2010年3月4日高雄甲仙地區發生了芮氏規模6.4的強震，造成許多房屋毀損，故本研究希望藉由甲仙地震資料，使用隨機式有限斷層法配合場址修正，進行強地動預估，希冀能在以後結合地質構造資料提供一個可用的方法模擬情境地震可能造成之地動情形。 有限斷層法是將一個斷層面切割成許多大小相同的子斷層，每個子斷層視為一個具有隨機ω^2頻譜之點震源，將每個子斷層錯動時的貢獻加總起來即為此次地震事件能量的總和。 本研究利用隨機式點震源模擬的方法將小地震的觀測值與模擬值之間的差異當作各測站的之經驗轉換函數，以及利用單站頻譜比法（Nakamura, 1989）當作轉換函數的兩種方式進行場址修正，而後將其結果與衰減公式模擬之結果比較。 結果顯示利用單站頻譜比當作轉換函數的場址修正方法對於PGA結果來說會比較好；利用經驗轉換函數的場址修正方法則在頻譜的擬合結果上較不錯，未來可針對情境地震選擇良好的場址修正方法。 The Jiashian earthquake occurred on March 4th, 2010 in Kaohsiung with local magnitude M_L=6.4. It caused many buildings damage. In this study, stochastic finite fault method was used with site correction to simulate strong ground motion. In future, we hope to apply this available method with geologic structure data to simulate the ground motion for other scenario earthquake. The stochastic finite fault method which can divide the fault plane of large event into many sub-faults, and each sub-fault is treated as a point source and assigned a ω^2 spectrum. The total energy of the large event is summed up from the contribution of each sub-fault. Site correction in this research had two kinds: 1) the bias between spectra of stochastic point source simulated and observed from small earthquakes to be an empirical transfer function 2) H/V ratio from observed earthquake records was treated as a kind of transfer function. The results also compared with that from the attenuation equation. The results showed H/V ratio site correction was better for PGA simulation and empirical transfer function correction was better than H/V ratio in spectrum. Hopefully, site correction will be selected for the simulation of scenario earthquake.
