天然含水層因形成機制不同，均存在不同層度的異質性；此外，現實條件下，有限的觀測數據往往不足，若使用有限的觀測數據代入地下水模式，推估地下水水流與污染傳輸過程，其推估結果必然存在不確定性，此不確定性的量化近年來已成為風險評估的重要依據。本研究將利用既有數值模式如MODFLOW、MT3DMS、MODPATH或FEMWATER 等，由選定的研究場址配合既有觀測數據(如地表水文資料、地質鑽探取樣、地下水位量測、抽水試驗分析、示蹤劑試驗等)，先以地質統計方法建立場址參數的空間分布結構，再以序率方法(Stochastic methods)量化分析各標的參數對水流與污染傳輸所引致之不確定性。研究成果將提供標準分析流程，包括模式理論、模式建立、模式運作與結果分析等；此外， 將以選定的標的參數如水力傳導係數、降雨補注率、邊界條件等，對總體不確定性貢獻度，做分級成果可做為未來不同場址條件分析時的重要準則。Natural aquifers are heterogeneous due to the complex mechanisms of surficial deposit procedures and hydrogeologic conditions. For realistic problems, the data availability for groundwater flow and transport simulations are generally limited because of the restricted resources. Such complex nature of aquifer heterogeneity and insufficient data usually lead to uncertainty in the problems of modeling groundwater flow and contaminant transport. The issues of uncertainty have been widely considered and are the key parameters for risk analysis in recent applications. This study will employ MODFLOW, MT3DMS, MODPATH, or FEMWATER to conduct uncertainty analysis for groundwater flow and transport. The field data such as surface hydrological data, borehole data, groundwater levels, pumping and tracer tests and the associated parameters, are first used for spatial analysis to obtain the statistical structures of variables that are considered to be spatial random variables. This study then uses the statistical structure to perform Monte Carlo simulation to assess the degrees of uncertainty for different random variables. The study will provide technologies on analyzing uncertainty of groundwater flow and transport, giving the classifications of uncertainty for different variable such as hydraulic conductivity, recharge, and boundary conditions. The results and procedures can give guidelines for future investigations on site specific conditions, especially for those the filed observations are limited. 研究期間：10101~ 10112
