|dc.description.abstract||本文於2013年春季在鹿林山背景觀測站(以下簡稱為鹿林山，2,862 m)與夏季龍潭空品測站(以下簡稱為龍潭)以即時氣膠水溶性監測儀(Particle-Into-Liquid-Sampler coupled to an Ion Chromatograph, PILS-IC)觀測PM2.5氣膠水溶性離子，並收集站址觀測的PM2.5質量濃度、PM2.5散光及吸光係數、氣膠粒徑分布、氣膠總數目、氣體污染物動態變化。
結果發現在鹿林山觀測的三次雲霧事件，CO、NOx有隨時間上升的趨勢，推斷是平地氣體污染物受到山下氣流抬升至鹿林山而形成雲霧，此時氣膠質量濃度、散光及吸光係數、氣膠粒徑分布及總數目都有相對應的增大或變化。氣膠水溶性離子NH4+、SO42-、NO3-有明顯高濃度出現，NH4+並顯示是由NH3轉化而來。NH4+與SO42-有中度以上相關性(R2>0.64)，因此，這兩種離子結合型態可能為硫酸銨或硫酸氫銨。從過剩ExNO3-與過剩ExNH4+計算判斷NO3-生成機制 (Mwaniki, et al., 2014; Schlager, et al., 1990)，發現在不同時段分別有HNO3 (g) 凝結在氣膠表面、N2O5水解、以及硝酸銨氣膠的形成。
鹿林山觀測的生質燃燒事件(biomass burning, BB)中CO、NOx、O3、PM2.5質量濃度、PM2.5散光及吸光係數、氣膠總數目都有增高現象。NH4+和SO42-在大多數時間關聯性很好，而且K+有顯著的濃度增加。在五次BB時段有六次雲霧事件，其中四個雲霧事件經由K+增益量判斷，雲霧與生質燃燒氣團來源相同。生質燃燒氣團長程傳輸的硫酸鹽轉化比值(Sulfur oxidation ratio, SOR)可達0.9，顯示硫化物主要以SO42-形式存在。SO2在各次生質燃燒事件都沒有明顯的濃度增加趨勢，表示生質燃燒不會產生大量SO2且SO2來自遠端各方面污染源經過了大氣的均勻混合。
龍潭為平地測站，觀測期間發生光化反應時，NH4+、SO42-、NO3-濃度增加，PM2.5佔PM10比重上升至50%，表示有細粒徑氣膠產生。另外，龍潭在高濃度硝酸鹽時段，除了正午是受到光化反應外，在夜間推測是受到外地傳輸及混合層降低的影響；由於NH4+和NO3-有中度以上相關，推論龍潭光化與高濃度硝酸鹽事件是以硝酸銨為硝酸鹽主要組成。由於夏季生質燃燒活動較少，龍潭Knss+/NO3-與Knss+/SO42-數值明顯低於鹿林山，而且SOR和硝酸鹽轉化比值(Nitrogen oxidation ratio, NOR)也是受到污染物傳輸路徑較短的緣故，普遍低於鹿林山。
比較Aerosol Inorganic Model II (Clegg et al., 1998a, b)模擬的氣膠含水量(Aerosol Water Content, AWC)和鹿林山吸濕差異移動度粒徑分布儀(Humidified Scanning Mobility Particle Sizer, H-SMPS)量測的氣膠吸濕參數(κ)，生質燃燒氣團由於會帶來大量較不吸濕的含碳氣膠使κ值低，但雲霧事件也可能因氣膠吸濕粒徑變大超過H-SMPS量測上限，使得剩下所採集氣膠都是較不吸濕因而降低κ值，整體而言，生質燃燒事件κ值仍較雲霧事件低。然而，在氣膠含水量方面，因生質燃燒氣膠有高濃度SO42-、NO3-，因此鹿林山生質燃燒事件模擬的AWC是大於雲霧事件。龍潭高濃度硝酸鹽事件κ值與AWC模擬都高於光化事件，由於龍潭高濃度硝酸鹽事件SO42-、NO3-相加總濃度是高於光化事件，這指出SO42-、NO3-是主導κ值與AWC的重要因子。
|dc.description.abstract||This work monitored water-soluble inorganic ions (WSIIs) of atmospheric PM2.5 using Particle-Into-Liquid-Sampler coupled with an Ion Chromatograph (PILS-IC) at the Lulin Atmospheric Background Station (LABS, 2,862 m a.s.l.) in spring and Longtan air quality monitoring station (Longtan) in summer in 2013. Meanwhile, PM2.5 mass concentration, PM2.5 scattering and absorption coefficients, aerosol size spectra, aerosol total number concentration, and dynamic variations of gaseous pollutants were also measured at LABS.
CO and NOx were observed to increase with time during the period of three fog events at LABS. It suggests that ground gas pollutants were transported by the uplift flow to form fog at LABS. The PM2.5 mass concentration, PM2.5 scattering and absorption coefficients, aerosol size spectra, and aerosol total number concentration were also increased in the fog events accordingly. The levels of NH4+, SO42-, and NO3- of water-soluble inorganic ions were enhanced and NH4+ was observed from the conversion of NH3. Moderately high linear correlation between SO42- and NH4+ (R2>0.64) indicates that the compound form of these two ions might be ammonium sulfate or ammonium bisulfate. From the calculations of ExNO3- and ExNH4+, three nitrate formation mechanisms (Mwaniki, et al., 2014; Schlager, et al., 1990) can be inferred to be condensed HNO3(g) onto aerosol surface, N2O5 hydrolysis, or the formation of ammonium nitrate particles in different times.
CO, NOx, O3, PM2.5 mass concentration, PM2.5 scattering and absorption coefficients, and aerosol total number concentration were all increased during the biomass burning events (BB) observed at LABS. The concentration between SO42- and NH4+ was consistently varied for most of the time and K+ concentration was significantly enriched. For four of the six fog events during the five BB time periods, they were influenced by the transported BB air flow judged by the enhanced K+ concentration. The sulfur oxidation ratio (SOR) could be as high as 0.9 to indicate that SO42- was the major species of sulfur. Interestingly, SO2 concentration was quite stable during BB observation, which implies that SO2 is produced from a variety of sources through uniformly mixing in the atmosphere rather than produced drastically from BB.
Longtan is located nearly on the ground; the concentrations of NH4+, SO42-, and NO3- were increased when photochemical events occurred. This will lift the fraction of PM2.5 over PM10 to above 50%, which indicates the production of fine particles. In addition, high concentration of nitrates is caused not only by photochemical reactions at the noon time but also possible distant transport and shallow mixing layer in the evening. Since NH4+ correlated more than moderately well with NO3-, the ammonium nitrate was inferred to be the main nitrate compound form in photochemical and high nitrate events at Longtan. Owing to less BB activity in summer, the computed values of Knss+/NO3- and Knss+/SO42- at Longtan were significantly lower than LABS. Similarly, values of SOR and nitrogen oxidation ratio at Longtan were also lower than LABS due to shorter transport distance of pollutants.
Simulated aerosol water content (AWC) from Aerosol Inorganic Model II (Clegg et al., 1998a, b) was varied consistently with the measured aerosol hygroscopic parameter (κ) from Humidified Scanning Mobility Particle Sizer (H-SMPS) in the fog events at LABS. The values of κ might be reduced by the great amount of less hygroscopic carbonaceous aerosol in the BB air flow; however, the growth of hygroscopic aerosol to exceed the upper limit of H-SMPS might also make the collected remaining aerosol less hygroscopic as to lower the values of κ in the fog events. All things considered, the values of κ in the BB air flow are smaller than that of fog events. In contrast, the simulated AWC from the BB air flow is greater than that of the fog events because of the higher levels of SO42- and NO3-. The values of κ and AWC are both greater in high nitrate events than that of photochemical events at Longtan caused by the higher total level of SO42- and NO3-. This demonstrats the dominant effect of SO42- and NO3- in the determination ofκ and AWC.