From strong plates to weak boundaries: strain localization in the lithospheric mantle with low- to high-temperature dislocation creep

Abstract. Plate-like behavior may be reproduced in numerical experiments of mantle convection if rheology combines a temperature-dependent viscosity capped with a low yield stress. Other processes limiting the mantle ductile strength have been proposed, among which low-temperature plasticity. We propose to investigate how such a rheology can promote deformation localization at the lithospheric scale. We design a 2-D thermo-mechanical model of plate extension to compare the modes of strain localization as a function of the rheological parameterization. Various experimental flow laws for olivine (diffusion creep and dislocation creep at low- and/or high- temperature) are tested, sometimes combined with a yield-stress formulation. We quantify the evolution of the deformation pattern throughout the lithosphere, and define new diagnostics to assess whether the decrease in plate viscosity is due to an increase in strain rate ('mechanical weakening') and/or in temperature ('thermal weakening'). Deformation localization leads to a new extensional plate boundary in two successive stages: (i) narrowing of the deforming zone, (ii) rapid thinning of the highly deformed lithosphere. Here, a rheology combining diffusion creep and yield stress results in a mantle weakening that is either fully mechanical, respectively fully thermal, for temperatures lower, respectively higher, than ~1300 K. Accounting for dislocation creep enables additional feedbacks between mechanical and thermal weakening within the deforming plate, thereby enhancing the efficiency of strain localization. We demonstrate the consistency of using a yield-stress approximation to cap the mantle strength for temperature below ~1000 K. We also show that dislocation creep and yield-stress are not interchangeable for most of the lithospheric mantle (1000–1500 K), and using only the latter may overestimate the duration of plate break-up. Finally, we discuss how our results compare to natural continental rifting cases.