空氣汙染對於台灣北部午後對流影響之分析與模擬研究

An analysis and modeling study on the impacts of air pollution to afternoon convection in northern Taiwan

楊謹瑜(Yang Jin-Yu)

指導教授 ： 王重傑

國立臺灣師範大學/理學院/地球科學系/碩士(2022年)

https://doi.org/10.6345/NTNU202200646

摘要

人為活動產生的污染可能是造成極端天氣現象加劇的因子之一，而風向、洋流和降雨量的變動都會導致全球各地天氣型態的改變，其中又以人為排放的溫室效應氣體與懸浮微粒為最主要因子，因此懸浮微粒不僅會對人類健康造成危害，也會對環境產生影響。

其中空氣中的懸浮微粒(如空氣汙染物：PM2.5 以及 PM10)會透過輻射作用以及雲微物理效應來改變雲的生命週期抑或是降水分布的變化，但迄今為止對於空氣汙染物增加會導致降水如何改變的估計仍有非常大的不確定性。因此本研究藉由中央氣象局的地面觀測資料與衛星資料來挑選出 2016 年至 2019 年台灣雨季：6 月到 9 月的午後對流個案與無降雨個案，接著再整合行政院環保署的空氣品質監測數據來計算個案期間降水量與空氣汙染物的相關性分析。深入研究不同時間段、不同濃度對於日累積降水量、降水強度與降水總時長的相關係數。最後使用日本名古屋大學所發展的雲解析風暴模式(Cloud-Resolving Storm simulator，以下簡稱為 CReSS)來進行不同粒子濃度的模擬。

分析結果顯示，PM2.5 在午後對流個案的空污粒子當日逐時之分布較為收斂，且粒子濃度的最大值落在當地時間中午 12 點到 13 點之間，整體分布為鐘形曲線，降水前會累積到最大值，降水後則有快速下降的趨勢，而 PM10 因為粒子粒徑尺寸大，被雨水沖刷的程度也更大，因此有更顯著的下降趨勢；無降雨個案除了分布較為發散之外，粒子濃度的最大值落在當地時間下午13 點到 14點之間，達到最大值濃度的時間較午後對流個案晚 1 個小時，其不同個案之間在凌晨與夜晚兩個極端值會有更大的差異性，顯示出午後對流個案比起無降雨個案，懸浮微粒的變化更有規律性。

PM2.5 以及 PM10 對於降水量的相關性檢定統計結果有三個重要發現：(一)上午 0000-1000 LST 空污粒子濃度最小值，會影響午後對流的降水強度，兩者的相關係數為 0.48。(二)1100-1500 LST 空污粒子濃度最大值，則會影響午後對流當日降水總時長，兩者的相關係數為 0.473。(三)最小值與最大值之間的升幅變化，會影響午後對流的降水總時長，兩者的相關係數為 0.564。

最後 CReSS 模擬結果與上述統計結果的第一點及第二點相符，整體而言，本研究能夠透過統計與模擬的方式來知悉兩個空氣汙染物與雲微物理作用的重要機制：當雲凝結顆粒數愈少時，微粒能夠抑制毛毛雨現象的程度愈明顯；當顆粒數愈多時，微粒能夠使雲的生命週期愈持久。

關鍵字

空氣汙染 ； 懸浮微粒 ； PM2.5 ； PM10 ； 午後對流 ； CReSS

並列摘要

Air pollution from human activities might be one of the factors that exacerbate extreme weather phenomena, and changes in wind direction, ocean currents and rainfall also affected in weather patterns around the world. Among them, greenhouse gases and suspended particulates that were made from human beings were the main elements.

Particle materials in the air (such as air pollutants: PM2.5 and PM10) will change the life cycle of clouds or changes in precipitation distribution through radiation and cloud microphysical effects. However, how precipitation changed was still uncertain. Therefore, this study uses the ground observation data and satellite data of the Central Meteorological Administration to select the afternoon convection cases and no rainfall cases from 2016 to 2019 in the northern Taiwan.

Using the monitoring data from the Environmental Protection Agency of the Executive Yuan to calculate the correlation analysis of precipitation and air pollutants during the case period. And then, the Cloud-Resolving Storm simulator (CReSS) developed by Nagoya University in Japan was used to simulate different particle concentrations.

The analysis results show that the hourly data of PM2.5 in the afternoon convection cases were relatively convergent, and the maximum concentration was at 12:00 and 13:00 local time. The overall distribution was a bell-shaped curve. It will accumulate to the maximum value before raining, but decrease rapidly after the precipitation, however, because of the large particle size, PM10 was also washed out by the rain vigorously, so its overall distribution was more significant downward trend than others.The distribution of no rainfall cases was divergent. In addition, the maximum concentration time was at 13:00 and 14:00 local time, it was 1 hour later than afternoon convection cases. The results above shows that change of aerosols was more regular in the afternoon convection cases than in the no rainfall cases.

There are three important findings in the statistical results of the correlation test of PM2.5 and PM10 in precipitation: (1) The minimum concentration of air pollution between 0000 and 1000 LST in the morning might affect the precipitation intensity of afternoon convection, and its correlation coefficient was 0.48. (2) The maximum concentration of air pollution between 1100 and 1500 LST might affect the total duration of precipitation on the day of afternoon convection, and its correlation coefficient was 0.473. (3) The change in the amplitude from the minimum value to the maximum value might affect the total duration of afternoon convective precipitation, and its correlation coefficient was 0.564.

Finally, the CReSS simulation results were consistent with the statistical ones. Overall, this study could understand the important mechanisms of the interaction between air pollutants and cloud microphysics through statistics and simulations: when the number of cloud condensation particles is smaller, the particles can inhibit the the more pronounced the drizzle phenomenon is; the greater the number of particles, the longer the particles can make the cloud's life cycle.

並列關鍵字

air pollution ； PM2.5 ； PM10 ； afternoon convection ； CReSS

參考文獻

王秋森、陳時欣、鄭曼婷、張艮輝(2016)。氣溶膠技術學(第三版)。新北市：大揚。

陳泰然、王重傑、楊進賢(2002)。台灣梅雨季對流降水之時空分佈特徵。大氣科學，30，83-96。

林品芳、張保亮、周仲島(2012)。弱綜觀環境下台灣午後對流特徵及其客觀預報。大氣科學，40，77-107。

Akter, N. and Tsuboki, K. (2012). Numerical simulation of cyclone Sidr using a cloud-resolving model: characteristics and formation process of an outer rainband. Monthly Weather Review, 140, 789–810

Andreae, M. O. (2007).Aerosols Before Pollution, Science, 315, 50-51