三維速度模型雙差分地震定位程式之開發及其在2013年台灣瑞穗地震序列之應用

Abstract

地震定位是地震觀測中最基礎的工作項目之一，藉由更精準的計算地震發生的位置和時間，可以成為其它相關研究中重要的參考依據。「雙差分地震定位演算法」是眾多地震重定位方法的其中之一。被廣泛應用於研究的雙差分地震定位程式hypoDD（Waldhauser and Ellthworth, 2000）僅支援使用一維層狀速度模型進行定位，對於測向波速變化較為複雜的區域，定位結果的改善較為有限。另外因為雙差分地震定位演算法本身的限制，對於地震群中較外圍的孤立地震，將無法得到其定位結果。本研究所使用的三維雙差分地震定位法（3D-DD）使用三維球面座標之虛擬曲折波線追跡法來計算理論走時，以得到更為精確的地震位置。本程式也將雙差分定位法與單事件定位法進行聯合逆推，可在進行定位的同時保留孤立的地震資料。我們以合成資料測試檢驗定位法之表現，測試結果顯示3D-DD能得到比hypoDD及單純的三維地震定位更加精確的定位結果。我們亦將此地震定位法應用於2013年10月31日於台灣花蓮瑞穗發生之地震序列之重新定位。以三維雙差分定位法重定位的結果和地震目錄中原先的震源位置相比，震源分布顯得較為密集、線性，而走時殘差、定位誤差也明顯降低。以重定位之震源位置配合台灣地區速度構造討論之結果，瑞穗地震序列應為呂宋弧前地殼之構造運動，於其西側邊界所產生之地震活動相關。

Double-difference algorithm is one of the methods used in earthquake relocation. The widely used double-difference relocation program is hypoDD (Waldhauser and Ellsworth, 2000; Waldhauser, 2001). It only supports using layered velocity model in the relocation process. To improve the relocation result by using higher resolution velocity model, we developed a double-difference earthquake relocation program using three-dimensional velocity model (3D-DD). Moreover, we also joint single-event relocation method with double-difference location algorithm during the inversion procedure, in order to prevent losing the relocation results when double-difference location algorithm is only used. The performance of the program was tested by a series of synthetic data tests, and the location error of 3D-DD is smaller than the error of hypoDD and a three-dimensional single event earthquake relocation program. We also apply our method on the relocation of Ruisui, Taiwan, earthquake sequence occurred on October 31st, 2013. The relocated hypocentral distribution is much more concentrated and linear. The location error and residual also became smaller after relocated by this method. By relocated hypocenters and velocity structure of Taiwan area, the cause of Ruisui earthquake may be the tectonic activity of the Luzon forearc crust.