Maps of the macroseismic intensity and ground‐motion measures for historical events are often irregular; the contours of constant seismic intensity or of constant ground‐motion measures do not necessarily fall on a circle or an ellipse. This is the case, for example, for the 1999 Chi‐Chi earthquake (Shin and Teng, 2001), California earthquakes (Toppozada et al. , 2002), Italian earthquakes (De Rubeis et al. , 2005), Greek earthquakes (Schenkova et al. , 2007), the 2008 Wenchuan earthquake (Wang and Xie, 2010; Wang et al. , 2010; Yuan et al. , 2013), and the Tohoku earthquake (Goto and Morikawa, 2012). The ground‐motion records from historical events are used as the basis for developing ground‐motion prediction equations (GMPEs; Boore et al. , 2014). Often the adopted mathematical forms for the GMPEs are based on implicit assumptions that the ground‐motion measures such as the peak ground acceleration (PGA), spectral acceleration (SA), and Arias intensity ( I A) are independent of wave propagation path and/or focal mechanism and geometries (Douglas, 2011; Foulser‐Piggott and Stafford, 2012; Atkinson and Adams, 2013). This is a practical and useful approach for seismic hazard and risk assessments of buildings during their life‐cycles (Hong and Goda, 2006; Goda and Hong, 2008) because available records are insufficient to develop geographically or site‐dependent ground‐motion prediction models for ranges of focal mechanism and geometries. In fact, the GMPEs for randomly oriented horizontal ground‐motion measures are most often developed and used (Boore et al. , 2006; Douglas, 2011) because the GMPEs for the bidirectional horizontal excitations are rarely available and the directions of the principal axes (with respect to the source‐to‐site orientation) of the records from a broad network of recording stations are not very consistent (Hong and Goda, 2007, 2010). Some of the questions for seismic hazard and risk analysis that need to be …
Read full abstract