臺灣山多平原少地勢陡峭，再加上河川沖刷因此形成許多沖積台地及平原，沖積地質經過時間的演化，地表下常夾雜不同厚度的低透水層。在地小人稠的臺灣，都會區快速發展地表不透水面積增加，導致雨季來臨時下游低窪地區淹水。如何在現有的都會空間中增加地表水入滲，將是目前當務之急。入滲速度快慢取決於土壤特性，若採用礫石樁工法貫穿地層中的低透水層，使地表水藉由礫石樁快速入滲。但是該工法所造成的入滲效益是否符合預期目標，需藉由數據量化加以分析討論。本研究目的透過現地及實驗室方法，配合數值模擬量化評估都會地區滯洪池增設礫石樁之入滲效益。研究場址位於台中市大里區工業區內的南區滯洪池，研究透過壓力鍋實驗得到現場土樣特徵參數，並進行現場入滲實驗得到飽和水力傳導係數，將實驗得到的土壤特徵參數和飽和水力傳導係數與地理數值高程運用FEMWATER數值模式，透過降尺度技術建立符合滯洪池尺度和礫石樁尺度之水文地質模式模型，設定滯洪池內不同積水深度、水力傳導係數以及地下水位面，量化分析不同礫石樁數量和距離等情境下入滲率變化。模擬滯洪池入滲率和礫石樁入滲特性，並針對現場規劃適當礫石樁數量。 實驗得到van Genuchen土壤特徵參數如下：飽和體積含水率為0.358（-）、殘餘體積含水率為0.067（-）、空氣入滲係數為0.643（1/L）及材料孔徑參數1.556（-），以及入滲試驗求得之飽和水力傳導係數（Ks）0.6（m/day）。模擬結果顯示滯洪池的入滲量影響主要來自現場水力傳導係數以及池內積水深度；模擬雙樁間距分析顯示樁距大於1m以上時，以現場年平均地下水位及材料分層條件，對入滲量變化沒有顯著改變；礫石樁尺度模型內增設一支礫石樁將使入滲量每天平均增加約2%。因此若由礫石樁尺度模擬結果回推滯洪池範圍下，增加2%入滲率則需於滯洪池中增設16支礫石樁，而研究場址已設置三支礫石樁其等效水力傳導係數為0.603（m/day），約可增加0.375%入滲率。未來可將進一步以現地實驗驗證分析結果，而礫石樁設置工法將可廣泛應用在台灣各地區以補注日漸減少的地下水資源。 ;The development of metropolitan leads to creasing impermeable surface areas, which therefore induce the risk of flooding in storm seasons. Installing gravel piles in the open areas such as parks or detention ponds in metropolitan areas provides alternatives to increase the infiltration of storm water. The objectives of the study are to quantify the infiltration efficiency of gravel piles installed in a retention pond. The numerical FEMWATER model associated with field and laboratory experiments are employed to assess the improvements of infiltration influenced by the number of gravel piles. The study area is located in the industrial park in Taichung Dali district. In this study, the pressure plate extractor and falling head tests were used to obtain the soil water characteristic curves and the saturated hydraulic conductivity K. Those soil characteristics parameters, saturated Hydraulic Conductivity and digital elevation model, will then employed in the FEMWATER numerical model for modeling Infiltration scenarios on different scales. A number of conditions, including different ponding depths in detention pond, hydraulic conductivity values, groundwater levels, and numbers and distances of gravel piles were used in the FEMWATER model to quantify the effects of the factors on the infiltration. Based on the van Genuchten soil characteristic formula, the results of pressure plate extractor test show that the soil sample has saturated soil volumetric water content of 0.358 (-), residual soil volumetric water content of 0.067 (-),the air entry suction of 0.643 (1/L), and the pore-size distribution of 1.556 (-). The saturated K is 0.6 (m/day) based on the in-situ infiltration experiment. The numerical simulations indicate that the infiltration is mainly controlled by hydraulic conductivity and the flooding depth in detention pond. Simulation results also show that the distance between two gravel piles didn’t have significant influences on infiltration rate when the distance is longer than 1 m. One gravel pile in gravel piles’ scaling model made can lead to 2% of increased infiltration compared with the no gravel pile condition. In the study area, the existing three gravel piles can increase the infiltration up to 0.375%.
