Variations in creep rate along the Hayward Fault, California, interpreted as changes in depth of creep

Variations in surface creep rate along the Hayward fault are modeled as cha nges in locking depth using 3D boundary elements. Model creep is driven by screw dislocations at 12 km depth under the Hayward and other regional faul ts. Inferred depth to locking varies along strike from 4-12 km. (12 km impl ies no locking.) Our models require locked patches under the central Haywar d fault, consistent with a M6.8 earthquake in 1868, but the geometry and ex tent of locking under the north and south ends depend critically on assumpt ions regarding continuity and creep behavior of the fault at its ends. For the northern onshore part of the fault, our models contain 1.4-1.7 times mo re stored moment than the model of Burgmann et al. [2000]; 45-57% of this s tored moment resides in creeping areas. It is important for seismic hazard estimation to know how much of this moment is released coseismically or as aseismic afterslip.