Viscoelastic earthquake cycle effects and the kinematics of Northern Tibet

Viscoelastic relaxation of the mid- to lower crust during the earthquake cycle can cause large changes in interseismic surface velocities when the relaxation time of the mid- to lower-crust is short relative to the earthquake recurrence time. These earthquake cycle effects have been proposed as a means of explaining discrepancies between fault slip rates inferred from GPS velocities over decadal time-scales and those estimated over thousands of years along the Altyn Tagh Fault (ATF). In this contribution, we explore a range of simple deformation models to determine if this process may explain this discrepancy. In particular, we interpret geodetic velocities throughout the greater Tibetan Plateau using three different models: (1) two-layer, three-dimensional models that consider the effects of homogeneous strain and viscoelastic earthquake-cycle effects, (2) block rotation, and (3) block rotation with a contribution from viscoelastic earthquake-cycle effects. In the earthquake cycle formulation, the mid- to lower-crust is idealized as a linear Maxwell material that deforms elastically over short time-scales but viscously over long time-scales. We find that the geodetic velocities are best explained by block rotation, rather than homogeneous strain with superposed earthquake- cycle deformation. In the case that homogeneous strain and earthquake-cycle effects are considered, the ATF may permissibly slip left-laterally at a rate of 9.4 - 16.2 mm/yr, and should be experiencing ~0.7 mm/yr of contraction, while the Kunlun Fault should resolve 15.0 - 16.2 of left-lateral slip per year. However, when block motion is assumed, the ATF may only accommodate between 8.8 - 10.8 mm/yr of left-lateral motion, while the Kunlun Fault should accommodate 15.3 - 15.7 mm/yr of left-lateral motion. We are currently exploring models that consider both block motion and earthquake-cycle effects with Bayesian statistical methods that allow prior information, such as the timing of paleoearthquakes and bounds on their recurrence intervals, to be used when inferring crustal properties and earthquake-cycle parameters from observed geodetic velocities. In addition, we are also exploring three-dimensional, three-layer models to determine if geodetic data permit a confined, low-viscosity zone in the mid- to lower-crust of Tibet to be present. Using the Bayesian methods with models that consider earthquake cycle effects, we will determine if earthquake cycle effects may possibly reconcile the slip-rate discrepancy between geodetic and geologic time-scales along the ATF.