本研究模擬Liao et al. [1]於1988年將羅吉斯迴歸模型應用於土壤液化機率模型的方法，使用Hwang et al. [2]於2021年所收集的509筆土壤液化數據，建立一個新的土壤液化機率模型，並透過四起地震驗證該模型是否符合實際情形；再根據工程地質探勘資料庫和強震測站場址工程地質資料庫收集台北盆地330孔鑽孔資料及高雄292孔鑽孔資料，代入本研究之土壤液化機率模型與Liao et al. [1]提出之土壤液化機率模型中，並進一步考慮台北盆地和高雄地區的活動斷層可能誘發Mw7.0的地震，利用新開發的土壤液化機率模型預測未來50年內這兩個地區的液化機率，將其繪製成液化機率圖。 ;Taiwan, situated along the Pacific Ring of Fire, experiences frequent seismic activity due to the convergence of the Eurasian Plate and the Philippine Sea Plate. This collision has transformed Taiwan into a continental island prone to earthquakes. Among these seismic events, many significant ones have triggered soil liquefaction phenomena, raising concerns about its impact.
In this study, we adopted the methodology proposed by Liao et al. [1] in 1988, which utilized the logistic regression model to develop a soil liquefaction probability model. Using the dataset of 509 soil liquefaction occurrences compiled by Hwang et al. [2] in 2021, we established a new model and validated its accuracy based on four seismic events. Additionally, we collected soil data from 330 boreholes in the Taipei Basin and 292 boreholes in Kaohsiung from geological exploration databases. The data, along with information from strong-motion station site databases, were used to evaluate both our model and the one proposed by Liao et al. (1988) [1]. Furthermore, considering the potential that an Mw 7.0 earthquake could be induced by active faults in the Taipei Basin and Kaohsiung, we produced liquefaction probability maps for both regions within the next 50 years using the newly developed liquefaction logistic regression model.
