Reverse migration of seismicity on thrusts and normal faults

In this work, the control exerted by the stress axes orientation on the evolution of seismic sequences developing in compressive and extensional regimes is analysed. According to the Anderson fault theory, the vertical stress is the minimum principal stress in compressional tectonic regimes, whereas it is the maximum principal stress in extensional regimes. Using Mohr diagrams and discussing the present knowledge about the distribution of vertical and horizontal stress with depth we show that, in absence of localised fluid overpressure, such changes imply that thrust and normal faults become more unstable at shallower and greater depths, respectively. These opposite mechanical behaviours predict, in a rather isotropic body, easier rupture at shallower level in compressional regimes later propagating downward. On the contrary, a first deep rupture propagating upward is expected in extensional regimes. This is consistent with observations from major earthquakes from different areas in the world. We show that the exceptions to downward migration along thrusts occur along steeply inclined faults and probably imply localised supra-hydrostatic fluid pressures. Moreover, we show that the inversion of the meaning of the lithostatic load has consequences also for the role of topography. High topography, increasing the vertical load, should inhibit earthquake development in compressional environments and should favour it in extensional settings. Although several factors, such as geodynamic processes, local tectonic features and rock rheology, are likely to control earthquake locations, stress distribution and tectonic regime, these model predictions are consistent with seismicity distribution in Italy, central Andes and Himalaya. In these areas, large to medium compressional earthquakes occur at the low elevation borders of compressional mountain belts, whereas large extensional earthquakes occur in correspondence to maximum elevations.ASI 2001, Cofin 2001 and GNDT fundingsPublished195-222partially_ope