Determination of atmospheric PM 10 concentration in Kandy in relation to traffic intensity

: Atmospheric particulate matter (PM) concentration was measured at 25 sites in Kandy. The city is surrounded by hills and is the second most populous city in Sri Lanka, where for 3 continuous peak hours the traffic intensity is in the range of 0 to 2640 vehicles/h. A high-volume air sampler with the cyclone attachment was used to measure the concentration of particulate matter less than 10 um in size (known as PM |0 ). The concentration of PM 10 in the atmosphere at the average breathing level is found to have 82% variance in common with the traffic intensity. Colour indices assigned for the filtered samples, which represent the intensity of the airborne carbon particles, of the high-volume air sampler have 91% variance in common with the corresponding colour indices assigned for the filtered samples of the low-volume personal air sampler.


INTRODUCTION
In this study, we investigated the atmospheric suspended particulate matter (SPM) pollution levels in Kandy and its surroundings. SPM, also known as aerosols, include mineral dust, acid species, black carbon (or soot) and polycyclic aromatic hydrocarbons. The primary objective of this study was to assess the SPM of size less than 10 pm, known as PM , which is respirable and hence could enter the lungs 5 . PM 10 concentrations measured using a high-volume air sampler (HVAS), were correlated with traffic intensity. The second objective was to correlate the SPM levels measured by the HVAS to that measured by a personal air sampler (PAS), in order to assess the limitations arising from using the low-cost, easy-to-transport PAS to measure SPM levels. The cost of engaging the battery-operated and lightweight PAS to qualitatively assess the intensity of SPM is much less when compared to the HVAS and is more convenient to use.

METHODS AND MATERIALS
Air sampling was carried out at 25 sites covering the range of traffic intensities from 0 to 2640 vehicles/h. The sampling sites were grouped into 5 clusters and brief descriptions of them are given in Table 1. A HVAS (Envirotech APM 460, India; ± 10 ug particle weight variation; ± 2% airflow rate variation) was used for atmospheric SPM sampling, which has a 'Corresponding author 246 M.A. Elangasinghe and R. Shanthini Cluster 5: sites with very low traffic intensity away from city, that serve as control sites. Comparison of filter samples obtained from HVAS and PAS has been carried out using a colorimetric method 9 .

Gannoruwa Secondary
Traffic flow count for the 3 h period in the two directions was monitored and recorded at each sampling site. Average traffic intensities for the 3-hour period of sampling are shown in Figure 1. The highest PM 10 concentration of 340 ug/m 3 recorded at Site 4.2 corresponds to 2640 vehicles/h, the highest traffic intensity. This site was a 3-way junction at Katugasthota situated along the Kandy-Matale national highway (A9) that feed traffic into the city from the northern part of the country. The detailed traffic information collected in this study showed that the intensity of heavy vehicles that transport goods was high during the time of sampling and that the traffic intensity of buses and school vans was gradually increased during the period 11.00 a.m. to 2.00 p.m. The road congested most of the time because of the narrow road structure.

RESULTS AND DISCUSSION
The second highest PM ]0 concentration of230 p.g/m 3 was recorded at Site 3.1, where the traffic intensity was as high as 1785 vehicles/h. This site was located along the Kandy-Peradeniya road in front of a popular Buddhist temple, known as the Gatambe Temple, close to the Peradeniya Police Station, Peradeniya Teaching Hospital, a bank, a school, the Botanical Garden and few other offices. Traffic from western and south-western parts of the country (such as Colombo, Nuwara Eliya, Gampola and Nawalapitiya) also passed by this site. The traffic congestion at this site was intensified by vehicles decelerating and accelerating near the Gatambe temple.
The third highest PM 10 concentration of 220 ug/m 3 was recorded at Site 4.1, and the traffic intensity at this site was 1260 vehicles/h. This site was located by the Kadugannawa bend along the Kandy-Colombo national highway (Al). Vehicles travelling towards Kandy have to climb up a sloping road while making a 300° angle turn at this bend. This caused the vehicles to burn more fuel at this site even though there was no traffic congestion. A number of vehicles let out black smoke upon acceleration. Figure 1 shows that the PM 10 concentrations, in general, were high along the national highways of the urban and near-urban environments of Kandy city (Cluster 1) and Peradeniya (Cluster 3).
Sites 2.1, 2.2 and 2.3 of Cluster 2 were all located within the Kandy city limits and had PM 10 concentrations in the range of 25 to 40 ug/m 3 . It was probably because the sampler was placed within the school premises away from the roads with high traffic. The measurements at these three sites which are away from high traffic roads may be taken as background PM 10 concentrations. Site 2.4 of Cluster 2 however, gave a comparatively high PM )0 concentration. It was perhaps due to the increased emissions stemming from the traffic that slowed down to take the sharp bend on the road adjacent to the school.  Gatamhe temple (Site 3.1), which is 100 m away from the Botanical Garden entrance, respectively. This is a good example of the impact of greenery on air quality.     Combining the use of a HVAS with a PAS could reduce the cost of assessing the PM pollution levels and thereby enable the extension of the PM pollution study to other major cities in Sri Lanka.