Slip system domains, 3. New insights in fault kinematics from plane-strain sandbox experiments

Abstract

This paper describes experiments on plane-strain deformation of sandpacks for different imposed bulk kinematic conditions varying between coaxial stretching and strike-slip simple shearing. Arrays of conjugate faults start to develop after about 10% uniform strain and accommodate most of subsequent deformation. Initial fault orientations seem to reflect attitudes of bulk principal stresses, according to the Coulomb criterion. Conjugate faults are distributed into domains where one fault set is largely dominant if not present alone. Boundaries of individual domains are incoherent, parallel to one fault set. Curved faults accommodate jumps in rigid rotation across domain boundaries. Dominoes which are synthetic with respect to the imposed bulk simple shearing component, always rotate towards the bulk shearing plane; whereas antithetic dominoes always rotate away from the bulk shearing plane. Domino boundaries can migrate after substantial (~30°) rotation. Block and fault rotations can be large, especially where bulk deformation is non-coaxial (for example, antithetic dominoes rotate up to around 35° for a bulk simple shear of γ = 0.85). Concurrently, fault orientations deviate rapidly from stress-controlled to strain-controlled orientations, major long-lived faults tending to track orientations of no finite extension of the bulk strain ellipsoid. An increase in the component of bulk simple shearing tends to produce large domino domains where one family of faults is dominant. Although restricted to horizontal plane strain, our experiments produce fault patterns which show basic similarities with those occurring in the brittle crust. Following ideas developed in two previous papers of this series, we emphasize that strong similarities exist between fault patterns and other deformation patterns involving localized slip or shear, such as kink-bands, lattice fabrics, or shear zone patterns, and that these similarities occur because of simple kinematic factors. We argue that geometric analysis of natural fault patterns could provide basic information about associated crustal kinematics, including bulk strain magnitude, orientation of bulk principal strains, and bulk deformation regime.