Dislocation Creep of Olivine: Backstress Evolution Controls Transient Creep at High Temperatures

Abstract

The evolution of viscosity during flow of mantle rocks at high temperatures is fundamental to a variety of geodynamic processes. For example, transient creep of the upper mantle has been identified as a major contributor to geodetically observed surface deformations during post-seismic creep (Freed et al., 2012;Masuti et al., 2016;Pollitz, 2005;Qiu et al., 2018), for which the strains are typically <10 −3 , and inferred viscosities are one to two orders of magnitude lower than the long-term, steady-state viscosity. Because transient viscosities also continue to evolve during postseismic deformation, they likely cause a time-dependent transfer of stresses to neighboring faults, rather than the instantaneous transfer assumed by popular calculations of Coulomb stress changes (e.g., Freed, 2005). Although sophisticated earthquake forecast models do incorporate time-dependent loading according to average plate motion rates (e.g., Field et al., 2015Field et al., , 2017, they still do not incorporate variable loading rates that would occur due to transient creep of the lithosphere. In addition, transient viscosities are expected to be important, although they have not yet been thoroughly considered, in other small-strain processes including flexure of the lithosphere near volcanic loads (Zhong & Watts, 2013) or in subducting slabs near trenches (Hunter & Watts, 2016), during which the strains rarely exceed 10 −2 .

Transient Creep in Subduction Zones by Long‐Range Dislocation Interactions in Olivine

“…For clarity, we emphasize that residual stress heterogeneity is the physical manifestation of long‐range dislocation interactions, which generate back stress as an emergent state variable (Hansen et al., 2021; Wallis et al., 2020). It is difficult to infer the magnitude of former back stress directly from observations of stress heterogeneity, but the presence of long‐range stress heterogeneity imparted by the dislocations does demonstrate the mechanism by which back stress can emerge.…”

“…Furthermore, we note that because the microstructure and micromechanical state that persists during steady‐state creep is the end product of evolution that occurred during the preceding transient, observations of material that were either still within the transient or had reached steady‐state can both provide information on the processes occurring during the transient. For example, stress heterogeneity and associated back stress accumulate during the transient but persist at a steady‐state (Hansen et al., 2021; Wallis et al., 2020). Therefore, the presence of long‐range internal stresses in the rocks from the Oman‐UAE ophiolite supports the back‐stress model of transient creep, regardless of whether the final phase of deformation recorded by the microstructure and micromechanical fields occurred during a transient or at steady state.…”

“…These models are generally parameterized based on power‐law flow laws for steady‐state creep, that is, they assume that the individual crystals are at steady state and the transient results entirely from grain interactions. However, single crystals of olivine deforming by individual slip systems, in which this stress‐transfer model cannot apply, also exhibit transient creep over the strain intervals (<1%) relevant to postseismic deformation (Cooper et al., 2016; Durham et al., 1979; Hansen et al., 2021; Hanson & Spetzler, 1994). This effect casts doubt on the relevance of flow‐law parameters determined at steady state to the behavior of rocks undergoing transient creep and on the relevance of the stress‐transfer model to the low strains at which intragranular processes may dominate the transient behavior (Durham et al., 1977; Hanson & Spetzler, 1994).…”

A Theory of Inter‐Granular Transient Dislocation Creep: Implications for the Geophysical Studies on Mantle Rheology

“…However, the degree to which GND contribute to transient creep under high-temperature, low-stress conditions is unclear. For example, in many studies on high-temperature, low-stress deformation of olivine (e.g., Bai & Kohlstedt, 1992;Darot & Gueguen, 1981;Karato & Lee, 1999;Karato et al, 1986;Karato et al, 1980), high density GNDs (far exceeding the dislocation density corresponding to the applied stress) reported by Hansen et al (2021); Wallis et al (2021) are not observed. This is likely due to the effect of dislocation recovery to form low energy configuration such as sub-boundaries and/or to the effect of dynamic recrystallization by grain-boundary migration (Karato, 1988;Karato et al, 1980Karato et al, , 1982.…”

“…Therefore one may expect creep behavior of a polycrystalline olivine similar to that of a polycrystalline ice. Indeed, strain softening is observed in some studies on olivine single crystals (e.g., Durham et al, 1979;Hanson & Spetzler, 1994), whereas strain hardening is observed in the transient creep of olivine aggregates (Chopra, 1997;Hansen et al, 2021).…”