愛氏震度為地震歷時記錄之加速度平方後積分所得到的地動參數，是一個可以包含地震波之振幅、頻率內涵和持續周期之地動參數。愛氏震度與許多地震災害有相關性，如建築物之破壞、液化及地震引致之山崩災害等。本研究以愛氏震度作為地振動指標，針對此進行衰減式之研究。合適的衰減式不僅可以瞭解一地區之地動衰減特性，並可以用來預估特定場址之地動值，提供工程耐震設計之用。 本研究以中央氣象局自由場強地動觀測網計畫（TSMIP）及中央研究院地球科學研究所SMART-1陣列之豐富強震資料，考慮場址特性、震源機制以及地體構造特性差異，以理論推導為基礎的衰減模型，增加場址變數Vs30之考慮，並以混合效應模型及最大概似度法進行非線性迴歸分析，完成適合台灣地區之地殼地震及隱沒帶地震愛氏震度衰減式。 研究成果顯示，混和效應模型搭配最大概似法能解決地震規模不確定性及迴歸資料分布不均之問題，衰減模型中加入Vs30及震源機制兩項參數可有效降低迴歸結果之標準差。地殼地震愛氏震度衰減式所推估之地動值在近距離較前人研究成果為高、遠距離則略低，可能反映了台灣地殼地震衰減較快之特性。隱沒帶地震之愛式震度衰減式無前人研究結果可進行比較，而與地殼地震衰減式相比，所推估之地動值明顯地較高，顯示地殼地震與隱沒帶地震之愛氏震度衰減情況並不相同。 Arias Intensity (AI) is a ground motion parameter of an earthquake record as the integral of the square of the acceleration time history. It incorporates the amplitude, frequency content, and duration of the ground-motion, and is likely to be a more reliable predictor of earthquake damage potential. AI correlates well with several commonly used demand measure of structural performance, liquefaction, and seismic slope stability. This study develops the Arias intensity attenuation relationship in Taiwan. A good attenuation equation can reflect the characteristics of the ground-motion attenuation for a region, and can be used to predict the ground-motion value of a specific site for seismic resistance design. Two local empirical attenuation relationships for the crustal and the subduction zone earthquakes respectively are developed to estimate AI as a function of magnitude, distance, fault mechanism, and continuous site variable-Vs30, based on strong ground-motion data from TSMIP and SMART1 array in Taiwan. Its functional form is derived from the point-source model, and the coefficients are determined through non-linear regression analyses using maximum likelihood method (MLE) and mixed-effects model. The results show that mixed-effects model with MLE can effectively solve the regression problem in the treatment of uncertainty of the earthquake magnitude and the data weighting. Vs30 and fault mechanism used in the attenuation model can reduce the sigma significantly. The AI value predicted by the crustal earthquake attenuation equation is higher in the near distance and lower in the far distance than previous researches. No previous research was done on the subduction-zone earthquake attenuation for AI. To compared with the crustal earthquake attenuation equation, the subduction zone earthquake attenuation equation predicts significantly higher AI value than the crustal one. This indicates that the AI attenuations behave differently between the crustal earthquakes and the subduction zone earthquakes; similar to other ground-motion parameter do.
