| 摘要: | 大氣氣膠對環境的衝擊及人體健康的效應，已經引起重視，並投入大量的經費與人力進行研究。但對於氣膠特性的解析仍然多憑藉人工採樣，使用濾紙採樣然後分析氣膠特性，僅可提供較粗略的時間與空間的解析度，而且對於採樣誤差不容易量化，因此提供精確且能避免採樣運送及儲存誤差的即時(Real-Time)自動連續量測是很重要的，台灣超級測站的建置即是希望能提供達到上述目的所需的監測結果。在台灣地區由於大氣溼度較高，考量大氣氣膠在高濕環境朝解其含水量對於質量濃度量測、酸性沉降、光學性質、全球氣候變遷與人體健康皆具有顯著的影響，因此估算大氣氣膠含水量有其必要性，目前含水量推估模式雖然已經發展出來，但都是以實驗室產生的氣膠為基礎，對於大氣氣膠含水量的估算仍然有待評估，而且目前對於大氣氣膠含水量直接量測得數據仍然相當有限，有鑑於此，本文使用一套能量測收集在濾紙上氣膠含水量量測系統 (Chang and Lee, 2002; Lee and Chang, 2002)，進行量測與模式模擬的比較，期望能初步探窺大氣氣膠含水量的面貌。 研究結果顯示出台北地區2002年春季PM2.5的平均濃度為37.2mg/m3，PM10的平均濃度為54.9mg/m3。有機碳成份的平均濃度為7.6mg/m3，元素碳成份的平均濃度為2.6mg/m3，黑碳的平均濃度為3.7mg/m3。硫酸鹽的平均濃度為7.1mg/m3，硝酸鹽的平均濃度為3.6mg/m3，若不考量黃沙時期的影響，3到5月各月月平均濃度變化不大。 在量測期間氣膠特性的變動顯示：氣膠質量濃度及主要化學物種濃度的隨時間變化濃度受上下班（學）的影響大，工作日及假日型態有明顯差異。 風速影響污染物的稀釋與擴散，本文發現：除了PM2.5-10濃度與風速成正比外，其他監測項目皆成反比。在高相對濕度，PM10由粗細粒徑氣膠共存，轉變為以細粒徑氣膠為主。從二次有機碳估算結果顯示細粒徑氣膠碳成份主要為二次有機碳，約佔總碳的56%，其次為元素碳佔25%，一次有機碳佔19%。本文以「近污染源氣膠化學性質」推估PM2.5污染來源 (Lee and Hsu, 1996)，結果顯示二次反應所佔百分比高於交通活動，二種來源加總可佔細粒徑氣膠約35%至95%不等。 對於氣膠含水量特性的探討，本文獲得具體成果如下：從含水量與各監測項目相關係數矩陣來看，含水量主要受相對濕度及氣膠成份的影響。從大氣氣膠含水量歷程分析來看，若將乾微粒直接增濕到當時大氣相對溼度，所量測到的含水量是低估的，因為沒有考慮到氣膠在大氣中的歷程。但若從最高溼度降濕到當時大氣相對溼度，所測得的微粒含水量，則有高估的可能。因此，必須考量微粒在增濕前的大氣相對溼度所保有的含水量，才能量測到正確的含水量。若氣膠已經增濕到比較屬於高濕度的階段，則三種方式量測到的水量彼此相當接近。 本文量測的氣膠含水量與模式推估值有良好的相關性，判定係數R2高達0.86，但模式值低估許多，造成差異的原因主要有：模式值未納入吸濕性有機物、使用水溶性離子未涵蓋大氣中易吸濕的所有無機性鹽類。當秤重溫度維持在22?27℃間，濕度控制在26?34%之間，乾氣膠成份中仍帶有3%至29%的水量。在回復採樣時大氣相對溼度氣膠含水量顯示，氣膠含水量佔細粒徑氣膠百分比平均為64%，最大為86%，最低也有30%。 The environmental impact and health effects of atmospheric aerosols have drawn much attention. From time to time, people invest a great deal of money and efforts to understand the role of aerosols. Until recently, filter based sampling and laboratory analysis is still the mainstream to resolve aerosol properties. This could only provide a rough temporal and spatial resolution of aerosol chemical contents and is hard for quantification of sampling artifacts. The real-time continuous measurement for aerosol properties could provide the solution of acquiring precise results without incurring transportation and storge errors like filter samples. The installation of Taiwan supersite is to achieve the aforementioned objectives. Owing to the high atmospheric RH, it is necessary to assess liquid water content (LWC) of aerosols for their effects on mass measurement, acid deposition, light attenuation, global change, and health effects. Although there are several thermal equilibrium models to estimate aerosol LWC, however, they are all built by the LWC of laboratory generated known aerosols. These models need more data from aerosols in real atmosphere to test. To date, the data of aerosol LWC are scarce; this study adopted a developed measurement system to detect LWC from collected filters (Chang and Lee, 2002; Lee and Chang, 2002). Measurement results and model estimates are compared in this study to understand more on aerosol LWC in the real atmosphere. The results of aerosol properties from continuous monitoring in the springtime of urban Taipei show as follows. PM2.5 and PM10 is with an average of 37.2mg/m3 and 54.9mg/m3, respectively. The average of organic carbon is higher at 7.6mg/m3 compared with 2.6mg/m3 for elemental carbon and 3.7mg/m3 for black carbon. The measured average for sulfates is 7.1mg/m3 and that of 3.6mg/m3 is for nitrates. For all these aerosol properties, excluding the data during yellow sand periods, the monthly averages vary very little from March to May. The time history of aerosol mass and major chemical species are influenced by the time shift of on and off duty. Concentration pattern in working days and holidays is deviated from each other significantly. Wind speed has been expected to influence the dilution and dispersion of pollutants. This study agrees with the above inference by showing a linear relationship between PM2.5-10 and wind speed and a reverse relationship between wind speed and other measurements. At high relative humidity (RH), the predominant size fraction changed from equal weight of fine and coarse particles into fine particles. An apportionment of carbonaceous materials shows secondary organic carbon is predominant in occupying 56%, followed by elemental carbon with 25%, and primary organic carbon with 19%. By applying a “near-source aerosol chemical properties” source apportionment to PM2.5, the contributions from secondary reaction is found higher than that from mobile vehicles. The contributions from the two sources could be summed up to 35-95% of fine mass. For the assessment of LWC of aerosol, aerosol LWC is found mainly determined by atmospheric RH and its chemical compositions. From the humidographs of aerosol LWC, the results will tend to underestimate if the measurement system is operated from dry state of arosol to the reconstructed RH. However, starting from a very high RH to the reconstructed RH will induce an overestimate in measuring aerosol LWC. The right procedure is to include aerosol LWC retained in the previous RH cycle. Meanwhile, for aerosol in the high RH stage, the measurements of aerosol LWC are close to each other. The aerosol LWC measured in this study is agreed well with ISORROPIA model estimate with a R2 at 0.86. However, the model underestimates aerosol LWC, which is probably due to not incorporating hygroscopic organics and all hygroscopic inorganic salts into the model. It is noteworthy to reveal that for temperature controlled in the range of 22-270C and RH within 26-34%, the conditioned particles before weighing still carry 3-29% LWC. The measurement of aerosol LWC at the RH when it was collected shows an average of 64% in the range of 86-30% of the fine mass. |
