本研究以地震資料豐富且地動場址特性已清楚之宜蘭平原作為測試場，利用微地動單站頻譜比之比值（RHVSR）作為地動波形之場址修正因子，藉由強震站實測波形資料進行場址效應修正，進而推估平原各微地動測點之地動波形，並探討可能影響此地動預估方法之各項因素。經測試之結果顯示，測站距離在18公里以內時，對地震動預估之影響不明顯，但位於地質邊界之測站可能受地下複雜之地質構造形貌影響，其誤差值明顯高於平原之測站。另外，根據震源距離測試之結果顯示對預估之影響不大。接著根據測試結果建立微地動測點與強震站間的匹配準則，並對所有微地動測點進行歷史地震之地動預估。結果顯示佈設在平原的微地動測點可反應平原之場址特性，相較於強震站，高密度之微地動測點能夠反應更細緻之場址變化，並呈現高精度之地動分布。根據地震主頻與對應頻率之場址放大分布之分析結果發現，不同地震之震波特性會導致震波進入平原各處之頻率內涵有所差異，預估之地動分布與場址放大分布相當一致，顯示本研究所預估之地震動具一定的可信度。針對不同歷史地震模擬結果進行觀察，發現不同位置之震源確實會影響平原各處對地震動之反應。對於震源位置相似之地震事件，平原內所表現的地震動分布相當類似。本研究透過對場址相依地動預估分析程序之測試與實際應用，顯示此方法確實可反應平原內各地區之場址特性，並可獲得可信之高精度地震動分布。;This study proposes a method by using microtremor H/V spectral ratio (HVSR) to correct the site effect and predict the ground motion of nearby sites. The ratio of HVSR ( RHVSR) can be seen as a correction factor to predict the ground motion of the sites around a strong-motion station. Here, we selected the Ilan plain of Taiwan as a test site, and evaluate the error and stability of this method. The results show that there is no significant effect of distances on the prediction of ground motion if the distance between two sites less than 18 km. But, the stations located at the geological boundary may be affected by the topography of the geological structure, and the error is obviously higher than that of the plain station. In addition, the test shows that the source distance has minor effect on the ground motion prediction. Based on the results of the test for our method, the matching rules between microtremor sites and strong ground motion stations are established, and the scenarios of several historical events in the Ilan plain were simulated. The scenarios proved that the microtremor sites distributed in the plain reflect the site characteristics well. Compared with the original observations of strong motion stations, the simulations of high-density microtremor sites can react to complex site characteristics and present high-accuracy ground motion distribution. It is found that the seismic characteristics of different earthquakes will lead to some differences in frequency contents of seismic waves entering the plain because of the variation of site amplification. The predicted ground motion distribution is consistent with the site amplification distribution, which shows that the predicted ground motion in this study has certain credibility. Based on the simulations of historical events, it is also found that the hypocenter in different locations will affect the response of ground motion in different places in the plain. Through the test and application of the ground motion prediction by using site correction of microtremor, this study shows that the method can obtain reliable high-accuracy ground motion distribution that reflect the site characteristics sufficiently.
