Nicholas C. Davatzes*, Peter Eichhubl**, and Atilla Aydin

Department of Geological and Environmental Sciences, Stanford University, Stanford, CA 94305-2115

* Present address: Earthquake Hazards Group, U.S. Geological Survey, 345 Middlefield Road, M. S. 977, Menlo Park, CA 94025
** Present address: Texas A&M University-Corpus Christi, Dept. of Physical and Life Sciences, 6300 Ocean Drive, Corpus Christi, TX 78412

Faults in sandstone are frequently composed of two classes of structures: (1) deformation bands and (2) joints and sheared joints. Whereas the former structures are associated with cataclastic deformation, the latter ones represent brittle fracturing, fragmentation, and brecciation. We investigated the distribution of these structures, their formation, and the underlying mechanical controls for their occurrence along the Moab normal fault in southeastern Utah through the use of structural mapping and numerical elastic boundary element modeling. We found that deformation bands occur everywhere along the fault, but with increased density in contractional relays. Joints and sheared joints only occur at intersections and extensional relays. In all locations, joints consistently overprint deformation bands. Localization of joints and sheared joints in extensional relays suggests that their distribution is controlled by local variations in stress state that are due to mechanical interaction between the fault segments. This interpretation is consistent with elastic boundary element models that predict a local reduction in mean stress and least compressive principal stress at intersections and extensional relays. The transition from deformation band to joint formation along these sections of the fault system likely resulted from the combined effects of changes in remote tectonic loading, burial depth, fluid pressure, and rock properties. In the case of the Moab fault, we conclude that the structural heterogeneity in the fault zone is systematically related to the geometric evolution of the fault, the local state of stress associated with fault slip, and the remote loading history. Because the type and distribution of structures affect fault permeability and strength, our results predict systematic variations in these parameters with fault evolution.