P-WAVE AND S-WAVE DISPLACEMENTS FROM KINEMATIC DISLOCATION MODELS

Numerical dislocation models based on Haskells (1969) formulation are used to estimate the amount of normal motion necessary to produce the high P/S spectral ratios observed from strong-motion records of the Guerrero, Mexico, subduction zone (Castro et al., 1991). While this depends on the nature of the assumed dislocations, a normal motion with amplitude of less than 10% of the amplitude of shear slip is sufficient to produce P/S values comparable with the observations. We model a planar fault with random patches distributed on the fault plane and a nonplanar fault in which a dilatational jog connects en-echelon fault segments (a structural system proposed by Sibson, 1985, 1989, which introduces normal motions on the fault that depend mainly on the geometry of the fault). For the planar fault model the magnitude of the normal motion is prescribed. Both models introduce complexity in both body-wave displacement time histories, although for the P waves this complexity is accentuated at higher frequencies (f > 1 Hz). For the nonplanar model, a jog with an angle of 10 introduces a normal component of 18% of the slip on the fault.