台北盆地在地震時受盆地幾何形狀與土層特性共同影響，造成場址效應，易增強長週期之地振動。反應譜中的轉角週期可整合震源、路徑與場址效應，簡潔描述地震運動的頻率內涵，並充分反映場址效應。憑藉台北盆地內高密度實測資料，轉角週期可直接推求並用於地震微分區與耐震設計評估。然而，強震站空間分布仍無法完整覆蓋所有行政分區，無測站之區域可能產生估計誤差，本研究利用台北盆地大量之鑽井、地質資料，作為轉角週期空間內插輔助變數，以推估台北盆地轉角週期之空間分布。 本研究蒐集1992至2024年間台灣地區中大型地震事件紀錄(地震矩規模大於五，或盆地內任一強震站觀測到PGA大於10 gal的地震事件)，在統一波形處理流程下完成反應譜資料庫，並據此計算各強震站之反應譜與轉角週期。研究依震源特性將地震分為地殼、隱沒帶界面型與隱沒帶內部型三類，並取不同震源特性之平均值反應譜作為搜尋轉角週期時的目標反應譜。結果顯示，轉角週期在地殼地震與隱沒帶界面型地震均呈現明顯的區域差異，盆地西北側至北北東一帶的轉角週期最長，可達1.9秒以上，而東南側邊緣則最短，約為0.3秒；隱沒帶內部型之空間差異則不顯著，轉角週期介於0.2秒至0.8秒。本研究以各強震站不同地震類型之最大轉角週期作為代表性轉角週期，其空間分布與盆地沉積層結構一致，且以隱沒帶界面型地震為台北盆地長週期地震反應主要控制來源。 相關性分析進一步顯示，轉角週期空間分布與軟弱土層深度分布呈高度相關，與剪力波速呈低度相關，擁有越深厚軟弱沉積層之地區轉角週期越長，顯示地質條件可用以解釋反應譜形狀的區域變化。本研究進一步將上述地質參數納入迴歸克利金空間內插，以地質資料輔助推估無強震站區域之轉角週期分布，與未納入地質輔助變數的一般克利金及通用克利金相比，迴歸克利金能獲得更為細緻且具物理意義的轉角週期分布與地震微分區成果。;The seismic response of the Taipei Basin is jointly influenced by its basin geometry and soil conditions， resulting in pronounced site effects and enhanced long-period ground motions. The corner period (T0) of the response spectrum integrates source， path， and site effects， providing a concise description of the frequency content of ground motion and effectively reflecting site effects. Owing to the dense strong-motion network in the Taipei Basin， T0 can be directly derived from observed records and applied to seismic microzonation and seismic design evaluation. However， strong-motion stations do not fully cover all administrative districts， which may introduce estimation uncertainties in areas without observations. To address this limitation， this study incorporated extensive borehole and geological data from the Taipei Basin as auxiliary variables for spatial interpolation of T0. This study compiled a comprehensive record of moderate-to-large earthquakes (MW ≥ 5 or local PGA ≥ 10 gal) in Taiwan from 1992 to 2024. A unified waveform-processing workflow was applied to establish a response spectrum database. Earthquakes were classified into crustal， subduction interface， and subduction intraslab events according to their source characteristics. For each category， the T0 was determined using the mean spectrum as the target. The results show that T0 exhibits clear regional variations for both crustal and subduction interface earthquakes. The longest values occur in the northwestern to north-northeastern part of the basin and exceed 1.9 s， whereas the shortest values， approximately 0.3 s， are found along the southeastern basin margin. In contrast， subduction intraslab earthquakes show relatively limited spatial variability， with T0 values ranging from 0.2 to 0.8 s. By selecting the maximum T0 at each station among different earthquake types， a representative corner period map of the Taipei Basin was derived. Its spatial distribution is consistent with the sedimentary structure of the basin， and subduction interface earthquakes are identified as the primary controlling source of long-period seismic response in the Taipei Basin. Correlation analysis further indicates that the spatial distribution of T0 is strongly associated with the depth of soft sedimentary layers but only weakly related to VS30. Areas underlain by thicker soft sediments generally have longer corner periods， suggesting that geological conditions can explain regional variations in the shape of the response spectrum. These geological parameters were subsequently incorporated into Regression Kriging to estimate T0 in areas without strong-motion stations. Compared with Ordinary Kriging and Universal Kriging， which rely only on observed values and their spatial relationships， Regression Kriging produces a more detailed and physically meaningful distribution of corner periods and corresponding seismic microzonation results.
