水力傳導係數(K)及比儲蓄係數(Ss)為影響地下水流動，地下水污染傳輸以及工程開挖施工等的重要水文地質參數。本研究利用水力剖面探測方法，配合含水層鑿井資料與抽水試驗數據，使用SSLE(sequential successive linear estimator)模式推估K與Ss在空間上較精細的分布。本研究以模式模擬配合現地抽水試驗同時進行，於模式模擬測試例方面，首先利用虛擬含水層之隨機水力傳導係數分布場，進行二維與三維暫態水流流場測試並推估含水層參數，對於SSLE模式的運作過程以及推估結果作概念性地說明，之後為了更進一步了解SSLE模式對不同含水層特性推估時之準確性，本研究對SSLE推估方法進行更深入的評估，透過改變不同生成隨機場的參數，例如改變變異數及x方向之相關長度來比較推估結果，以相同模擬區域及邊界條件，改變水力傳導係數分布場之變異數(0.1, 0.5, 1.0, 2.0)，與x方向相關長度(20m, 40m, 60m, 80m)，各產生20組隨機水力傳導係數場，再以SSLE模式，針對此不同異質性程度的含水層進行參數推估。結果顯示K值變異性越高則推估的誤差相對提高；在某一相關長度範圍內若有三組以上抽水反應觀測數據，即可以較完整地描述含水層之異質特性。在現地實驗部分，研究場址為高雄縣大寮鄉輔英科技大學地下水井場，利用水力剖面探測方法進行水平跨孔式與垂直跨孔式抽水試驗，藉以得到多組獨立抽水反應資料，以SSLE模式透過資料點在空間上位置結構特性，以及抽水試驗造成之含水層系統反應，分別模擬現地尺度下的K與Ss值在二維與三維空間上的分布情形。模式邊界條件之定水頭假設與現地條件之差異會對推估結果造成高估之影響，但將推估結果與地層資料柱狀圖比較，發現以此模式推估之水力傳導係數分布形態與實際情形大致相符，故此模式可以定性地推估出現地尺度下含水層中水力傳導係數高低分布情形，並能夠幫助了解該場址水文地質參數的空間分布狀況。 Hydraulic conductivity (K) and specific storage coefficient (SS) are key parameters to predict groundwater flow and contaminant transport and to evaluate the stability of excavation sites. This study uses the concept of hydraulic tomography surveys, which integrates the information from direct measurements of aquifer properties and pumping test data to inversely estimate the spatial distributions of hydraulic conductivity (K) and specific storage coefficient (SS) with higher resolution. The inverse model sequential successive linear estimator) SSLE is employed in this study to conduct the inversion of aquifer parameters. This study starts with two synthetic examples (horizontal 2D and vertical 3D cases) to introduce the concept of the SSLE inverse model, and the associated measurement procedures on sites. A variety of cases, including different variances of hydraulic conductivity and correlation lengths in x direction, are used to assess the effect of different degrees of the aquifer heterogeneity on the estimation results. The results indicate that the higher variance of aquifer properties the lower accuracy of the SSLE estimation results. Additionally, three pumping data within a correlation length can well characterize the aquifer heterogeneity. The inverse model is then applied to cross-hole pumping test data obtained from Fooyin University (FU) well field. Base on the concept of hydraulic tomography surveys, such independent pumping test data are then integrated in the inverse model to estimate the 2-D and 3-D spatial distributions of hydraulic conductivity and storage coefficient. Although the specified of constant head boundary condition may not fit well with the conditions on FU site, the estimation results show that the pattern of estimated K distribution agrees well with the well logs obtained from FU site.
