What comes first: The fault or the ductile shear zone?

Abstract Crustal scale fault zones extend below the brittle-ductile transition as ductile shear zones. Here we address the question of which regime, brittle or ductile, initiates and controls the overall system of shear localization. Observations of crustal scale but low displacement conjugate strike-slip faults show that they are typically nearly orthogonal, as expected from plastic shear criteria. Sub-crustal scale conjugate strike-slip faults, however, have acute dihedral angles in accordance with the Coulomb fracture criterion as Anderson's theory predicts. We modeled the crustal scale system with strain weakening rheologies that follow the Coulomb and von Mises criteria respectively, within the brittle and ductile regimes. We find that when the strain weakening rate in the ductile regime exceeds a critical value the entire system shears in the von Mises mode with orthogonal conjugate shears forming at all depths, in accordance with the observations. There is, as observed, no deflection of fault orientation at the brittle-ductile transition. Anderson's theory of faulting thus breaks down for crustal scale faults. This difference in behavior was already evident in data presented in Anderson's 1951 book, but its significance was not understood at that time.