| 摘要: | 場址效應是影響地震震度的因素之一，在地震危害度分析中亦是相當重要的部分。地表下三十公尺內之平均剪力波速（Vs30）是一個被廣泛用於評估地盤特性的參數，舉凡近年提出之地盤分類法或是地動預估式皆使用Vs30作為描述場址特性的參數。臺北盆地是臺灣人口密集且經濟活動頻繁的地區，然而其盆地的地形以及覆蓋其上的鬆軟沖積層使得重大地震發生時會因為場址效應產生更大危害。過往雖有研究探討臺北盆地之Vs30分布，然而隨著臺北盆地內的鑽井資料日益增加，有必要重新繪製臺北盆地之Vs30分布圖。
臺北盆地在進行Vs30推估時有著鑽井資料量多的優勢，可藉由經驗公式將鑽井資料轉換成剪力波速以獲得更多的Vs30資料。舊有的剪力波速經驗式大多以標準貫入試驗N值（SPT-N值）或深度做為主要影響參數。然而標準貫入試驗常受人為操作或設備影響，使得N值在使用上有標準不一致之問題。過去研究顯示，砂土之孔隙比、有效應力與剪力波速間有高度的相關性，因此本研究嘗試建立一新的剪力波速經驗式，並以孔隙比與垂直有效應力作為迴歸式之自變數，期望能以較低的推估誤差取代舊有的剪力波速經驗式。
本研究蒐集2000年至2004年分布於全臺灣的強震站鑽井資料（土層物性試驗資料與剪力波速測量資料）進行迴歸分析，建立可應用於臺北盆地礫石、砂、粉土與黏土之剪力波速經驗公式。過程中對剪力波速資料進行檢核，並依照材料是否屬於沖積層樣本及土壤類型進行篩選與分類，藉此降低剪力波速推估的不確定性。
本研究藉由此經驗公式對臺北盆地內數量眾多的工程鑽井進行各深度土壤之剪力波速推估，以獲得更多的Vs30資料。最後以Kriging with varying local mean方法對各鑽井之Vs30進行空間內插，完成臺北盆地Vs30分布圖。
研究結果顯示，本研究建議之砂、黏土與粉土剪力波速經驗式與前人研究相比有較低不確定性與較不偏估的優勢。在臺北盆地Vs30分布上，大部分地區之Vs30落在210m/s至300m/s之間，但在北投、士林、中山、松山、信義、大安等地區之Vs30低於210m/s，而在盆地西南側大漢溪流域（樹林、土城）以及盆地東南側新店溪流域（新店、文山、永和及中和區東側）等地區之Vs30較高，約在210m/s至440m/s之間。除此之外，前人研究認為中山、中正、大安與松山等區之Vs30小於180 m/s，然而隨著本研究將新的強震站資料加入分析，結果顯示此Vs30低區應不存在。
最後本研究計算臺北盆地內工程鑽井處因剪力波速推估方法造成的Vs30不確定性，結果顯示此類不確定性絕大多數落在50m/s至70m/s間，越靠近盆地邊緣與盆地南側因地質材料多為礫石或岩層，因此Vs30之不確定性較高。
;Site effect is one of the influential factors of seismic intensity. It also plays a crucial role in seismic hazard analysis. The average shear wave velocity of the upper 30 meters of a soil profile (Vs30) has been widely used for assessing site characteristics, such as site classification and ground motion prediction equation used Vs30 to describe the site conditions. Taipei Basin is an area with high population density and active economic activities in Taiwan. Site effect happen here which increase the earthquake disaster because of the basin topography and the overlying soft sediments. Although the distribution of Vs30 has been proposed by previous studies, as the borehole data gradually accumulates, it is necessary to develop a new Vs30 map in Taipei Basin.
There are a lot of borehole data in Taipei Basin. Those data can be converted into shear wave velocity(Vs) by empirical equations then get more Vs30 data. The existing empirical equations usually use Standard Penetration Test N value or depth as independent variable. However, Standard Penetration Test is easily affected by man-made operations or differences of equipment, which makes the N value inconsistent. Previous studies have pointed out that there is a strong correlation between void ratio, effective stress and Vs in sandy soil. Therefore, this study tried to establish new empirical equations which use void ratio and vertical effective stress as independent variable, and we expect the new equations can replace the old one with lower prediction error.
In this study, we constructed the Vs empirical equations for gravel, sand, clay and silt in Taipei Basin. The soil physical properties test data and Vs measurement data we used come from EGDT(Engineering Geological Database for TSMIP). In order to reduce the uncertainty of Vs estimation, the quality of Vs measurement data was checked and the soil data was separated by (1) whether the sample belongs to Holocene sediment; (2) classification of soil sample.
By using this empirical equation, the Vs of each depth of the numerous engineering borehole in Taipei Basin were computed to obtain more Vs30 data. Finally, the “Kriging with varying local means” method was applied to spatial interpolation and the distribution of Vs30 in Taipei Basin was mapped.
Compared with previous studies, the Vs empirical equation for sand, clay and silt suggested in this study have lower prediction uncertainty and less bias. Regarding the distribution of Vs30 in the Taipei Basin, the Vs30 at most of the area is around 210m/s to 300m/s. In Beitou, Shilin, Zhongshan, Songshan, Xinyi and Daan district, Vs30 is lower than 210m/s. In the southwest and southeast part of the basin, Vs30 is relatively higher, ranging from 210m/s to 440m/s. In addition, previous studies believe that there is a low-Vs30 area which Vs30 is lower than 180m/s in Zhongshan, Zhongzheng, Daan and Songshan. However, as this study adds new data to the analysis, the results show that this low-Vs30 area should not exist.
Finally, we calculated the Vs30 uncertainty caused by Vs estimation at engineering borehole site in Taipei Basin. The results show that this kind of uncertainties at most of the engineering borehole sites are 50m/s to 70 m/s. Nevertheless, there are higher uncertainties when borehole is located in southern basin or near the edge of basin, because most of the materials in these area are gravel and rock. |
