| 摘要: | 人類對於石油資源的需求隨著經濟發展而與日俱增，石油除了作為供給能量的燃料之外，亦可進一步精煉成多種副產品，作為其他產品的製造原料。然而因人為操作管理不當所導致的洩漏、管線破損或非法棄置，對於環境生態的衝擊也日趨嚴重，造成人類健康風險上的威脅。石油碳氫化合物所造成的污染因其密度比水輕、水中溶解度低的特性，被稱為輕質非水相液體(Light Non Aqueous Phase Liquids，簡稱 LNAPLs)。土壤採樣、地下水監測井設置、地下水採樣、現地快速篩測及實驗室化學分析等，為目前污染場址調查時最常採用的方法。從現地資料收集、樣品採集、專一或多種化學物定性/定量分析，到最終數據闡釋及場址污染概念模式建立，需要多樣專業技術的跨平台整合才有辦法達成，而這些複雜且昂貴的成本花費，為的就是能清楚且準確的掌握場址污染的範圍、程度與發生的來源。前述的污染調查技術有個共通點，就是往往需要高密集度的採樣才能完整地瞭解場址污染狀況，降低因採樣佈點所造成的漏失風險。如何在調查目的、調查品質與調查經費中取得平衡，為場址調查計畫執行時最重要的關鍵。透地雷達(Ground Penetrating Radar, 簡稱 GPR)為一種近地表高解析度電磁波探測技術；經由透地雷達探測，可快速且大面積地收集場址地表下的資訊，透過資料處理技術做即時性的判釋，在進場調查前期提供場址污染及水文地質相關訊息，有效地在空間上區分且縮限調查範圍，讓接續的調查規劃與資源作更有效率的配置整合，發揮場址調查的最大效益。
特性分析(Attribute analysis)為一種廣泛使用於震波探勘的資料處理技術，在本研究中被應用於透地雷達資料處理，透過特性分析使原本訊號中細微的變化加
以顯現，並用以偵測石油碳氫化合物污染所可能造成的反應。特性分析在透地雷達訊號上的運算快速，所有的資料處理可在現地以一台個人電腦完成，做即時的資料分析與闡釋。在本研究中以三個不同污染程度與類型的場址作為案例分析，用以評估透地雷達與特性分析技術在 LNAPLs 污染場址調查的適用性。
根據實際採樣分析的結果，透地雷達特性分析可顯現並標示污染物所造成的
細微物性變化，透過特性分析中不同特性(Attribute)學理上的定義，用以解釋可
能的成因，並進一步推測可能的污染概念模式。在未飽和層(Vadose zone)土壤中LNAPLs 殘留相(Residual phase)會在地下水位面或毛細帶產生大範圍的污染髒污帶(Smeared zone)，當雷達波反射/ 折射於此區帶時，在瞬時相位(Instantaneous phase)上會造成較劇烈的擾動；而在瞬時頻率(Instantaneous frequency)也可觀察到較高且尖銳的頻率反應，進一步探討成因，可能是由殘留相污染物所形成的薄層(Thin beds)所產生的雷達波干涉現象。
本研究將各個不同案例場址所收集的雷達資料與現地實際採樣、分析的結果作比對，證實透地雷達探測污染物所造成細微電性變化的能力，將地球物理探測結果與實際觀察做連結，發展一套結合地球物理探測技術並且具有實務可行性的場址調查方法與策略方向。;Fossil fuel is still the primary source of energy and can be refined to other side products as raw materials or solvents. Spill, uncontrolled disposal, storage tank/pipeline leakage of fuel oil and solvents are threatening to biological system and impacting our living environment and even for human health risk. Hydrocarbon contaminant whose density is lighter than water would be considered as LNAPLs (Light Non Aqueous Phase Liquids). Common approaches including soil sampling, monitoring well installation, groundwater sampling, in-situ screening, and laboratory analysis are conducted during ordinary site investigation. From field sample collections, specific or multi-compounds analyses to data interpretations, all these complicated and expensive efforts are performed in order to detect and confirm the source of contaminant. One common problem which relies on the above mentioned conventional point measurements usually unable to provide clear range of contaminated area without dense sampling. The important issue is how to maintain a good balance among the investigation purpose, field operation quality control and budget. Near-surface GPR (Ground Penetrating Radar) survey provides an economical geophysical approach that offers large scaled, quick survey, real time interpretation and determination of possible occurrence of contaminated site. With preliminary GPR survey, areas of the site can be categorized according to the probability of being contaminated, and further, the range can be narrowed down.
Attribute analysis, a commonly used data processing technique in seismic exploration, is implemented for GPR data to reveal detailed information in detecting the presents of the hydrocarbon contaminant. Three different types of known hydrocarbon contaminated sites are investigated in order to examine its feasibility. GPR profiles acquired from each sites in conjunction with field sampling and laboratory analyses results proposed in our study would provide a useful approach for site investigation, monitoring and remediation. Detailed GPR attribute analysis demonstrates that the proposed approach is capable of detecting minor attribute changes caused by the contamination phases and thus providing very useful information for re-occupied or even for time-lapsed site investigation and monitoring.
The residual phase of LNAPL in the vadose zone will produce a large area of smeared zone. Instantaneous phase which reflects the polarity change shows relative aggressive change across water table in those seriously contaminated area. High and spiky instantaneous frequency can be observed around reflected signals from water table or capillary fringe which could be related to the detection of thin beds. Studying GPR signal attributes is extremely useful as the proposed methodology could provide as a strong indicator for detecting hydrocarbon contamination events. |
