubricants Meeting: Dearborn, MI, May 1996. (29) Ayala, A.; Kado, N. Y.; Okamoto, R. A.; Holme´n, B. A.; Kuzmicky, P. A.; Kobayashi, R.; Stiglitz, K. E. Diesel and CNG Heavy-Duty Transit Bus Emissions over Multiple Driving Schedules: Regulated Pollutants and Project Overview, Paper 2002-01-1722; Society of Automotive Engineers, Spring Fuels & Lubricants Meeting & Exhibition: Reno, NV, May 2002. (30) Melendez, M.; Taylor, J.; Zuboy, J.; Wayne, W. S.; Smith, D. Emission Testing of Washington Metropolitan area Transit Authority Natural Gas and Diesel Transit Buses; NREL/TP-54036355. National Renewable Energy Laboratory: Ann Arbor, MI, December 2005. (31) Zhai, H.; Frey, H. C.; Rouphail, N. M. A vehicle specific power approach to speed- and facility- specific emissions estimates for diesel transit buses. Environ. Sci. Technol. 2008, 42 (21), 7985–7991.

ES900535S

VOL. 43, NO. 21, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8455