Initiating Localized Deformation in the Mantle

Phil Skemer, Washington University in Saint Louis

Shear localization is an essential feature of mantle deformation, particularly along plate boundaries. However the physical bases for localized deformation at high temperatures and pressures are poorly understood. There are a number of plausible strain-weakening mechanisms that may contribute to the initiation and evolution of shear zones, including grain-size reduction, compositional gradients, texture development, and viscous shear heating . Geologically, mantle shear zones are often identified on the basis of fine-grained, weakly textured, mylonitic microstructures. These microstructural observations are widely interpreted as evidence that dynamic recrystallization, grain-size reduction, and the associated transition to grain-size sensitive deformation is the predominant strain-weakening process in mantle rocks. To assess these ideas about shear localization I will present results from two studies. I will first describe some recent laboratory experiments. These experiments, conducted at 1 GPa and 1200°C on coarse grained synthetic harzburgite, demonstrate that the serial processes of dynamic recrystallization, phase mixing, and grain-size sensitive deformation, do not occur at small strains. Hence, I will argue that this commonly invoked strain-weakening mechanism is a consequence of localized deformation, rather than its cause. I will then compare our laboratory results with field observations from the Josephine Peridotite (SW Oregon). In this study, field measurements are used to construct strain profiles across several mantle shear zones to determine the magnitudes and gradients in strain accumulation. Measurements of water concentration in nominally anhydrous minerals show that gradients in water concentration exist on a 10-100 m scale, giving rise to spatial variations in viscosity of up to a half order of magnitude. These water concentration measurements are also correlated with the locations of shear zones and the observed olivine CPO. Using empirical flow laws we model the formation of mantle shear zones using water concentration variation to generate perturbations in the strain field. We also include in the models the effect of viscous anisotropy due to the progressive re-orientation of olivine CPO. On the basis of the laboratory experiments, field observations, and modeling, I will attempt to show that no single process can explain shear localization in the mantle. Rather, the initiation and evolution of mantle shear zones requires several independent serial processes.