Assessment of potential long-range transport of particulate air pollution using trajectory modeling and monitoring data

Abstract

Quantification of the long-range transport (LRT) contribution to ambient air pollution levels at a location is a challenging task and is normally done with a high uncertainty. In the lack of accurate emission data over the large regional domain for dispersion modeling, this study attempts to use both trajectory analysis and monitoring data to assess the potential contribution of LRT to particulate air pollution (PM) in the Bangkok Metropolitan Region (BMR). The 10-day backward trajectories of air masses arriving at BMR from January 2002 to December 2004 were determined using Hybrid Single-Particle Langrangian Integrated Trajectory model version 4 (HYSPLIT4) and were categorized by k-means clustering into 6 clusters. Subsequently, PM levels in the BMR associated with each air mass cluster during this period were analyzed. Clusters 1 and 6 were observed with the highest and 2nd highest average PM 10 and PM 2.5 levels in the BMR, respectively, which commonly have a longer air mass pathway over populated South East Asia (SEA). The third highest PM levels were associated with air masses from the east (clusters 2 and 5), which enter the BMR via the Gulf of Thailand without passing the SEA regions. The other two clusters (3 and 4) are characterized with a long pathway of air masses over the Indian Ocean and the lowest PM levels. High PM days, which are defined based on the spatial coverage of high PM levels in the BMR, were identified and analyzed for the possibility of long-range transport contribution of PM. The potential source contribution function (PSCF) and air mass trajectories show that on high PM day, the air masses commonly originated and passed over populated regions before arriving at the BMR, which suggests a possible LRT contribution. Considerations are made for surface ozone, SO 4 2−/SO 2 and average SO 2, PM 2.5/PM 10, and weekday–weekend traffic emission within each air mass trajectory cluster to reveal the possible LRT contribution.