土壤滲透係數（permeability）及孔隙率(porosity)常被用來作為氣相流體在非飽和土壤或岩石中流動能力的指標，為非飽和層中兩項非常重要的基本參數。傳統上欲得知此兩參數常使用水力試驗來量測，但在現地使用水力試驗量測非飽和層中土壤或岩石參數在技術上具有相當大的困難性，因此許多研究已發展出使用注入氣體的方式取代水力試驗來量測非飽和土壤參數。過去研究中所發展的注氣試驗（pneumatic test）大多僅能量測整個土樣之平均氣體滲透係數，本研究將使用相似於水力剖面掃瞄法（hydraulic tomography）之概念，應用於一維與二維砂箱跨孔式注氣試驗數據，推估砂箱內非飽和土壤參數的空間異質性分佈。跨孔式注氣試驗（pneumatic cross-hole test）乃將穩定氣體注入非飽和砂樣中，利用多次變換注氣點、觀測點及注入之氣體流量，記錄通過砂樣之氣壓變化，可獲得多組氣壓變化數據，最後套用前人研究中所使用之解析解計算出其氣體滲透係數（gas permeability），並參考ASTM D854-83規範（土粒比重試驗）間接量測出各砂樣之孔隙率。再結合剖面掃瞄法概念所發展出來之序率反推估模式，對實驗室尺度之注氣試驗數據進行土壤氣體滲透係數及孔隙率之推估。砂箱實驗結果顯示，即使在相同的砂樣中，注氣量的大小及觀測位置影響推估氣體滲透係數的結果甚巨，相同砂樣設置甚至有10倍以上的推估差異。垂直二維模式測試例顯示此模式可非常準確的推估出氣體滲透係數的空間分佈情形，但孔隙率推估結果稍差。套用實驗數據後的推估結果並不如預期，僅氣體滲透係數獲得合理之參數範圍，砂箱之砂樣分佈型態未能完全推估出來，此外，由於實驗操作方式或砂箱設備問題，使得氣壓初期變化過快過大，造成無法推估出孔隙率於砂箱中的分佈。 Soil permeability and porosity are key indicators usually used for characterization of gas phase flows in unsaturated soil or rocks. Classical approaches to obtain these two parameters usually employ hydraulic tests. However, the characterization of unsaturated flow parameters for soil or rocks by mean of hydraulic tests is fraught with technical difficulties. To resolve the difficulties, much literature has conducted field tests with gas rather than with water to characterize the flow parameters of unsaturated soils. Transitional pneumatic tests can only characterize average gas permeability for the tested soil sample. This study uses the concept similar to the one in the hydraulic tomography surveys and applies to cross-hole pneumatic injection tests in laboratory sandbox. Such measurements from pneumatic tests will then be used for estimations of unsaturated soil parameters in sandbox. Pneumatic cross-hole test injects gas in one location and monitoring the variations of air pressures at other locations in the sandbox. The same injection event is repeated at different injection locations. The sequential pneumatic tests can yield a series of gas pressure measurements. Such measurements are first for permeability analyzed by employing approaches that are proposed by previous studies. The estimation of porosity follow the procedures in ASTM D854-83（Test Methods for Specific Gravity）. Same cross-hole measurements are then used in the stochastic pneumatic inverse model to estimate gas permeability and porosity distributions of soil in sandbox. The one-dimensional sandbox experiments show that the flow directions can significantly affect the final estimation of gas permeability with same measurement locations of two pressure probs, the permeability value can varies in one order. The stochastic pneumatic inverse model can predict well the spatial distributions of gas permeability and fairly well the pattern of porosity in two-dimensional synthetical case with laboratory scale. However, the data from sandbox measurements only result in a fair estimation of permeability ranges. The inversion of porosity distribution for 2-D sandbox is unavailable for the presence of soil distribution.
