1909台北歷史地震之地動模擬

Ground motion simulation of 1909 Taipei historical earthquake

廖怡雯(Yi-Wun Liao)

指導教授 ： 馬國鳳 ； 李憲忠

國立中央大學/地球科學學院/地球科學系/碩士(2014年)

摘要

1909年4月15日於台北地區深處發生一規模7.3之深源地震，稱為1909台北地震。雖然此地震在當時所造成之傷亡不大，但若一類似事件再發生對於現今台北地區所造成之衝擊仍須被關心與注意，因此本研究使用數值模擬方法模擬1909台北地震，並探討不同震源機制、不同目標規模 (Mw6.8與Mw7.3) 以及不同震央位置之模擬結果在台北地區所產生之地表震動之差異。
本研究首先使用譜元素法 (Spectral element method) 模擬1909台北地震對台北地區所造成之地表震動。根據前人研究所重定位之震源位置以及其所提出1909台北地震可能之震源機制解，我們假設此地震有兩種破裂模式：點震源破裂模式與有限斷層破裂模式。在點震源破裂的假設之下，伸張型震源機制解 (δ= 20°, λ= -90°, φ= 90°) 較隱沒帶類型震源機制解造成較高之地表震動，而兩者皆在台北盆地西部有較高之PGV (Peak ground velocity) 值分布。而有限斷層假設之模擬結果，相較於擁有相同震源機制解(伸張型震源機制解)之點震源破裂假設之結果，在台北盆地內PGV高值分布較廣且有較長震動時間。
除了使用譜元素法模擬台北地震可能造成的地表震動外，本研究亦使用混合式的模擬方法，在分別使用譜元素法以及經驗格林函數法 (Empirical Green’s function method) 模擬低頻與高頻部分地震波後，將兩者結合得到寬頻段地震波資訊。本研究選擇2013年10月11日發生在台北地區之下，芮氏規模4.9之深源地震 (深度143.8公里) 作為經驗格林函數，將寬頻測站網之紀錄分別由芮氏規模4.9 (中央氣象局) 和地震矩規模4.2 (RMT：Real-time moment tensor monitoring system) 修正至目標規模。台北測候所之地震報告早期地震測站台北(TAP)與基隆(KEE)站之觀測PGA (Peak ground acceleration) 分別為59.3以及67.0 gal。而本研究模擬結果顯示，模擬目標規模為Mw7.3且應力降較高(100 bars)之假設之模擬結果(TAPB/TAP：59.9 gal；WFSB/KEE：140.4 gal)與文獻紀錄之觀測值較符合，1909台北地震之規模可能高達Mw7.3。在分析地動值與地動搖晃時間之結果中，台北盆地內之測站(TAPB)地動值並非最高，但卻有較長之搖晃時間。

關鍵字

1909台北地震 ； 譜元素法 ； 經驗格林函數法 ； 混合式模擬法

並列摘要

The 1909 Taipei earthquake (M6.8) occurred beneath the Taipei Metropolitan Area (TMA), the possible seismic hazard similar to the 1909 Taipei earthquake would require special attention. According to the hypocenter relocated from historical archived travel times of the 1909 Taipei earthquake by Kanamori et. al (2012), we simulated ground motions in TMA by using Spectral Element Method (SEM). A 3D velocity structure with topography and sediment layers in Taipei basin was also included in the simulation.
As the results of point-source simulations, the tension type mechanism may cause larger ground motions than the subduction zone type mechanism. The higher peak-ground-motion values occurred in the western part of Taipei basin with both two types of focal mechanisms. This distribution might correspond to the thicker sediment layers in the western part of the basin. The finite fault simulation showed that the high PGV area spread wider than the results from point-source assumption, and the coda waves became more obvious and caused longer shaking duration at station TAP than point-source assumption with the same focal mechanism.
SEM could provide us good prediction of low frequency ground motions, however, for real earthquakes, waveforms consist of broad frequency band signals. In order to obtain the information in a broad frequency band we simulated waveforms with hybrid method, which is a combination of SEM and empirical Green’s function method (EGFM). For higher frequency components, we chose a recently happened deep event (2013/10/11 M4.9 Depth:143.8 km) in Taipei area as the empirical Green’s function for the 1909 Taipei earthquake. We corrected the amplitude of the broadband records with hypocentral distance from depth 143.8km to 75km and modified the magnitude from ML4.9 (CWB) and Mw4.2 (RMT) to target magnitude (Mw6.8 and Mw7.3) by using EGFM. An east-dipping (δ=344°, λ=57°, φ=103°) finite fault was considered in the SEM simulation to obtain the lower frequency components. We also moved the hypocenter to the relocated hypocenter from Kanamori et al. (2012) and simulated both higher and lower frequency components.
By combining the time series of lower (SEM) and higher (EGFM) frequency components, the model of modifying the magnitude from Mw4.2 to target magnitudes (Mw6.8 and Mw7.3) has more reasonable PGA values and response spectra curves than from ML4.9 to target magnitude (Mw6.8 and Mw7.3). The observed PGA values from literature are 59.3 gal at station TAP and 67.0 gal at station KEE. According to the simulation results, the PGA values (TAPB/TAP：59.9 gal；WSFB/KEE：140.4 gal) from the model with target magnitude Mw7.3 and higher stress drop are closest to the observed values. For the analysis of strong shaking durations and PGA values, the station inside Taipei basin (TAPB) has lower PGA values but longer strong shaking durations.

並列關鍵字

1909 Taipei earthquake ； Spectral element method ； Empirical Green's function method ； Hybrid simulation method

參考文獻

楊世緯，2012，利用經驗隔鄰函數法探討板塊介面及內部地震對台灣北部地區強地動之影響：中央大學地球科學系，碩士論文.

顏銀桐，2011，台灣地區有限斷層錯動量分布尺度分析及模擬：國立中央大學地球科學系，博士論文.

Reasenberg, P. and Oppenheimer, D., 1999. FPFIT: http://earthquake.usgs.gov/research/software/.

Aki, K., 1967. Scaling law of seismic spectrum: Journal of Geophysical Research, 72, 1217-1231.

Chen, K., Huang, W., and Wang, J., 2007. Relationships among magnitudes and seismic moment of earthquakes in the Taiwan region: Terrestrial Atmospheric And Oceanic Sciences, 18, 951-973.