Using short-term postseismic displacements to infer the ambient deformation conditions of the upper mantle

To interpret short-term postseismic surface displacements in the context of key ambient conditions (e.g., temperature, pressure, background strain rate, water content, creep mechanism), we combined steady state and transient flow into a single constitutive relation that can explain the response of a viscoelastic material to a change in stress. The flow law is then used to investigate mantle deformation beneath the Eastern California Shear Zone following the 1999 M7.1 Hector Mine earthquake. The flow law parameters are determined using finite element models of relaxation processes, constrained by surface displacement time series recorded by 55 continuous GPS stations for 7 years following the earthquake. Results suggest that postseismic flow following the Hector Mine earthquake occurs below a depth of similar to 50 km and is controlled by dislocation creep of wet olivine. Diffusion creep models can also explain the data, but require a grain size (3.5 mm) that is smaller than the inferred grain size (10-20 mm) based on the mantle conditions at these depths. In addition, laboratory flow laws predict dislocation creep would dominate at the stress/grain size conditions that provide the best fit to diffusion creep models. Model results suggest a transient creep phase that lasts similar to 1 year and has a viscosity similar to 10 times lower than subsequent steady state flow, in general agreement with laboratory observations. The postseismic response is best explained as occurring within a relatively hot upper mantle (e.g., 1200-1300 degrees C at 50 km depth) with a long-term background mantle strain rate of 0.1-0.2 mstrain/yr, consistent with the observed surface strain rate. Long-term background shear stresses at the top of the mantle are similar to 4 MPa, then decrease with depth to a minimum of 0.1-0.2 MPa at 70 km depth before increasing slowly with depth due to the pressure dependence of viscosity. These conditions correspond to a background viscosity of 10(21) Pa s within a thin mantle lid that decreases to similar to 5 x 10(19) Pa s within the underlying asthenosphere. This study shows the utility of using short-term postseismic observations to infer long-term mantle conditions that are not readily observable by other means.