| Abstract: | 本文使用環保署十個空氣品質監測站埔里(參考站)、線西(工業站)、三義(背景)、永和(交通站)、馬祖(一般站)、大里(一般站)、西屯(一般站)、沙鹿(一般)、忠明(一般站)、豐原(一般站)，在2015-2018年每小時資料，進行空氣品質分析。首先使用Catmull-Rom Spline和Linear法，進行遺漏值之插補；其次，利用聯立分量迴歸模型分析，分別使用0.1、0.5、0.9分位，進行PM2.5影響因子分析。實證結果發現:在PM2.5汙染嚴重情況之0.9分位時: (1)化學因子方面:PM2.5(-1)、PM10、O3、CO，在所有監測站中，皆會惡化PM2.5； (2)氣象因子方面:AT在線西、永和、馬祖、大里、西屯、沙鹿和忠明呈現U字型，現階段會增加PM2.5；RH在永和、大里、忠明和豐原呈現倒U字型，現階段會降低PM2.5；WS只在線西、永和呈現U字型，現階段會增加PM2.5；(3)東北季風虛擬變方面，永和、馬祖、大里、忠明的東北季風，皆會惡化PM2.5。
This study use The Environmental Protection Administration(EPA) Ten Air Quality mornitoring Station, Puli(reference), Xianxi(industry), Sanyi(background), Yonghe(traffic), Matsu(general), Dali(general), Xitun(general), Shalu(general), Zhongming (general) and Fengyuan(general),hourly data between 2015 and 2018 to analysis air quality. First, using Catmull-Rom Spline and Linear method, interpolate the missing data；then using Simultaneous Quantile Regression model with 0.1, 0.5, 0.9 quantile，PM2.5 impact factor analysis. Empirical study discovered at 0.9 quantile of serious PM2.5 pollution: (1) Chemical factors: PM2.5 (-1), PM10, O3, CO, in all monitoring stations, will worsen PM2.5 (2) meteorological factors: at Xianxi, Yonghe, Matsu, Dali, Xitun, Shalu and Zhongming present U-shaped, and will increase PM2.5 at this stage; RH will have inverted U-shaped in Yonghe, Dali, Zhongming and Fengyuan, and will reduce PM2.5 at this stage; WS will only display U-shaped on the Xianxi and Yonghe. At this stage, PM2.5 will be increased; (3) North east monsoon deteriorate PM2.5 in Yonghe, Matsu, Dali and Zhongming. |
