COPULA-BASED CORRELATION ANALYSIS OF INTENSITY MEASURES OF MAINSHOCK-AFTERSHOCK GROUND MOTIONS

This study selects 662 mainshock-aftershock (MS-AS) ground motions from PEER NGA-West2 ground motion database, to analyze the correlation of the intensity measures (IMs) among the MS-AS ground motions. The correlation coefficients of these IMs among the MS-AS ground motions are calculated, and their optimal probability models are determined according to the K-S test, the AIC criterion and the BIC criterion. Meanwhile, the AIC criterion and the BIC criterion are used to determine the optimal copula functions among the MS-AS ground motion IMs, and their joint distributions are built based on the copula functions. On this basis, the conditional distribution and the conditional mean of the aftershock ground motion IMs are obtained, given those of the mainshock. The results show that: the significant duration has the highest correlation amongst the 34 selected IMs, the joint distributions can be built using the Copula function with reasonable accuracy, and the Copula conditional mean can be used to predict the IMs of the aftershock ground motions, given the IMs of the mainshock ground motions.