3-D finite-element modelling of deformation and stress associated with faulting:: effect of inhomogeneous crustal structures

Modelling of deformation and stress caused by earthquake faulting (or fault motion) based on a homogeneous elastic half- space model does not include crustal rigidity layering and lateral variations. Geological data and seismic surveys indicate that the crust is elastically inhomogeneous, and that rigidity layering and heterogeneities are likely to affect the magnitude and pattern of deformation and stress. We use the finite- element method (FEM) to investigate deformation and stress generated by strike- slip, tensile and thrust faulting, considering rigidity layering in the crust. For a fault with a dimension of 25 x 10 km and a slip value of 0.5 m cutting through the top 10 km of the crust, the deformation and stress change associated with the three modes of faulting are calculated for an island arc crustal structure represented by three models of rigidity layering. The results indicate: (i) The maximum effect in horizontal displacements due to crustal layering for the investigated models is 19.7 mm and 28.8 mm for strike- slip and tensile faulting, respectively. For thrust faulting, the effect is up to 58.3 mm. This remarkable effect is due to the offset of layers by thrust faulting, which dynamically changes the crustal structure both horizontally and vertically. (ii) The maximum effect in vertical deformation due to crustal rigidity layering for strike- slip and tensile faulting is 5 mm. In contrast, the effect for thrusting is 30.3 mm. (iii) The effect in the shear stress change at a depth of 7.5 km for the different styles of faulting ranges from 0.02 MPa to 0.10 MPa. An additional modelling analysis of subduction thrusting at the Japan trench with a detailed structural model based on seismic surveys further confirms the significant distortion caused by a homogeneous model that ignores crustal structural information. Overall, the effect of crustal rigidity layering is at least 20 per cent for surface deformation and 15 per cent for stress changes at depth. In light of the present accuracy of global positioning system (GPS) observations ( +/- 2 mm) and the estimate of stress change required for earthquake hazard analysis (ca 0.01 MPa), the results from the numerical experiments indicate that crust inhomogeneity (rigidity layering and lateral variations) should be included in any quantitative models of faulting.