Intensity measures considering directivity of near-fault ground motion and optimal selection for bridges -- case study in the Wushi earthquake

This paper proposes omnidirectional ground motion intensity measures (OGMIMs) to intuitively characterize two-dimensional ground motions, using the magnitude 7.1 Wushi earthquake as a case study to investigate the directional characteristics and spatial distribution patterns of near-fault ground motions. By identifying the intensity measures (IMs) in orthogonal extreme-value directions with OGMIMs, probabilistic seismic demand models (PSDMs) in the form of vector-valued IMs [IM max, IM min] are constructed and compared with those based on scalar-valued IMs. The findings indicate that for reverse dip-slip faults, the predominant direction of the maximum acceleration-based OGMIM, namely peak ground acceleration (PGA) and compound acceleration-based IM (I a ), aligns with the fault slip direction on the horizontal plane, while velocity-based intensity measures, including peak ground velocity (PGV), the composite velocity-based index (I v ), and the multi-parameter index (PGV/PGA), are consistent with the strike-normal direction of the fault. PSDMs constructed with vector-valued IMs exhibit higher goodness-of-fit than scalar-based ones. Moreover, velocity-based IMs represented by I v and the root-mean-square velocity (v rms ) are most suitable for simply supported and continuous girder bridges. Analysis of input angles indicates that the optimal fitting direction aligns with the fault, whereas the optimal direction for predicting girder displacement runs parallel to it.