細懸浮微粒(fine particulate matter，PM2.5)污染物對於人體健康影響嚴重，已成為各國監測與管控的重要空氣污染物。本研究旨在定義高高屏地區的典型 PM2.5污染事件，探究其區域氣象條件的日循環變化，以及伴隨的東亞大尺度環流條件。結果顯示，典型的事件約持續 36 小時，且在「當日」主污染期於下午 6 點結束前，有一從「前一日」清晨 6 點開始的前兆期。區域平均 PM2.5 濃度變化呈現半日波的震盪：11 時和凌晨 1 時最高，18 時和 06 時較低。NOx（nitrogen oxides，氮氧化物）、CO（carbon monoxide，一氧化碳）、THC（total hydrocarbons，總碳氫化合物）、以及 NMHC（non-methane hydrocarbons，非甲烷碳氫化合物）等源自交通及工業活動的汚染物，其濃度隨時間的變化，領先 PM2.5 約 3 至 4 小時。然而，NOx 和 CO 的事件平均濃度與其氣候平均值之間並無明顯差異；而屬於光化學反應前趨物質的 THC和 NMHC 以及較易溶於水的 SO2（sulfur dioxide，二氧化硫）的事件平均濃度反而低於其氣候平均值。
事件發生時普遍增温，白天臨海地區出現南風距平，入夜後山區出現東風距平，以及低對流層的沈降氣流等氣象條件變化，導致 PM2.5 滯留在邊界層內的時間拉長。溫度與 PM2.5 濃度變化在白天（晚上）大致呈現負（正）相關。日循環的溫度變化與因人類活動而呈現的 PM2.5半日波變化彼此互相影響。我們也發現伴隨污染事件發生時，東亞大尺度環流變化主要有高壓出海，寒潮暴發以及鋒面過境等 3 種型態。這些在冷季被模式高度預測的綜觀尺度波動，提供了針對 PM2.5 污染事件進行展期預報的基礎，縱使就區域觀點而言，它們都頗為符合所謂的高壓迴流天氣。

This study aims at identifying the typical PM2.5 contamination events over the Kaohsiung-Pingtung areas and detecting the associated diurnal variations of regional meteorological conditions and large-scale circulation settings over East Asia. The results show that the typical event lasts about 36 hours, and there exists a precursor period initiating at 6 a.m. during the previous day before the principal period ending at 6 p.m. of the current day. The temporal evolution of area-averaged PM2.5 concentration shows a semidiurnal oscillation by peaking at 1 a.m. and 11 a.m., but with two dips occurring at 6 a.m. and 6 p.m. The accompanying changes of NOx, CO, THC, and NMHC, which are mainly generated by traffic and industrial activities, lead PM2.5 3-to-4 hours. The event-averaged concentration of NOx as well as CO shows no significant difference compared to the climatic mean value. In contrast, the THC and NMHC (precursors of the photochemical reaction) and the hygroscopic SO2 show significant lower event-averaged concentrations.
When a PM2.5 event occurs, changes of meteorological conditions including the wide-spread rising temperature, daytime southerly wind anomalies near the coasts, nighttime easterly wind anomalies near the mountains, and the sinking air in the lower troposphere as a whole, prolong its duration time inside the boundary layer. Changes of temperature and PM2.5 concentration, qualitatively, are negative- (positive-) correlated during the daytime (nighttime). It is found that the diurnal nature of temperature evolution likely interacts with the human-induced semidiurnal signal of PM2.5. The associated changes in the large-scale circulation over East Asia have three major patterns: the off-shored high pressure, cold surge, and frontal passage. The identified patterns are well predicted by the current weather model and thereby providing the basis for the extended forecast on the occurrence of PM2.5 event, though they all satisfy the so-called high pressure reflux weather within the regional perspective.