In this study, the cluster analysis method is applied to the daily averaged wind fields and sea level pressure observed from surface weather stations in Taiwan to classify the synoptic weather pattern and study the corresponding characteristics of air pollutants, including fine particulate matter (PM2.5), coarse particulate matter (PM10), and ozone (O3), in Taiwan. The study period is from January 2013 to March 2018. The classification identified six weather types: clusters 1, 2 and 3 (C1-C3), which are typical winter weather types and associated with high air pollutant concentrations; C3, which is under the influence of weak synoptic weather, associated with the lowest wind speed and the highest PM2.5 and PM10 concentrations, and represents the most prevalent weather type that is prone to the occurrence of PM2.5 events; C4, which occurs mostly during seasonal transition months and is associated with the highest O3 concentrations; and C5 and C6, which are summer weather types with low air pollutant concentrations.
Further analysis of the local wind flow in the Taiwan area using the 0.3-degree ERA5 reanalysis dataset and surface-observed wind indicates that the land-sea breeze embedded within the synoptic weather forms in western Taiwan in the C3 cluster, which is favorable for air pollutant accumulation. On the other hand, when the prevailing northeasterly wind is obstructed by the Central Mountain Range, the southwestern Taiwan being situated on the leeside of the mountain, often exhibits the worst air pollution problem due to the stagnant wind condition.