| 摘要: | 本文於2019年秋季(10月)在南投縣鹿林山(海拔2,862公尺)以及2020年春季(3至5月)在鹿林山和台中市中山醫學大學(海拔50公尺)分別採集氣動直徑為1.0 μm、2.5 μm、10 μm (PM1、PM2.5、PM10)大氣氣膠，並解析氣膠化學成分和光學特性。
2019年鹿林山秋季PM2.5和 PM10質量濃度分別為 7 ± 3 和9 ± 3 μg m-3，2020年春季PM1、PM2.5和 PM10質量濃度則分別為 18 ± 9、21 ± 10、24 ± 11 μg m-3，兩季氣膠粒徑分布都偏向由細粒徑主導。春季氣流軌跡分類以生質燃燒類型(BB)發生頻率最高，PM2.5 總碳成分與K+的線性相關性高(R2 =0.93; n=25, p < 0.01)，顯示受到生質燃燒傳輸影響；弱生質燃燒人為排放類型(WBB-AN)的SO42-濃度比秋季自由大氣類型(F-FT)高0.8 μg m-3，可視為境外傳輸人為排放增多的貢獻量。分析2003-2020年觀測數據，春季最高PM2.5濃度出現在2004年，BB類型的人為排放影響有逐年上升趨勢。
中山醫PM1、PM2.5、PM10的質量濃度分別為23 ± 9、31 ± 11、47 ± 16 μg m-3，中山醫細粒徑顆粒也是主導氣膠質量濃度，但不及鹿林山顯著。中山醫PM2.5低濃度(CSMU-LC)主要污染源是工業活動，工業活動貢獻量隨粒徑變小(PM1)而增加；相對地，交通活動是中山醫高濃度(CSMU-HC) PM2.5主要污染源，交通活動貢獻隨粒徑變大(PM10)而增加，高濃度的形成主要受擴散不佳及發生光化學反應影響。
以Revised IMPROVE方程式估算鹿林山及中山醫氣膠化學成分貢獻的大氣消光係數(bext)與儀器測量結果相關性都很好(R2 =0.84, n=20, p < 0.01; R2 =0.80, n=17, p<0.01)，顯示氣膠化學成分與bext有密切關係。鹿林山秋季bext由硫酸銨(AS)和空氣分子(RS)共同主導，春季則由有機物(OM)主導。中山醫bext主要由AS和硝酸銨(AN)貢獻，CSMU-HC bext高於CSMU-LC主要原因為AN占比提升。吸收光徑370 nm和880 nm黑碳濃度差異提供另類方法推測受生質燃燒影響，所推估出鹿林山受到較強烈生質燃燒日期，與氣流軌跡分類日期相近。鹿林山氣膠光學厚度(AOD)受PM2.5中AS及元素碳影響較大，且鹿林山高AOD採樣日通常伴隨較細的粒徑分布。
綜合而言，歷年數據分析顯示春季最高PM2.5濃度出現在2004年，2020年春季鹿林山仍受到東南亞生質燃燒傳輸影響，秋季bext由AS和RS共同主導，春季則由OM主導。中山醫採樣期間bext主要由AS和AN貢獻，CSMU-LC受工業源影響大，CSMU-HC則是受交通源、擴散不佳、光化學反應影響。
;This study collected PM1, PM2.5, and PM10 (particulate matter with an aerodynamic less than or equal to 1, 2.5, and 10 μm) in fall (October) of 2019 and spring (March to May) of 2020 at Mt. Lulin (2862 m above sea level). PM1, PM2.5, and PM10 were also collected at Chung Shan Medical University (CSMU) (50 m a.s.l.) in the spring of 2020. The chemical composition was resolved, and optical properties were computed from the collected PM.
The mass concentrations of PM2.5 and PM10 at Mt. Lulin in fall of 2019 were 7 ± 3 and 9 ± 3 μg m-3, respectively, and the mass concentrations of PM1, PM2.5, and PM10 in spring of 2020 were 18 ± 9, 21 ± 10, and 24 ± 11 μg m-3, respectively. Fine PM both dominated the particle size distribution of the two seasons. In spring, the biomass burning (BB) type occurred highest in the airmass trajectory classification. High linear correlation (R2 =0.93; n=25, p <0.01) between the total carbon content and K+ of PM2.5 indicated the influence of transported BB smoke. In contrast, the SO42- concentration of the weak BB anthropogenic type (WBB-AN) was 0.8 μg m-3 more than that of the free troposphere type (F-FT) in fall, which could be counted as an additional contribution from anthropogenic emissions by transboundary transport. The highest spring PM2.5 concentration appeared in 2004, and the impact from anthropogenic emissions of BB type had an upward trend year by year from 2003 to 2020.
The mass concentrations of PM1, PM2.5, and PM10 at CSMU were 23 ± 9, 31 ± 11, and 47 ± 16 μg m-3, respectively. The fine particle size at CSMU also dominated the aerosol mass concentration but was not as significant as that of Mt. Lulin. The main PM2.5 pollution source of CSMU-LC was industrial activity, which contributed more with reduced particle size (PM1). In contrast, traffic activity was the primary pollution source of CSMU-HC, and the contribution of traffic activity increased as the particle size became larger (PM10). The formation of high concentration events was influenced by poor ventilation and photochemical reactions.
The atmospheric extinction coefficient (bext) contributed by the aerosol chemical components, estimated by using the Revised IMPROVE equation, correlated well (R2 = 0.84, n = 20, p <0.01; R2 = 0.80, n = 17, p <0.01) with measurements both at Mt. Lulin and CSMU. The results indicate that aerosol chemical composition is closely related to bext. The bext was jointly dominated by ammonium sulfate (AS) and Rayleigh scattering (RS) at Mt. Lulin in fall, while it was led by organic matter (OM) in spring. The bext at CSMU was mainly contributed by AS and ammonium nitrate (AN). The higher value of bext at CSMU-HC than that of CSMU-LC is because of the greater proportion of AN. The difference in black carbon concentrations between 370 and 880 nm provided an alternative method for inferring dates influenced by intense BB because the dates were similar to those of the air trajectory classification method. The aerosol optical depth (AOD) at Mt. Lulin was greatly affected by the AS and elemental carbon (EC) in PM2.5, and finer particle size spectra usually accompanied high AOD sampling days.
In summary, the highest PM2.5 concentration at the Mt. Lulin site in spring occurred in 2004 from historical data. Mt. Lulin is still under the influence of the transported BB from Southeast Asia in spring 2020. The bext was dominated by AS and RS in fall and OM in spring in 2020. During the CSMU sampling period, the bext was primarily contributed by AS and AN with industrial activity as the main pollution source of CSMU-LC in contrast to traffic activity at CSMU-HC plus poor ventilation and photochemical reactions. |
