• journal_title：Bulletin of the Seismological Society of America
• Contributor：Kun‐Sung Liu ; Yi‐Ben Tsai ; Po‐Shen Lin
• Publisher：Seismological Society of America
• Date：2013-06-01
• Format：text/html
• Language：en
• Identifier：10.1785/0120120065
• journal_abbrev：Bulletin of the Seismological Society of America
• issn：0037-1106
• volume：103
• issue：3
• firstpage：1823
• section：Articles

The purpose of this study is to derive attenuation relationships of the peak ground acceleration (PGA) and peak ground velocity (PGV) in Ilan, Taiwan, including factors of fault type and site effect. A total of 2852 accelerograms recorded from 92 shallow earthquakes with M_w magnitude ranging from 4.0 to 7.7 are used to develop the regional attenuation relationships. Two models are used: Model 1 uses records from 65 strong‐motion sites. The results will be especially useful for the Central Weather Bureau (CWB) early warning system to quickly assess and report the PGA and PGV maps for effective emergency‐response operations. Model 2 uses records from the 46 strong‐motion sites with available V_S30 to incorporate a site‐effect term, aiming to reduce the standard deviation of the predicted ground motion for engineering applications. First, the results show that the fault‐type amplification factor decreases faster with increasing magnitude for the strike‐slip fault type than the reverse or normal fault types. In addition, from the intraevent residuals as function of distance of horizontal PGA for reverse earthquakes, we found clusters of higher PGA values at a distance of 110–130 km from events in southwestern areas. This was probably due to Moho reflection along paths crossing the Central Mountain Range. Second, the PGV intraevent residual is more closely correlated with V_S30 than that of the PGA because the latter is not a simple function of V_S30. Finally, by comparing the standard deviations of the total residuals between the observed and predicted values before and after incorporating the fault‐type and site‐effect (V_S30) terms, the change of standard deviation for PGA is 2.3%. In contrast, the PGV standard deviation is significantly decreased by about 11.6%. Evidently, the Model 2 attenuation relationships, especially for PGV, would be better suited for engineering applications.