| 摘要: | 本文以正矩陣因子法(Positive Matrix Factorization, PMF)及條件機率函數(Conditional Probability Function, CPF)推估2017年「細懸浮微粒(PM2.5)化學成分監測及分析計畫」(以下稱為「2017年計畫」) 6個測站PM2.5污染來源，並以潛在源貢獻因子法(Potential Source Contribution Function, PSCF)探討每個測站硫酸鹽長程傳輸軌跡起源。針對「2017年計畫」在PM2.5高濃度(≧35 μg m-3)時，硝酸根離子在PM2.5占比提高及台灣西部測站有時出現的凌晨PM2.5高濃度現象，本文彙整上述發現的各種污染物濃度及環境條件。此外，本文也推估原生有機碳和二次有機碳濃度以瞭解彼此濃度分布。
污染源推估結果顯示板橋站以F3 (二次硫酸鹽與工業鍋爐)為最重要污染源，推論受樹林等工業區影響。忠明站最重要污染源為F4 (二次硝酸鹽)，推測與國道1號等交通排放有關。斗六站也是以F3 (二次硝酸鹽)為最重要污染源，可能受到雲林科技園區與國道3號擴散不佳影響。嘉義站最重要污染源為F1 (二次硝酸鹽)，從污染來向推論受嘉太工業區和交通排放影響。F3 (二次硝酸鹽與揚塵)是小港站最重要污染源，推論受大發、臨海工業區等影響。具體來說，花蓮站是以F4 (二次硫酸鹽與工業鍋爐)為最重要污染源，推論與海上船舶及美崙工業區有關。
以PSCF追溯各測站(二次硫酸鹽)因子軌跡起源，僅有板橋站來自大陸山東及江蘇省，其他測站多以本地污染源為主。忠明站傳輸污染與台中焚化廠、台中工業區與彰濱工業區等有關；斗六站與斗六工業區等污染源混合影響有關。嘉義站可能受台南南部科學工業園區（南科）影響；小港站則受仁武垃圾焚化廠、大發工業區、中國鋼鐵、林園工業區等混合影響；花蓮站在東北方海面上有污染貢獻，推論受船舶影響。
當台灣西部發生高濃度PM2.5，NO3-呈現高濃度占比，主要與風速低造成的污染物累積與前日污染殘留影響有關。氣膠水溶性離子成分使用ISORROPIA模式進行模擬，指出以(NH4)2SO4、NH4NO3結合型態為主，顯示有氨氣充足現象；這可由台灣北至南部5個測站過剩NH4+與NO3-莫耳濃度有很好相關性(R2) 0.79、0.89、0.89、0.89、0.90得到驗證。至於偏離過剩NH4+與NO3-線性關係較大的離群值，推測可能與夜間高硝酸根離子和NO2濃度的N2O5水解等有關。顯然地，凌晨PM2.5高濃度現象受到前一天氣體污染物濃度影響，其中，以NOX濃度較高，SO2濃度雖然較低，但和前一天濃度差異百分比也不小，NOX和SO2管制具有相同重要性。
透過EC-tracer方法推估板橋站的二次有機碳占OC約32-38%、忠明站約23-41%、斗六站約44-56%、嘉義站約30-47%、小港站約30-57%、花蓮站約28-46%，顯示台灣地區PM2.5 有機碳組成以原生有機碳為主，來源主要是交通活動相關排放與生質燃燒。
綜合而言，除了板橋站有部分SO42-來自中國北方外，台灣西部地區PM2.5以本地污染源貢獻為主，特別是NO3-。高濃度PM2.5事件的高NO3-濃度占比與風速低及較高NOX有關。至於夜間高濃度NO3-可能與高NO2和較高環境相對濕度導致N2O5水解有關。
;This study used Positive Matrix Factorization (PMF) and Conditional Probability Function (CPF) to identify PM2.5 source contributions of the six stations in the “PM2.5 chemical composition monitoring and analysis study” in 2017 (henceforth referred to “2017 Study”). In addition, the Potential Source Contribution Function (PSCF) was applied to investigate the trajectory origins of sulfates through long-range transport to each site. Given the proportion of PM2.5 nitrate ion was higher in a high PM2.5 event (≧35 μg m-3) and frequent occurrence of high PM2.5 in the western area of Taiwan in the midnight in “2017 Study”, this study summarized the pollutant concentrations and environmental conditions associated with the above findings. Moreover, this study estimated primary and secondary organic carbon concentrations to understand the relative distribution of each other.
The source apportionment results showed that F3 (secondary sulfate and industrial boiler) was the main source factor under the influence of the Shulin Industrial Zone at the Banqiao station. For the Zhongming station, F4 (secondary nitrate) was the most important source factor caused by Freeway 1 and other traffic emissions. F3 (secondary nitrate) was the most significant source factor affected by Yunlin Technology Industrial Park and Freeway 3 during poor ventilation of the Douliu station. Similarly, F1 (secondary nitrate) was the source factor contributed the most and was related to Jiatai Industrial Zone plus intensive traffic emissions at the Chiayi station. For the Xiaogang station, F3 (secondary nitrate and soil dust) was the major source factor under the influences of the Dafa and Linhai Industrial Zones. Specifically, ship emissions from the sea and Meilun Industrial Zone accounted for the worst source factor of F4 (secondary sulfate and Industrial boiler) at the Hualien Station.
PSCF modeling for the trajectory origins of “secondary sulfate” of various stations only to find that the Banqiao Station was under the influence of Shandong and Jiangsu Provinces in China. In contrast, local pollution sources affected other stations. For example, Taichung Incineration Plant, as well as Taichung and Changbin Industrial Zones affected the Zhongming Station. The mixed sources including Douliu Industrial Zone influenced the Douliu Station. The Southern Taiwan Science Park might have affected the Chiayi Station. Moreover, a mix of sources affected the Xiaogang Station such as Renwu Refuse Incineration Plant, Dafa Industrial Zone, China Steel Corporation, and Linyuan Industrial Zone. The ship emissions in the northeastern direction to the Hualien station might have caused the pollution.
The high NO3- proportion in high PM2.5 concentration events were mainly under the influences of pollutant accumulation due to low wind speed and the pollutant left-overs from the previous day. ISORROPIA modeling indicated the abundance of NH4+ such that the main compound forms were (NH4)2SO4 and NH4NO3 of the water-soluble inorganic ions. This was evidenced by the excellent correlations (R2) between excess molar concentrations of NH4+ and NO3- were 0.79, 0.89, 0.89, 0.89, and 0.90 in the five stations from north to south in Taiwan. For the outliers that deviate from the linear relationship between excess NH4+ and NO3- might be related to the occurrence of N2O5 hydrolysis under high NO3- and NO2 concentrations at night. Apparently, the pollutant concentrations of the previous days affected the midnight high PM2.5 concentrations. Among various gas pollutants, NOX concentrations were the highest with great deviations with that of the previous day; however, the control of NOX and SO2 were of equal importance as SO2 was also with great deviations with that of the previous days.
The estimates from the EC-Tracer method showed that the proportions of secondary organic carbon over organic carbon (OC) were 32-38% at the Banqiao station, 23-41% at the Zhongming station, 44-56% at the Douliu station, 30-47% at the Chiayi station, 30-47% at the Xiaogang station, and 28-46% at the Hualien station, respectively. It implies that primary OC is the major OC in Taiwan. The sources are mainly associated with emissions from transportation activities and biomass burning.
In summary, local pollution sources are mainly responsible for PM2.5 concentration especially for NO3- in the western area of Taiwan except part of SO42- transported from northern China. Low wind speed and higher NOX were associated with the high proportion of NO3- in high PM2.5 concentration events. As for the high concentration of NO3- at night, it might be related to N2O5 hydrolysis under high NO2 and relative humidity in the environment. |
