Haibin Yang, Louis N. Moresi , Mark Quigley

The San Andreas Fault (SAF) is one of the dominant components of the transform boundary between the Pacific and the North American Plate. Although the fault is verti-cal-to sub-vertical at shallow (<10 km) depth, it variably dips at angles of ca. 40-70º to the southwest near the western Transverse Range and to the northeast in its southern seg-ment at depths of ca. 10-20 km, and thus can be described as having a listric geometry at the crustal scale. The mechanism controlling the fault dip direction variation at depth along SAF is not well understood. We utilize a 3D, finite element thermomechanical, vis-coplastic model to simulate lithospheric deformation associated with the SAF. The Big Bend of the fault near the western Transverse Range is taken as a geometrical initial condition. Numerical experiments demonstrate that regional lower crust strength variation along the SAF strike is an important control on fault dip direction. For two blocks separat-ed by transpressional faults, viscous lower crustal material moves from the high viscosity (strong) block to the lower viscosity (weak) lower crustal block. Fault-plane-normal flow in the viscous lower crust forces fault dip direction at brittle-ductile transitional depth to rotate in the flow direction. Geophysical data suggest that the Great Valley (eastern block) and south coast area (western block) have stronger lower crust than their opposing fault blocks and that the SAF dips towards the weaker block in both instances. Our self-consistent model also sheds light on the left-lateral Garlock Fault, which intersects the right-lateral SAF in the western Transverse Range. To maintain a vertical dip in the shal-low crust and structural-kinematic connectivity with the migrating fault zone at visco-elastic depths, the SAF may exhibit spatiotemporal transience in upper crustal fault loca-tion and geometry over geological time scales, including strain zone widening and dis-persions. This behavior is expected to have ramifications for SAF earthquake behaviors including rupture nucleation locations and segmentation and adds complexity to expecta-tions that strain should localize on sub-vertical strike-slip faults with increasing fault ma-turity.