108

4,900 千元

空氣品質監測

2019/01/22

2019/12/31

鄭芳怡

監資處

陳亦絢

國立中央大學

大氣空氣

臭氧，空氣品質預報，天氣型態分類，預報誤差校正

O3, Air quality forecasting, weather pattern classification, bias correction

馮智勇

邱玉琴

03-4227151#65556

uamn0209@mail2000.com.tw

本計畫執行重點包含：(1)分析近3至5年臺灣各空品區發生高臭氧濃度事件，含O3 8小時(O3-8hr)，以及中、長程傳輸時之天氣系統及氣象時空變化特性; (2)分析近5至10年颱風路徑與污染物濃度（PM2.5、PM10、O3）時空分布及變化的關係; (3) 使用天氣型態分類結果與污染物濃度關係，建立預報參考數據或指標; (4) 使用空氣品質模式預報結果，建立天氣型態分類，測試其應用於預報系統之可行性。 O3濃度長期與近五年資料分析，顯示O3平均值濃度為長期增加趨勢，但在前5 %的高O3濃度，在中南部地區於2007年之後呈現遞減趨勢。2018-2019空污季節，臺灣整體O3濃度較前年增加，推測受到高溫影響、同時雨量也比前年度少，加上靜風天數增加。 透過數個高O3污染個案模擬（包含境內與境外污染事件），了解臺灣地區O3與其前驅物(NOx和VOC)之間的反應。境內污染事件，顯示西半部多數地區屬於NOx飽和區域。 空品模式預報校正方法，使用了天氣型態分類重新定義預報誤差，再行建立線性迴歸方程。此新方法，反應出模式在氣象場、排放量、大氣化學模式本身的系統性誤差特性，也因此能有效降低預報誤差，提升各地PM2.5濃度預報度。

臭氧，空氣品質預報，天氣型態分類，預報誤差校正

正式報告_定稿(全).pdf

38MB

2020/2/1 上午 12:00:00

The objective of this study includes: (1) investigate the O3 problem in Taiwan through the long-term and short-term data analysis; (2) analyze the impact of the typhoon on Taiwan’s air pollution problem; (3) develop the bias-correction method based on the weather pattern dependent method; (4) investigate the performance of the new bias-correction method. The analysis of the O3 concentration from air quality monitoring stations appears a long-term increasing trend. For the top 5 percentiles of the high O3 concentrations, there appears a decreasing trend since 2007. The most recent air pollution season (2018/10-2019/04) shows higher O3 concentration comparted to that of the previous air pollution season (2017/10-2018/04). This might be related to the higher temperature, less rainfall and lower wind speed. The analysis of the several high O3 events indicates that the O3 concentrations in the air are mostly in the NOx saturated regime in western Taiwan. The new bias correction method is developed by categorizing the bias into different weather patterns. The bias characteristics represent the error from the meteorological model, emissions and the air quality models. The new method is able to enhance the air quality forecasting capability.

O3, Air quality forecasting, weather pattern classification, bias correction