Sensitivity study of near-source ground motion

We studied the sensitivity of near-source ground motions for hypothetical events on a thrust fault (M_w 6.6 to 7.0) and a strike-slip fault (M_w 7.0 to 7.1) to five earthquake source parameters. We systematically varied the rupture speed, maximum slip rate, hypocentre location, distribution of final slip, and fault depth. We used the finite element method to discretize a homogeneous or layered half-space into an unstructured mesh to model the wave propagation in the domain surrounding the fault. Our sensitivity study of near-source ground motion indicates it is very important to include directivity effects when modelling near-source ground motion. In the thrust fault scenarios a double velocity pulse sweeps along the surface in the direction of the propagating rupture. For most of the scenarios the peak velocity, filtered to periods longer than 2.0 sec, exceeds 1.0 m/sec over an area of 100 square kilometres. In the strike-slip scenarios a complex series of pulses involving the shear wave and Rayleigh waves propagates in the direction of the rupture with the most severe motion confined to a narrow region along the fault. The peak, filtered velocity exceeds 1.0 m/sec over an area of 700 square kilometres. We found the ground motions strongly sensitive to the material properties and fault depth, moderately sensitive to the hypocentre location, rupture speed, and maximum slip rate, and relatively insensitive to the distribution of final slip. The shape of the near-source factor, N_v, from the 1997 Uniform Building Code does not correlate with the zone of severe shaking in the case of blind thrust faults, because the maximum displacements and maximum velocities tend to occur up-dip from the top of the fault