Environ. Sci. Technol. 2005, 39, 9130-9139
Quantification of Particle Number Emission Factors for Motor Vehicles from On-Road Measurements L I D I A M O R A W S K A , * ,† M I L A N J A M R I S K A , † STEPHEN THOMAS,† LUIS FERREIRA,‡ KERRIE MENGERSEN,§ DARREN WRAITH,§ AND FRASER MCGREGOR‡ International Laboratory of Air Quality and Health, School of Civil Engineering, and School of Mathematical Sciences, Queensland University of Technology, 2 George Street, Brisbane, QLD 4001, Australia
The database on particle number emission factors has been very limited to date despite the increasing interest in the effects of human exposure to particles in the submicrometer range. There are also major questions on the comparability of emission factors derived through dynamometer versus on-road studies. Thus, the aims of this study were (1) to quantify vehicle number emission factors in the submicrometer (and also supermicrometer) range for stop-start and free-flowing traffic at about 100 km h-1 driving conditions through extensive road measurements and (2) to compare the emission factors from the road measurements with those obtained previously from dynamometer studies conducted in Brisbane. For submicrometer particles the average emission factors for Tora Street were estimated at (1.89 ( 3.40) × 1013 particles km-1 (mean ( standard error; n ) 386) for petrol and (7.17 ( 2.80) × 1014 particles km-1 (diesel; n ) 196) and for supermicrometer particles at 2.59 × 109 particles km-1 and 1.53 × 1012 particles km-1, respectively. The average number emission factors for submicrometer particles estimated for Ipswich Road (stop-start traffic mode) were (2.18 ( 0.57) × 1013 particles km-1 (petrol) and (2.04 ( 0.24) × 1014 particles km-1 (diesel). One implication of the conclusion that emission factors of heavy duty diesel vehicles are over 1 order of magnitude higher than emission factors of petrol-fueled passenger cars is that future control and management strategies should in particular target heavy duty vehicles, as even a moderate decrease in emissions of these vehicles would have a significant impact on lowering atmospheric concentrations of particles. The finding that particle number emissions per vehiclekm are significantly larger for higher speed vehicle operation has an important implication on urban traffic planning and optimization of vehicle speed to lower their impact on airborne pollution. Additionally, statistical analysis showed that neither the measuring method (dynamometer or onroad), nor data origin (Brisbane or elsewhere in the world), is associated with a statistically significant difference between the average values of emission factors for diesel, * Corresponding author phone: +61 7 3864 2616; fax: +61 7 3864 9079; e-mail:
[email protected]. † International Laboratory of Air Quality and Health. ‡ School of Civil Engineering. § School of Mathematical Sciences. 9130
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005
petrol, and vehicle fleet mix. However, statistical analyses of the effect of fuel showed that the mean values of emission factors for petrol and diesel are different at a 5% significance level.
Introduction Vehicle emissions are most commonly characterized in terms of their emission factors, which are defined as the amount of a chemical species emitted per unit mass of fuel burned or per a defined task performed. For particulate matter, emission factors are expressed as mass or number of particles emitted per kilometer driven (g/km or particle number/km) or per unit of mechanical energy delivered (g/kWh or g/bhphr). The current vehicle emission standards for particulate matter refer to the total mass emitted by vehicles (eg., refs 1-3), and therefore the existing data from studies conducted worldwide are most extensive for total mass emissions. Significantly fewer studies have been published on particle number emissions factors from vehicles. Motor vehicle emissions are one of the most significant sources of ambient submicrometer particles (which are most commonly quantified in terms of number emissions or concentrations), since in terms of size, most particles emitted by vehicles are in this size range. There has been an increasing number of epidemiological and clinical studies indicating that the smallest particles in the vehicle emission spectra could be associated with higher health hazards than larger particles (4). The size of airborne particles determines in which parts of the respiratory tract the particles are deposited; submicrometer particles have a high probability of deposition deeper in the respiratory tract (4) and are likely to trigger or exacerbate respiratory diseases. These particles also have higher burdens of toxins, which, when absorbed in the body, can result in health consequences other than respiratory health effects. Therefore, it has been increasingly important to gain knowledge about particle number emission factors. There are a number of methods for determination of motor vehicle emission factors, and the most common experimental ones include (i) direct measurements of pollutant emissions from vehicles placed on a dynamometer and run through certain driving cycles and (ii) indirect estimation through measurements of pollutant concentrations in the close vicinity to a road link. Both these methods have advantages and limitations. In particular, the first method provides emission factors of a relatively small number of vehicles, but for very wellcontrolled conditions, that, however, may not be representative for vehicles on a road. Particle number emission factors using this method have been measured in several studies (5-11), and the number of vehicles included in the studies ranged from 1 to more than 60. The second method does not enable control of either meteorological or traffic conditions (other than choosing the time or locations for the measurements) and requires models to calculate emission factors from the concentrations measured. However, its big advantage is that it provides emission factors of the whole motor vehicle fleet on this road link and, thus, for real world conditions. There have been several studies reported on application of this method for measuring of particle number emission factors, for example, refs 7, 12-16. 10.1021/es050069c CCC: $30.25
2005 American Chemical Society Published on Web 11/03/2005
The question commonly asked is how do the emission factors quantified using the two methods compare for the same vehicle fleet (for example, vehicles operating within a particular urban region). Over the past few years a number of studies were conducted in Brisbane, Australia, to quantify particle number emission factors through dynamometer studies for selected types of diesel and petrol vehicles (5, 6, 11, 17, 18). There has also been a model developed for quantification of vehicle emission factors from on-road measured pollutant concentrations (15). These investigations provided a foundation for the current study, whose aims were formulated as follows: (1) to quantify vehicle number emission factors for selected driving conditions through extensive road measurements and application of the model and (2) to compare the emission factors from the road measurements with those obtained previously from dynamometer studies conducted in Brisbane. The second aim was extended to include a comparison with emission factors from studies conducted elsewhere and published in the literature. Although the main focus of this work was on particle number emissions in the submicrometer range, which is by far the most significant component of the total particle number concentration, for completeness of the data, measurements were also conducted for particle number emissions in the supermicrometer range.
Box Model A model (referred to as the box model) based on a particle number or mass balance equation developed, calibrated, and validated (15) to estimate the emission factors of a car fleet for on-road conditions was used in this work. Parameters required as input to the model are (1) time series of targeted pollutant concentration on the road and background; (2) wind direction and velocity; (3) traffic data (vehicle number, traffic volume, car speciation, etc). An output of the model is pollutant emission per vehicle per kilometer. The average emission factor is calculated as
EFAV ) EFAV )
EFtotal 1 N(t) dx
(CB(t) - C0(t))(vxH + vzW) dx N(t) dx
(1)
(2)
where CB(t) is the particle concentration at the road [particles m-3 or mg m-3]; C0(t) is the particle background concentration [particles m-3 or mg m-3]; H is the height of the box (semiempirical model parameter) [m]; W is the width of the box (usually the width of the road) [m]; dx is the differential of the box (segment) length [m]; vz is the normal component of the vertical wind velocity [m s-1]; vx is the normal component of the horizontal wind velocity [m s-1]; N(t) is the traffic flowrate [cars s-1]; EFtotal is the total emission rate [particles s-1 or mg s-1]; EFAV is the average emission factor [particles km-1 car-1 or mg km-1 car-1]. Only the overall or total emission factors for the submicrometer (18-880 nm) and supermicrometer (0.5-20 µm) particles is provided in the paper. Determination of the sizesegregated emission factors, for example, for ultrafine particles, that could be valuable for the health impact assessment, has not been investigated, since it was not the aim of the study and would be beyond the scope. The box model represents a simplified version of previously used techniques for determination of pollutant concentration inside an imaginary enclosure encompassing an investigated segment of a road. Contrary to the other models, it does not require as input concentrations of a pollutant measured at different heights but a concentration value
FIGURE 1. Map of the locations of the reference site AMRS QUT and the sampling sites at Ipswich Rd and the SE freeway at Tora St. measured at one height only. It is very important for longterm monitoring programs, when it is usually impossible to set up a number of monitors for each pollutant at different heights, and a common practice is to employ only one monitor for each pollutant. The gain from long-term monitoring campaigns is that the trends in seasonal, daily, and meteorological changes in emission levels can be more reliably established. The presented model is a very convenient tool for such applications.
Experimental Section Two monitoring sites (Tora St and Ipswich Rd) and one reference site (AMRS) were selected for the measurement program. Continuous measurements of particle characteristics at each of the two monitoring sites were conducted over about 3 months at each site. Traffic data in terms of total vehicle flow for the two sites for the entire monitoring period was obtained from the Main Roads department, QLD government. In addition, traffic was also videotaped for a number of days at each site to enable vehicle speciation. Particle concentration data and traffic data were used as input parameters for the box model to calculate vehicle emission factors. Description of the Measurement Sites. The monitoring site at Tora St, located about 10 km from the Brisbane CBD, represents typical freeway traffic conditions with the majority of vehicles traveling on the South-East (SE) freeway in a freeflowing traffic mode at a steady speed of approximately 100 km h-1 at off-peak periods. Due to the distance of the site from the city on one hand and the major suburban locations on the other, it was assumed that by the time vehicles reach this location, engines are operating in a steady temperature mode. The freeway consists of four lanes, with two lanes in each direction. The road is flat in this area, and the topography of the site could be characterized as a semiopen, street canyon type as the road was cut through rock formation with the height of the walls about 5 m. The second site at Ipswich Rd, which is about 3 km from the CBD in a high urban density area, represents typical urban city type traffic conditions with the majority of vehicles traveling in a stop-start mode between two sets of traffic lights for pedestrian crossings, with the maximum speed of 60 km h-1. The road is flat in this area, and the topography of the site could be characterized as an urban canyon street type as the road passes between buildings of height of about five levels. The road consists of six lanes, with three lanes in each direction. The locations of the reference and the sampling sites and the surrounding areas are presented in Figure 1. The fleet composition at the two sites varied, with heavy duty diesel vehicles contributing up to 20% of all vehicles at Ipswich Rd and about 5% at Tora St. Almost all passenger cars in Australia are fueled by petrol. VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9131
FIGURE 2. Periods and parameters monitored at the Tora St and Ipswich Rd sites. The reference site was the Air Monitoring and Research Station (AMRS) operating at the Queensland University of Technology Gardens Point campus in the Brisbane CBD, and is part of the South East Queensland Monitoring Network (19). At this site particle and gaseous concentrations have been monitored on a continuous basis since 1995. The reference site was at a distance of about 10 and 3 km, respectively, from Tora St and Ipswich Rd. While ideally the reference and the measurement sites should have been close by, it would have been impossible within the scope of the project to duplicate the existing, continuously monitoring reference site, equipped with the advanced instrumentation for particle monitoring. Extensive tests were conducted to establish the relationship between background concentrations measured at the AMRS and those at the sampling sites. Particle Measurements. The monitoring instrumentation was housed in an air-conditioned cabinet located on bridges across the traffic routes. There was negligible local vehicle traffic on the bridge at Tora St and only pedestrian traffic on the bridge over Ipswich Rd. At the latter site the overpass is located between the traffic lights, with a distance of about 20 m on both sides. Sampling tubes were suspended 3 m from the bridge to provide particle concentration data required by the box model (15). The box height parameter H, used in the box model eq 1, was determined experimentally for each monitoring site. The H values were 8.4 and 7.9 m for the Tora St and Ipswich Rd sites, respectively. The effect of particle losses in the sampling lines was assessed and the measured concentration values corrected. The following instrumentation was used for the measurements of particle number concentrations: • Two scanning mobility particle sizers (SMPS), each consisting of the electrostatic classifiers (EC TSI model 3071A) and condensation particle counters (CPC TSI models 3010 and 3022) for the measurements of particle number concentrations and size distributions in the size range of 0.0180.880 µm (termed submicrometer particles in this paper). • Aerodynamic particle sizer (APS) (TSI model 3320) for the measurements of particle number concentrations and size distributions in the size range of 0.5-20 µm (termed supermicrometer particles in this paper). The measurements of particle concentrations were conducted for the two on-road sites during a monitoring program conducted from September 1998 to October 1999, with about 3 months of monitoring at each site. Figure 2 presents the time period parameters that were monitored at each location. Data for particle concentrations and size distributions at the AMRS reference site were collected through the entire period of the road-monitoring program. Both the SMPS and APS instruments were set up for continuously repeating scans of particle size distributions in 5 min time intervals. The sampling interval is relatively long compared to fast changes 9132
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005
in traffic conditions; however, since the monitoring was continuous over a long period of time, the 5 min sampling was selected in order to average the effect of underestimations and overestimations of individual measurements. This was since the focus was on determination of the average emission factors for the whole fleet and not necessarily of the individual cars. Realistically, the minimum sampling time for the SMPS system used in conditions as in this study is about 3 min. Increasing scanning time provides better resolution and broader size range. It also provides sufficient time for mixing of the car fleet emissions. During a pilot study both 3 and 5 min sampling was conducted, and the differences between data (size distribution, concentration levels) were found to be not significant. The instruments were calibrated prior to the measurements using a known size aerosol produced from a TSI condensation monodisperse aerosol generator (model 3475), and the results were corrected for particle losses in the sampling lines. A correction factor, determined as a ratio of total particle concentration measured (i) with and (ii) without the sampling line, was applied to instruments’ reading to correct for the effect. Monitoring of Meteorological Parameters. A portable weather station (Davis Instruments’ weather monitor II) was used to monitor local meteorological parameters including horizontal wind direction and velocity at the two road sites during measurements of pollutant concentration. Continuous monitoring of wind characteristics as well as temperature and humidity was conducted at the AMRS. The emission factors were calculated using local wind data, which provided more accurate characterization of traffic emissions’ dispersion. Only horizontal wind component values were used in emission factors calculation, since the vertical wind speed components, quantified at each site at the beginning of each monitoring period, were found negligible compared with horizontal wind speed components. To avoid a cumulative effect of the traffic emissions transported along the road, only the data set measured for the wind direction outside of the (20° alignment with the road and wind velocity higher than 3 km/h were considered for the calculation of emission factors. Monitoring of Traffic Volume and Fleet Composition. Traffic flowrate data were obtained from two permanent traffic counters located at the Ipswich Rd and SE freeway at Tora St located approximately 20 m from each sampling point. The counters (electronic inductive loops) were operated by the Main Roads department, QLD government. For both sites the electronic data were validated against video-recorded data at the beginning and during the monitoring campaigns. The electronically recorded and visually observed values were in close agreement. Traffic speciation data was also collected at intermittent intervals for 4 days at Tora St and Ipswich Rd. This was done by video-taping traffic at the sites and then analyzing the tapes to obtain information on selected vehicle categories. Data Preparation. Each set of measured data including particle, traffic, and meteorological characteristics was preliminary screened for their validity and completeness. Screening involved assessment of data normality, type of the distribution (particle concentration), and identification of outliers, using boxplot and nonparametric methods. Collected data was exported into an Access database for easy manipulation, retrieval, and processing. Only subsets containing a complete set of data were used for calculation of the emission factors.
Results and Discussion Relationship between Background Concentrations at the Reference and Monitoring Sites. Assessment of the monitoring data collected at the AMRS for the purpose of its
suitability as background data for the two road sites was conducted in two steps. First, investigations were conducted to determine whether the measured concentrations are affected by a very local source, which is the nearby freeway, and if so, to quantify the effect. Second, comparison was made between concentrations measured at AMRS and the monitoring sites at the time of minimum traffic on the road, therefore with the least impact from vehicles, as well as simultaneously measured at AMRS and at a location upwind from the Tora St site. Impact of the Freeway on the AMRS Concentrations. The AMRS is located approximately 150 m east from the SE freeway, and although the distance and vertical elevation (air sampled from sixth floor) are relatively large, thus decreasing the effect of the freeway, the reference data could under certain conditions (west and southwest winds blowing from the freeway toward the AMRS) be affected. The effect was investigated and quantified in a week long intensive measuring campaign, in which concentration levels of particles were measured simultaneously upwind and downwind from the freeway. A second sampling location, in addition to the AMRS, South Bank, was set up westerly from the freeway, approximately 350 m from the AMRS. The sampling site was on the opposite bank of the Brisbane river, running in this area immediately next to the freeway (the freeway is located between the AMRS and the river). Data analyses showed that for the west and southwest wind conditions (i.e., the AMRS was downstream in relation to freeway) occurring usually early in the morning and late in the afternoon, the AMRS concentration levels are affected by the emissions from the freeway. For all other wind directions, the AMRS and the South Bank concentrations were the same within the limits of experimental error. A correction factor of 2.45 ( 0.57 was derived and later used for the AMRS data to remove the effect of the freeway on the concentrations, so they could be used as a background for the current study. Comparison between AMRS and Road Sites Background Concentrations. As discussed above, in this study the local background data (which would be upwind road concentrations) was to be substituted with data measured at the AMRS reference site. Thus, an assumption was made that these are closely related. The validity of this assumption and quantification of the relationship was achieved by (i) comparing the reference (AMRS) and two road sites data obtained for no traffic conditions at the test sites observed during the night time between 1:00 a.m. and 4:00 a.m. and (ii) conducting upwind concentration measurements at the Tora St site and comparing the data with the AMRS concentrations. The correlation between night time concentrations was assessed using the nonparametric Spearman rank correlation test and showed significant correlation at the 0.01 level between AMRS and Tora St (Spearman’s rho ) 0.424; n ) 577) and AMRS and Ipswich Rd (Spearman’s rho ) 0.481; n ) 639) road sites. In addition, parallel measurements at the AMR reference site and a residential area approximately 1 km away from the Tora Street, representing the local background conditions, were conducted over a period of 2 days. The ratio between particle concentration levels measured at the AMRS (corrected for the freeway contribution) and the Tora St background levels was 1.02 ( 0.35 (n ) 88). On the basis of the above it was concluded that the AMRS concentration data, after correction for the freeway contribution for winds from the west and southwest, can be used as a background for the two road sites. Median Emission Factors for the Whole Traffic Fleet. The mean is a useful representative value of the emission factor (EFAV) if the distribution of the emission factors is symmetrical. If not, then the median emission factor (EFMED) is a more appropriate way of presentation of the results.
FIGURE 3. Median emission factors for submicrometer particles determined for different time intervals for the Tora St and Ipswich Rd monitoring sites. WD, week day; WE, weekend. The 25th and 75th percentile ranges are indicated by Y-axis error bars.
FIGURE 4. Median emission factors for supermicrometer particles determined for the Tora St monitoring site. WD, week day; WE, weekend. The 25th and 75th percentile ranges are indicated by Y-axis error bars. Inspection of the data showed that the distribution was not symmetrical, so EFMED is reported. The box model was subsequently employed to calculate the submicrometer particle number EFMED and the interquartile ranges (IQR) for the two measurement sites for the entire data sets and under varying time constraints (see Figure 3). The various time constraints were expected to reflect different driving conditions (for example slower traffic during rush hours) and different vehicle fleet mixes (for example a greater proportion of petrol vehicles during rush hours). Comparison between the emission factors calculated for the two sites was also expected to provide information that may also be related to the driving mode. In particular traffic at Tora St is mostly free flowing at high speed, while traffic at Ipswich Rd is mostly stop-start. The box model was also applied for the calculation of supermicrometer particle number EFMED and IQR for both sites, for the entire data sets, and under varying time constraints. The results obtained for the Tora St site are presented in Figure 4. The concentrations of the supermicrometer particles measured at Ipswich Rd were of the same order as those measured at the AMRS reference site (i.e., background levels, average of ∼1.7 particles cm-3), which made it impossible to calculate EFMED for these particles at Ipswich Rd. A number of conclusions can be made from the analyses of the results presented in Figures 3 and 4. First, it can be seen that by measurement of only a small number of VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9133
parameters at the two road sites and application of the model, it was possible to derive EFMED for the entire vehicle fleet traveling at these locations for different time periods. Second, for the submicrometer particles, the emission factors measured at Tora St are higher than those at Ipswich Rd. For example the EFMED quantified for Tora St for the entire data set was 1.11 × 1014 particles km-1 per vehicle, compared to 5.67 × 1013 particles km-1 per vehicle for Ipswich Rd. The majority of the vehicles on the motorway at this location travel at speeds of the order of 100 km h-1, while on Ipswich Rd vehicles travel at low speeds, under stop-start conditions. It can thus be concluded that EFMED is about 2 times higher for the vehicle speeds of the order of 100 km h-1, compared to the situation at traffic lights, including stopstarts and low-speed travel. This trend is expected, because at higher speeds (and thus loads) vehicles are known to consume more fuel and emit more particulate matter than at idle or at lower speeds, as demonstrated, for example, in refs 5 and 12. Third, for the submicrometer and supermicrometer particles the EFMED for both locations during the morning peak periods (6:30-8:30 a.m.) are at a minimum compared to other periods. The explanation of this finding is that on one hand vehicles are traveling slower during this period because of frequent traffic congestion and the emissions are lower at lower speeds, and on the other, the fraction of petrol vehicles (of lower emission factors) at this time has been shown to be higher than that of diesel vehicles (of higher emission factors). The finding that particle number emissions per vehicle-km are significantly larger for higher speed vehicle operation has an important implication on urban traffic planning and optimization of vehicle speed to lower their impact on airborne pollution. The emission factors were determined from the measured particle number concentration, traffic characteristics, and wind velocity. The most dominant factors affecting particle concentration levels of traffic emissions and thus the emission factors were traffic flowrate and wind velocity. Some laboratory studies indicate that the relative humidity and ambient air temperature may also play an important role in generation of the secondary, nucleation mode particles. The effect was investigated using a range of statistical methods including classical PCA/FA, linear and tree regression (CART), and correlation techniques. The results confirmed the laboratory findings, showing a weak but statistically significant (p ) 0.05) inverse relationship between ambient air temperature and a direct relationship between RH and particle number concentration measured at both monitoring sites. A more detailed analysis of the dependency for size-segregated particle emissions showed that the most pronounced effect occurred for nuclei mode (smaller than 30 nm) particles. The results indicate that emission factors are affected by the ambient air parameters, showing a statistically significant increase in the emission factor values for relatively humid (relative humidity >80%) and cold (temperature 0.010 µm EAA (electrical aerosol analyzer)
this study
Cadle et al. 1999 (8)
Cadle et al. 2001 (9)
chassis dynamometern
1.57 × 1014
fleet mix SMPS (n ) 3) diesel SMPS (n ) 12) SMPS (n ) 5) petrol SMPS (n ) 6) fleet mix SMPS (n ) 46) diesel buses SMPS (n ) 1) petrol SMPS fleet mix petrol (mean ( SE) diesel (mean ( SE) SMPS fleet mix petrol (mean ( SE) diesel (mean ( SE) petrol FTP summer (n ) 25) winter (n ) 67)
(1.75 ( 1.18) × 1014
Ketzel et al. 2003 (14)
on-road study inverse modelingq
0.010-0.7 µm
Kittelson et al. 2004 (12)
on-road datar
0.008-0.3 µm
diesel FTP ELPI petrol (n ) 38) diesel (n ) 7) SMPS (n ∼ 1000 estimate based on 6 month) fleet mix SMPS
0.003-0.9 µm
avg ( STD DMPS fleet mix
Farnlung et al. 2001 (10)
Kristensson et al. 2004 (13) tunnel measurementss
0.01-10 µm (ELPI)
NEF (particles km-1)a
SMPS (n ) 12) diesel SMPS (n ) 11) petrol SMPS (n ) 8800)
diesel FTP (n ) 12) >0.003 µm (UCPC) non-oxygenated petrol 0.03-10 µm (ELPI) FTP oxygenated petrol petrol FTP
(NFRAQS)o
on-road study chassis dynamometerp
method/ instrumentation
1.74 × 1011
5.9 × 1014 (3.87 ( 2.49) × 1014 (4.5 ( 2.0) × 1013 (2.82 ( 0.68) × 1014 (3.23 ( 2.16) × 1014 (2.44 ( 1.41) × 1014 j 2.1 × 1013 (1.11 ( 0.90) × 1014 j (1.89 ( 3.40) × 1013 (7.17 ( 2.80) × 1014 (5.67 ( 2.80) × 1013 j (2.18 ( 0.57) × 1013 (2.04 ( 0.24) × 1014 2.43 × 1014 1.06 × 1014 3.42 × 1014 1.68 × 1013 1.93 × 1013 4.01 × 1014 1.7 × 1013 1.0 × 1014 (2.8 ( 0.5) × 1014 8.70 × 1013 2.73 × 1014 2.24 × 1014 (1.95 ( 0.96) × 1014 (4.6 ( 1.9) × 1014
a Results are presented as (mean ( SD), unless specified otherwise. b The bus fleet consisted of 12 diesel-powered buses tested at a steady-state mode at a speed of 80 km/h and intermediate engine power (0.5PMax). c Eleven petrol-powered cars tested on a dynamometer at steady-state mode at 100 km/h. d Fleet consisting of approximately 94% LDV and 6% HDV. Vehicle speed approximately 100 km/h. e Three diesel-powered trucks tested on a dynamometer at steady-state at a speed of 40-80 km/h and an engine power of 0.50PMax, mode 5. f Twelve diesel-powered BCC buses tested at a steady-state mode at a speed of 40-80 km/h and an engine power of 0.25PMax. g Five unleaded petrol tested on a dynamometer at steady state at 100 km/h. h Fleet consisting of approximately 86% LDV and 14% HD ( 2.49 driving at ∼100 k/h. i The bus fleet consisted of approximately 300 diesel-powered BCC buses running at an average speed of 60 km/h and an engine power of 0.25PMax (estimate). j Median ( Q (semiquartile). k Six-cylinder sedan car ULP (tested at 120 km/h speed). l Fleet consisting of approximately 94% LDV and 6% HDV driving at speed of about 100 km/h (Tora St data). m Fleet consisting of approximately 85% LDV and 15% HDV measured under varying urban driving conditions including acceleration, constant speed, deceleration, “stop and go” or idling) (Ipswich Rd data). n Average emission factors determined from the FTP test for gasoline vehicles measured in summer (n ) 25) and winter (n ) 67). Average diesel FTP emission factors measured indoors at winter (n ) 12). o Average emission factors determined from the FTP test for diesel vehicles (refs 2, 3 cited in Cadle et al. (9)). p Thirty-eight petrol and seven diesel new vehicles (mileage
