| dc.description.abstract | Taiwan is located in the Circum-Pacific seismic belt, which is a continental marginal island caused by the collision between the Eurasian Plate and the Philippine Ocean Plate. Under the interaction of the plates, there are a number of earthquake events in Taiwan every year. In history, there have also been many massive earthquakes that caused economic losses and casualties in Taiwan, so the earthquake is a very important issue in Taiwan. However, up to now, it is still impossible to predict the location, time and scale of earthquakes as accurately as the earthquake. Even so, the earthquake early warning system (EEW) is the most effective method to provide real-time earthquake warning information.
Earthquake early warning system most often uses intensity and PGA as early warning index, but the standard drop of intensity classification is too large to be accurate in judgment, and PGA is not as good as CAV in showing the damage degree of structures. Therefore, this study establishes three kinds of CAVSTD-GMPE models suitable for Taiwan: log method, ln method and new log method, and applies them to Taiwan′s earthquake early warning system.
The results of the CAVSTD-GMPE model established and earthquake early warning application in this study are as follow: (1) the new log method is recommended for the prediction accuracy of ground motion parameters, and (2) for the credibility of earthquake early warning, the log method is recommended. (3) the ground motion prediction equation established by regression analysis with maximum accuracy is better than or equal to the ground motion prediction equation established by ordinary least square method in earthquake early warning. | en_US |