| 摘要: | 利用地震儀接收地震所造成的地表運動情形，並利用震源逆推的方法可以計算地震的資訊，包含了地震的位置、規模大小、震源機制等，若能更進一步在地震發生後利用近場資料快速解算這些資訊，將能對破壞性地震波尚未到達的遠場區域做地震預警的準備，或對沿海的海嘯預警。
快速決定地震規模的方法為測量某頻段的地震波振幅值，然而對於大型地震，斷層需要更多時間釋放更大的能量，因此地震波高頻段的振幅將受到限制，在地震大於某規模後逐漸呈現定值，稱為地震規模的飽和（saturate）。反之低頻段的地震波較不受此限制，因此為了避免規模的飽和以得到正確的規模，有效記錄大地震的低頻地表運動訊號是不可缺少的重要關鍵。然而受限於地震儀對低頻訊號的不敏感，再加上大地震可能造成近場的地震儀發生超格（clipping）的問題，因此，利用近場地震資料快速解算震源參數，對大地震而言仍有許多困難。
全球衛星定位系統（Global Positioning System, GPS），提供了大地震近場的地表運動資訊，其低頻記錄的優勢克服了地震儀對低頻地表運動記錄不敏感的問題，另一方面，GPS的記錄沒有超格的問題，因此能夠有效地記錄大地震所造成的地表運動情況。
為了能在大地震發生後穩定的解算震源機制以評估可能造成的災害，本研究建立以GPS靜態位移資料為主的震源機制逆推模型。並使用實際GPS資料逆推2013年發生於台灣的三起地震，分別為2013/03/27 ML 6.2與2013/06/02 ML 6.5的南投地震，與2013/10/31 ML 6.4的瑞穗地震。另外，本研究還分析了兩個規模較大的地震事件，分別為發生於2002年花蓮外海的331地震與2003年的成功地震。所有結果皆與各地震單位計算結果相符，顯示簡化地表運動過程的GPS靜態位移觀測量有足夠能力還原震源機制。甚至，基於GPS對低頻訊號有好的攫取能力，以及沒有記錄超格的問題，我們預期此方法能更穩定的提供大地震的真實震矩規模。
;Inversion of the earthquake source information by using seismic data becomes effective and performs well in the last decades. Such information, including source location, earthquake magnitude and focal mechanism, are important for earthquake and tsunami early warning studies.
When earthquakes occur, a preliminarily calculation for earthquake magnitude is based on the amplitude of short period seismic waves in the near-field. However, for larger earthquakes, the energy need more time to release, which constrain the amplitude of high-frequency energy to a constant. Such a phenomenon is called magnitude saturation, which leads us underestimating the true magnitude of large and also devastating earthquakes. In order to calculate the magnitude closer to the real magnitude, immediately after an earthquake, the near-field and low-frequency information from the source are important and necessary.
Global Positioning System (GPS) takes advantage on the resolving power of low-frequency rupture behavior and overcomes the records clipping in the near-field, which is an ideal instrument for large earthquake. In order to calculate the focal mechanism, source location, and magnitude in a few seconds following an earthquake, this research builds a GPS-based focal mechanism inversion method for Taiwan. To test whether the model works properly, I test three local magnitudes ML >6 earthquakes occurred in 2013, Taiwan. Two earthquakes are of ML 6.2 and 6.5 occurred in Nantou County, in the central Taiwan with close epicentral locations within a distance about 8.5 km, namely the 0327 and 0602 earthquakes, respectively. Another earthquake occurred near Ruisui town, in the eastern Taiwan with ML 6.4, called the Ruisui earthquake. I also test two additional events for a larger occurred in 2002 of ML 6.8, called 331 earthquake; and the other occurred in 2003 of ML 6.4, called Chengkung earthquake. All the results showed good agreements with the results from GCMT and BATS, which evaluate the source parameters from dynamic wave information. Such success shows that GPS measurements alone can provide important information for inverting earthquake source model parameters. |
