應用條件模擬與分佈型水平衡模式於濁水溪沖積扇地下水抽水量推估不確定性之研究

Abstract

地下水抽水量為地下水管理之重要依據，而抽水量之推估一般多藉由數值模擬而得，惟實務上由於水文地質參數如透水係數之不確定，常使抽水量推估亦充滿高度不確定性，以往之相關研究亦已反應出這個問題，而這對相關管理措施的評估亦帶來相當的不便。因此，如何對抽水量推估之不確定，能有進一步的分析與瞭解，乃是地下水管理上重要的課題。 有鑑於此，本研究乃整合條件模擬與本研究自行發展之分佈型水平衡模式，應用蒙地卡羅模擬的理念，對抽水量推估之不確定性進行量化的分析。其中條件模擬乃以地質統計中的半變異元為基礎，合成多組透水係數網格值，並確保每組合成值皆能符合觀測資料之半變異元，而分佈型水平衡模式則可在給定水位及參數下，迅速模擬系統之抽水量。 本研究以濁水溪沖積扇為研究區域，整個模式共有3個含水層、2個阻水層，其中半變異元模型為高斯模型，而相關參數從F1至F3的Nugget分別為0.391、0、0；Range則分別為23778、6400及7478公尺；Sill則分別為1.7、2.134及2.77。進一步蒙地卡羅模擬分析成果顯示，並將不確定性以5%及95%的信賴區間作表式，第一含水層之年淨補注量5%及95%信賴區間為6.26億噸與4.22億噸之間；第二含水層年抽水量5%及95%信賴區間為5.60億噸與3.56億噸；第三含水層年抽水量5%及95%信賴區間為2.68億噸與1.18億噸。平均而言，第一層之淨補注量為5.15億噸，第二層及第三層之抽水量分別為4.56億噸及2億噸。由前述成果可看出將具有相同半變異元的透水係數場輸入至分佈型水平衡模式，確會為區域抽水量之推估帶來相當的不確定性，此可為後續相關管理決策之參考。

Groundwater pumping information is very important information for groundwater management. However, practically, the estimated pumping information is highly uncertain due to uncertainty in hydrogeological parameters such as conductivity coefficient. This uncertainty leads to an important issue of groundwater management. This study develops a distributed water balance model integrated with groundwater simulation model for simulating the groundwater flow. For the uncertainty of the of the estimated pumping, Monte-Carlo simulation is applied. Multiple grid sets are generated using semivariogram method. With given groundwater level and parameters, the groundwater pumping is estimated. The proposed method is applied to Choshuihsi River alluvial fan. Three aquifers (F1 to F3) and two aquitards are included in the simulation model. Gaussian distribution is assumed in the semivariogram model. For three aquifers: the values of Nugget are 0.391, 0, and 0; the values of Range are 23778, 6400, and 7478; the values of Sill are 1.7, 2.134, and 2.77. The 5% and 95% confidence interval for three aquifers are: (1) net recharge of 626 million and 422 million tons of water for F1; (2) net pumping of 560 million and 356 million tons of water for F2;(3) net pumping of 268 million and 118 million tons of water for F3. In average, F1 has net recharge of 515 million tons; F2 has net pumping of 456 million tons; and F3 has 200 million tons, respectively. These results could provide critical information for future groundwater management.