集集地震對濁水溪沖積扇水文地質特性之影響

Abstract

水力傳導係數是分析水文地質、地下水流及污染傳輸極為重要之參數，集集地震引致地下水位異常變化，亦可能造成土壤結構及水力傳導係數之改變。本研究是將集集地震引致同震地下水位變化視為大自然本身執行之水文試驗，藉由分析水位變化，以期能瞭解集集地震對本區地下水水文地質環境之影響。首先分析將本區劃分為非侷限含水層、侷限含水層與山區之同震水位離震央距離與流通係數之關係，然後探討觀測井同震水位歷線變化與水文地質特性之關係，但均未獲得具體之相關性。再藉由MODFLOW模式模擬地震後地下水位之回復，UCODE模式進行參數優選，探討地震後水力傳導係數及儲水係數之改變或邊界條件之改變。以彰化溪湖、員林及溪州三處土壤液化區之地下水觀測站井進行抽水試驗及微水試驗，分析921地震後該三處站井之流通係數及儲水係數之改變，並與921震前水利署於民國82-83年量測之流通係數及儲水係數比較。 抽水試驗結果顯示流通係數部分除溪湖（二）觀測井減少4.2 %其餘皆為增加，溪州（二）觀測井變化最大增加80.52 %，員林（二）觀測井增加61.02 %；儲水係數部分於溪州（二）觀測井減少83.27 %，而以溪湖（二）觀測井減少60.07 %。本研究經由地震加速度及土壤液化分析求得地震前後土壤孔隙率變化與流通係數變化一致，土壤壓縮係數變化與儲水係數變化一致，顯示地震後員林(二)有土壤液化現象且地震加速度大使土壤孔隙率增加，導致流通係數增加，溪州(二) 地震加速度大使土壤孔隙率增加，導致流通係數增加，而溪州(二)及溪湖(二)之土壤壓縮性減少，導致儲水係數減少。以地震後地面水與地下水水文條件改變進行模式模擬，由模擬與觀測水位歷線擬合結果顯示員林(二)及溪州(二)觀測井模擬結果優於溪湖(二)觀測井。模擬結果顯示集集地震對濁水溪沖積扇水文地質參數影響很小，集集地震主要是增加濁水溪流出水量及東側山區邊界之流出水量。後續研究可考慮地震發生當時應力引致不排水荷重及土壤孔隙率隨時間變化之關係，以改善模擬與觀測結果之擬合。

Hydraulic conductivity (K) is an important parameter for the investigations of the hydrogeology, groundwater flow and contaminant transport. Soil structures and hydraulic conductivities maybe vary due to the 1999 Chi-Chi earthquake in the Choushui River alluvial fan. In this study, the changes of groundwater level induced by the Chi-Chi earthquake are regarded as the water-level fluctuations of a hydraulic test made by nature. We respect to understand the effect of groundwater hydrogeological properties in Chi-Chi earthquake by analyze the water level changes. First, this study analyzed the relationship of coseismic water-level change versus the distance from epicenter and transmisivity of wells in unconfined aquifers, confined aquifers and mountain region. The recovery of water levels in specified observation wells and hydrogeological properties was analyzed. The analyzed results indicate that the change of coseismic groundwater levels does not correlate well with the distance from epicenter and transmisivity. The numerical model, MODFLOW, was used to simulate the recovery of coseismic water-level changes and the automatically calibrated model, UCODE, was used to determine the changes of K/S on the boundary conditions after the Chi-Chi earthquake. This work performed the in-situ pumping test and the slug test in the liquefaction area of Shi-Hu, Yan-Lin and Shi-Chou monitoring wells after Chi-Chi earthquake occurrence. The transmissivity and storage coefficients were determined and compared with those measured before the Chi-Chi earthquake in 1993-1994. The results of pumping test show that the transmissivity increases with the highest of 80.52% at the Shi-Chou 2 observation well and increases 61.02% at the Yan-Lin 2 observation well. However, the transmissivity decreases to original one of 4.2% at the Shi- Hu 2 observation well. Furthermore, the storage coefficients decrease with the highest of 83.64% at the Shi-Chou 2 observation well, and decrease 60.07% at the Shi-Hu 2 observation well. The analyzed results of the seismic acceleration and soil liquefaction reveal that the changes of the porosity and soil consolidation coefficients agree with the changes of the transmissivity and storage coefficients, respectively, after the earthquake. At the observation well of Yan-Lin 2 where the soil liquefaction occurred, the soil porosity and the transmissivity increase due to the large seismic acceleration. At the observation well of Shi-Chou 2 where the soil liquefaction does not occur remarkably, the soil porosity and the transmissivity also increase because of the large seismic acceleration. Meanwhile, at the observation wells of Shi-Chou 2 and Shi- Hu 2, the storage coefficients decrease owing to the decreased soil consolidation coefficients. The simulation of the coseismic water-level recovery at the monitoring wells of Yan-Lin and Shi-Chou performs better than that at the monitoring well of Shi-Hu when the simulated conditions of groundwater inflow into rivers and groundwater loss in mountains are imposed on the model. Some factors, such as the transient changes of porosity and undrain loading, do not be considered in this model. A robust model containing these factors should be considered to develop and obtain on improved simulation and observation result.