Pedestrian Exposure to Ultrafine Particles in Hong Kong Under Heavy Traffic Conditions

Vehicle generated particle emissions represent a major source of air pollution in urban environments. Recent studies show that monitoring particulate matter in the ultrafine particle (UF) size range (diameter < 100 nm) is critical for assessing adverse health effects. A consensus is emerging that particle number concentration, rather than particle mass, may constitute a better predictor of health effects resulting from exposure to particulate matter (Oberdorster et al., 1990; Pekkanen et al., 1997; Peters et al., 1997; Laden et al., 2000). In this study, a water-based condensation particle counter (WCPC) was used to measure particle number concentrations at a busy intersection of Lai Chi Kok Road and Nathan Lane, located in the center of the urban mega city Mong Kok of Kowloon, Hong Kong. Individual particle numbers and traffic patterns revealed that spikes in the particle number concentration coincided with vehicle acceleration. The highest average particle count (~9.0 × 10 particle/cm) was observed in an area next to the Hong Kong Environmental Protection Department measurement station (Site A), followed by ~5.5 × 10 particle/cm measured at the roadside walkway at the Pioneer Shopping Center (Site B), and ~4.5 × 10 particle/cm in front of the SKH Kei Wing primary school (Site C). The highest particle counts occurred when vehicles accelerate, after stopping at a signal light or a bus stop; a peak concentration of 5.4 × 10 particle/cm was observed during acceleration of a heavy-duty diesel bus. Peaks with particle number similar to this were reported for a Los Angeles freeway which has the highest percentage (25%) of diesel vehicles (Zhu et al., 2007).

Ultrafine particles can enter the circulatory system and are capable of causing both acute and chronic adverse health effects (Nemmar et al., 2002;Donaldson and Stone, 2003;Oberdorster et al., 1990).Studies have shown that particles in the UF size range are able to penetrate cellular organelles such as mitochondria (Li et al., 2003;Li et al., 2004).
Measurements at a major Los Angeles freeway showed that the concentration of smaller UF particles can be up to 25 times greater than ambient counts (Zhu et al., 2002a, b).A study of in-cabin passenger exposure to UF particles on Los Angeles roads and freeways revealed dramatic increases in particle number concentration from nearby vehicle exhaust (Miguel et al., 2003), and that particle size distributions, observed both in-cabin and on road, were bi-modal with diameters around 10-20 nm and 50-70 nm (Zhu et al., 2007) et al., 1993;Pope el al., 1995)

EXPERIMENTAL METHOD
A water-based condensation particle counter    1 summarizes the average particle number concentrations for each of the sites along with the distance of the WCPC inlet probe from the curb.The highest peak 8.26 × 10 5 particle/cm 3 (Fig. 2(a)) occurred at 14:12:12 hour at Site A, collocated with the HKEPD station.At this site, the observed particle count averaged 9.0 × 10 4 particle/cm 3 during the sampling period (Table 1).
An average particle count of 5.5 × 10 4 particle/cm 3 was measured at Site B by the Pioneer Shopping Center (Fig. 2(b)), and were generally lower in front of the SKH Kei Wing Primary School (Site C) averaging 4.5 × 10 4 particle/cm 3 (Fig. 2(c)).In Figs. 3 a-b, the darkly shaded portions represent periods of red light and lightly shaded portions refer to periods of green light.When traffic patterns are superimposed on the particle count vs. time data (Fig. 3(a-b)), we clearly see that the peaks coincide with changes of traffic lights at site A.
The ability of the WCPC to take data at a 1second time resolution was integral to the ability to clearly view these peaks.

DISCUSSION
As expected from the high volumes and types of vehicles at any given moment, the peaks for the particle number counts show a diversity of forms and shapes.Incomplete mixing and complex plume dispersion patterns present a situation with many variables to consider.Expanded examples of peaks can be seen in Figs.3(a-b) for measurements taken at site A. Many of the observed spikes of particle counts are small and typically last for a short period of time (10 to 45 seconds).Often, these spikes quickly ascend to a maxima of 1.6 × 10 4 particle/cm 3 before descending rapidly (Fig. 3(a)).The observed patterns regularly follow green lights; the peaks are clearly discernable from baseline count measurements.
In contrast to the small spikes observed in For Site A, the average PM count during the totaled green-light periods is generally higher (~9.4 × 10 4 particle/cm 3 ) compared to the average counts during red-light periods (~7.8 × 10 4 particle/cm 3 ), whereas the total average is 9.0 × 10 4 particle/cm 3 .Such patterns are consistent with previous findings that vehicle acceleration is associated with increased particle emissions (Maricq et al., 1999a, b;Imhof et al., 2005).Due to variations in the general flow of traffic, the lag time between the peaks and the light change is somewhat variable and peaks do not always fall neatly into 'stop' and 'go' periods.Some of the traffic phenomena that may contribute to lag times include traffic jams, vehicles not following traffic laws, as well as periods with no vehicles at all.However, the general trend shows that high points of vehicular particle counts occur during the green light-while pedestrians are waiting for the light to change.On the other hand, many idling vehicles do not show such high particle counts.

CONCLUSIONS
Atmospheric dilution, increased distance from exhaust pipes, and time contribute to rapidly decreasing particle counts in the ultrafine and nanoparticle range after emission.
The average particle number concentration around the intersection was 6.2 × 10 4 particle/cm 3 .An order of 10 5 particles/cm 3 is often associated with acceleration of a heavy duty vehicle.High particle counts occur when vehicles accelerate.With a busy public transportation system, the large pedestrian population is vulnerable to exposure of toxic particulate matter.While these results are preliminary, they clearly suggest that reducing congestion would contribute greatly to improved air quality in this area, therefore diminishing pedestrian exposure to toxic UF particles.
to the SKH Kei Wing Primary School for allowing the measurements during a school session, and to Mike Sommer for graphical design work.
. Since the majority of Hong Kong's urban population spends time outdoors as pedestrians close to high volume of traffic, quantitative knowledge is needed regarding the extent of their exposure to vehicular UF particles that may cause adverse health effects.The Hong Kong Environmental Protection Department (HKEPD) operates an air quality monitoring station measuring various pollutants at the intersection of Lai Chi Kok Road and Nathan Lane.However, the respirable suspended particulate matter (RSP) measurements taken by this station are dependent on mass concentrations whereas increasing detrimental health effects are better correlated to decreasing particle size (Dockery

(
WCPC, TSI Inc. Model 3785, Saint Paul, MN) was mounted on a mobile platform at a height of 1.0 m from the ground.WCPC data were collected in 1-second intervals to provide high resolution temporal results.Power was supplied by a deep cycle 12 V battery and inverter to run the WCPC (Hering and Stolzenburg 2005a; Hering et al., 2005b) and a laptop.A 5 cm length of copper tubing (0.635 cm dia.) pointed towards the road was used to collect samples.Data acquisition was performed on site and in real-time with a laptop computer running the 32-bit TSI (2005) Aerosol Instrument Manager, a program that permits instrument control and data collection, management, and export capabilities.The measurements were conducted in central Kowloon, Hong Kong near several high volume pedestrian walkways around the intersection of Nathan Lane and Lai Chi Kok road.Nathan Lane and Lai Chi Kok road each have 6 lanes, 3 in each direction.Like much of Hong Kong, the intersection is a street canyon bounded on all sides by high-rise commercial and residential buildings.Due to this, wind patterns in the immediate vicinity of the intersection are complex.Air movement was generated by the vehicle traffic as opposed to a steady flow.To gauge the general traffic flow, we counted the number of vehicles that passed a predetermined line perpendicular to the road in front of the measuring site.The layout of the intersection and measurement sites is shown in Fig. 1.Locations labeled A, B, and C, identify landmarks: the HKEPD roadside monitoring station, the sidewalk of the Pioneer Center shopping mall, and the entrance of the SKH Kei Wing Primary School, respectively.The equipment-to-roadside distances are 1 meter for Site A, 5 meters for Site B, and 6 meters for Site C.This distance is defined as the distance of the WCPC instrument inlet to the roadside curb.The particle counts of each site are most greatly affected by the road immediately adjacent: the north side of Lai Chi Kok road at A, Nathan Lane at B, and the south side of Lai Chi Kok road at C. Occasional showers took place between 15:00 and 16:50 hour; and continuous rainfall soon after 17:12 hour until the end of the sampling campaign at 17:51 hour.RESULTSMeasurementswere conducted on 30 July 2005 totaling approximately 4 hours, starting at 14:00 hrs.Fig. 1 shows vehicular flow around the intersection and the average number of vehicles per hour: 282 vehicles/hr north or south-bound on Nathan Lane; on Lai Chi Kok road, 150 vehicles/hr south-east bound and 126 vehicles/hr north-west bound.Numerous traffic lights and bus-stops keep the traffic flowing at low speeds.

Fig. 1 .
Fig. 1.Location A is located next to the HKEPD air quality monitoring station; location B at the Pioneer Center; and location C at the SKH Kei Wing Primary School.For each site, the road immediately adjacent is the main influence of the particle counts observed.Vehicular flow around the intersection and the average number of vehicles per hour are shown: 282 vehicles/hr north or south-bound on Nathan Lane, and on Lai Chi Kok road, 150 vehicles/hr south-east bound and 126 vehicles/hr northwest bound.

Fig 3
Fig 3(a), larger peaks with high and prolonged particle counts dominate the distribution plots taken some 20 minutes later, as shown in Fig. 3(b).A sharp rise in particle count occurs approximately 3 seconds after the traffic light changes to green, as a result of vehicle acceleration.Six seconds later the particle count plateaus at 5.4 × 10 5 particle/cm 3 , after which it descends slowly over the next 35 seconds until the start of the red light period.

Table 1 .
Average particle number concentrations and rain occurrence at each location.