Evidence of Powerlaw Flow in the Mojave Desert Mantle

Laboratory experiments suggest that rocks in the lower crust and upper mantle (shallower than ∼200 km) should deform by dislocation creep, leading to a powerlaw rheology characterized by a strain rate proportional to stress raised to a power, n. Powerlaw creep has not yet been confirmed by geodetic observations. We use GPS campaign and continuous time-series data associated with 1992 Landers and 1999 Hector Mine earthquakes to infer rheologic properties of the Mojave lithosphere in southern California. The coupled nature of these earthquakes (20 km and 7 years apart) makes them ideal for a stringent rheology study in that a candidate rheologic model must satisfy the postseismic observations associated with both events. To infer the nature of viscous flow we developed a finite element model of this earthquake sequence that simulates coseismic slip associated with both events, a regional background strain rate, and temperature dependent powerlaw rheologies. We consider a range of powerlaws (for felsic and mafic, wet and dry rocks) reflecting uncertainty in the mineralogy of the lithosphere and in the extrapolation from laboratory to geologic conditions. Thermal gradients are constrained from surface heat flow measurements and regional seismic velocities. For comparison purposes, we also consider models with a Newtonian (strain rate linearly proportional to stress) rheology. Our results show that the spatial and temporal evolution of transient surface deformation following the Landers and Hector Mine earthquakes can be successfully explained by powerlaw flow (n = 3.5), predominantly in a warm and wet upper mantle. We can rule out Newtonian flow as a reasonable explanation of both the spatial and temporal patterns of postseismic transient motions, implying that the common assumption of Newtonian flow in numerical models of ductile deformation within the crust and upper mantle may be invalid. These results suggest that recovery-controlled dislocation creep is the dominant mechanism of viscous flow following earthquakes. The model results also preclude significant flow in the lower crust, supporting the contention that, at least beneath the Mojave Desert, the mantle is the weaker region. The stress dependence of powerlaw flow inferred by our calculations means that the viscosity of the upper mantle changes as a function of time after an earthquake. This has implications for models of regional stress changes and fault interaction. For example, the influence of earthquake induced stress changes on neighboring faults will evolve more rapidly early on, but will last many decades longer than would be inferred from a Newtonian model. Furthermore, as viscosities are lowest where stresses are highest, a powerlaw rheology leads to a more localized shear zone beneath faults where coseismic stresses are highest. For example, our calculations show that a Newtonian model of post-Landers relaxation leads to a broad, diffuse shear zone in the mantle beneath the Landers rupture zone about 250-300 km wide. In contrast, post-Landers relaxation of a powerlaw rheology leads to a relatively narrow shear zone 70-90 km wide, with much of the shear concentrated in a central zone only 15 km wide.