Normal fault interaction caused by coseismic and postseismic stress changes

We study coseismic and postseismic stress fields caused by a normal faultin g earthquake in a self-gravitating, stratified, viscoelastic spherical Eart h over distances from a few to hundreds of kilometers. We investigate the c ontribution of postseismic relaxation on the induced Coulomb stress for ext ensional tectonic settings accounting for the effects of the Earth stratifi cation. We use a numerical code based on the spherical self-gravitating Ear th model developed by Piersanti et al. [1995, 1997]. We study how postseism ic relaxation can modify the state of stress at the base of the seismogenic layer where large earthquakes are believed to nucleate. We compare our res ults with those obtained by means of a three-dimensional dislocation model in an elastic half-space, which does not account for the time-dependent pos tseismic stress transfer. The viscoelastic relaxation process modifies the coseismic stress changes during time periods from several decades to centur ies. The postseismic stress is generally greater than the coseismic stress change. Postseismic relaxation increases the Coulomb stress near the causat ive faults and tends to reduce the stress shadow areas. The temporal evolut ion of Coulomb stress reveals that in addition to the viscosity value, the thickness of the elastic layer controls the time at which the relaxation pr ocess is completed. A larger thickness of the elastic layer yields a faster relaxation in the first few decades after the seismic event but smaller po stseismic stress amplitudes at longer timescales. One of the most significa nt results of this study is the extreme sensitivity of the timescales of th e viscoelastic relaxation to small changes in the thickness and depth of th e shallowest viscoelastic layer as well as in variation of the viscosity. S uch a result suggests that the interpretation of the time evolution of the postseismic signals only in terms of viscosity values could lead to mislead ing conclusions.