名竹盆地地下水流數值模式之建立與應用

Abstract

名竹盆地為丘陵與台地環繞而成之區域，位於濁水溪沖積扇扇頂上游之延伸區域屬濁水溪中游段，其間有清水溪及東埔蚋溪支流匯流至濁水溪本流被視為補注潛勢極高之地區。然而由於濁水溪沖積扇之地下水資源缺乏妥善經營與管理造成地層下陷、海水入侵等危害，並危及高鐵行車安全，是故開發新水資源甚為迫切。本研究目的即為建立一地下水流數值模式以供各種開發方案對地下水影響之評估工作。 建置地下水流數值模式前須先蒐集水文地質相關資料，並進行地下水系統辨識，了解其邊界、含水層與地下水系統流入與流出量，茲以地下水歷線法估算名竹盆地地下水系統之入流量與出流量，入流量以氧同位素質量守恆法細分雨水入滲補注及河水入滲補注，出流量經由估算則得到隘口出流量、抽水量以及流失量，而後將地下水系統辨識得到之水文量結果作為MODFLOW模式輸入與率定之依據。 為評估短時距內地下水與河水間之交換量以及豐枯水期之水文變化，本研究採模擬時限2008/11/1∼2009/12/31共14個月，以1日為一時間間距(time step)進行暫態(transient)模擬。MODFLOW中並引入河流演算套件(Stream-Flow-Routing package，簡稱SFR)，經由給定河川流量後以曼寧公式進行河道演算，得到每個河段(reach)所鏈結之數值網格(cell)代表的河川水位，接著依河川水位與地下水位之差值，以達西定律(Darcy’s Law)計算每日地下水與河川水之交換量。建置完成之模式可用於井群抽水開發情境模擬，本研究設置一種井群排列以及三種抽水量情境，模擬抽水開發對研究區域地下水文之影響。 研究結果顯示，名竹盆地大部分的地下水出流量為透過隘口出流，在隘口束縮與含水層通水斷面小之情形下，亦有大部分之地下水因隘口束縮壓力擠出至地表，以地面水形式流出隘口，在MODFLOW模擬時限內隘口地下水日出流量為0.17∼0.26百萬噸之間，14個月總量0.98億噸；隘口地面水日出流量為0∼0.14億噸之間，總量2.13億噸；雨水日入滲補注為0.02∼7.79百萬噸之間，總量2.25億噸；河水日入滲補注為0.064∼6.52百萬噸之間，總量1.64億噸；地面水與地下水日交換量為-0.12∼0.0168億噸之間，總量-0.68億噸（正為河水入滲補注地下水，負為地下水出滲河道與地表）。而井群抽水開發情境模擬結果顯示情境I抽水2億噸下能誘導河水入滲增加36.47%；情境II抽水4億噸下能誘導河水入滲增加47.98%；情境III抽水8億噸下能誘導河水入滲增加82.16%，顯示井群抽水量與誘導河水入滲補注呈現正相關，但抽水量愈大隘口地下水出流量損失亦越多，在情境III損失達80.47%，情境II為45.68%，情境I為23.77%。

Ming-Chu Basin surrounded by hills and terraces is located at the top extension of Jhuoshuei River alluvial fan and on the midstream segment of Jhuoshuei River. Within the basin, Qingshui River and Dongpurui River both converge to Jhuoshuei River, so Ming-Chu Basin is considered as a high potential recharging region of groundwater aquifer. However, due to lack of proper operation and management of groundwater resources, land subsidence, seawater intrusion and other hazards are induced, and the safety of high-speed rail is threatened. Therefore the development of new water resources is urgent. The purpose of this study is to establish a groundwater flow numerical model for impact assessment of a variety of development scenarios in groundwater. Building a groundwater flow numerical model requires collecting hydrogeologic data and identifying the groundwater system, including the boundary, stratification, and hydrologic input and output. In this paper, we utilize groundwater hydrograph method to evaluate each hydrologic quantity. The natural recharge from rainfall infiltration and stream leakage is further separated with the help of isotope analysis method. Through groundwater hydrograph method we could acquire recharge, pumping, loss and outflow from mountain pass, separately. These hydrologic quantities are input data of groundwater flow model (MODFLOW) for model calibration and flow simulation. To assess the short-term hydrological changes as well as the exchange between aquifer and stream in the rainy season or dried season, MODFLOW with Stream-Flow-Routing package (SFR) are used to simulate the groundwater level variation from 2008/11/1-2009/12/31, a total of 14 months, at a time step of 1 day. The stream level is calculated by given stream inflow rate and Manning's equation. Then Darcy's law is applied to evaluate the exchange capacity of the stream and aquifer during each time step considering the difference between stream level and groundwater level. The calibrated model is then used for well pumping exploitation scenario assessment. One pumping well cluster arrangement and three pumping rate are set and analyzed to assess the impact of exploitation on the groundwater system of Ming-Chu Basin. The results indicated most of groundwater outflow through the narrow-pass. In a combined influence of narrow-pass shrinkage and aquifer cross section reduction, daily outflow through aquifer cross section is between 0.17-0.26 million tons; while daily outflow through narrow-pass as surface water is between 0.00-0.14 million tons. Daily recharge from rainfall infiltration is 0.02-7.79 million tons; daily recharge from stream leakage is 0.06-6.52 million tons. Daily exchange between surface water and groundwater is -0.12-0.02 million tons (positive for stream water infiltrating to groundwater; negative for the groundwater exfiltrating to stream). The pumping effect of well cluster in different scenarios are: Scenario-I for 200 million tons pumping per year, could induce an increase in stream leakage by 36.47%; Scenario-II, 400 million tons, could increase leakage by 47.98%; Scenario-III, 800 million tons, increases leakage by 82.16%. The results exhibit a positive correlation between quantity of groundwater pumping and that of induced stream leakage. Nevertheless, the larger the pumping, the smaller the daily outflow through aquifer cross section.