A COMPARATIVE STUDY ON THE OPTIMAL SAMPLING STRATEGIES OF MULTIVARIATE SPATIALLY AND TEMPORALLY CORRELATED GROUND MOTIONS

This paper introduces three methods for the optimal sampling of multivariate spatially and temporally correlated time history of ground motions representing the regional earthquake hazard, and a comparative study is conducted to evaluate the performances of the three strategies. In the field of regional seismic risk assessment, optimal sampling methods are used to reduce the number of required scenario-based analyses. Most of the existing optimal sampling methods are developed for the intensity measures (IM) such as Peak Ground Acceleration (PGA) or Spectral Acceleration (SA). However, these IMs can only partially represent the nature of the earthquake events. On the other hand, the complete time history of ground motions enables much more accurate assessment of the seismic performance of the structures in a region. To this respect, three optimal sampling methods are introduced for the effective generation of a relatively small set of ground motion time histories, which can preserve the stochastic properties of a significantly larger set of ground motions. Performance assessment with different metrics for the three strategies is presented to demonstrate the pros and cons of the three strategies quantitatively.