Feasibility of the mobile-lid regime controlled by grain size evolution

One of the most discussed issues of whole-mantle geodynamic models is the need of an 'ad hoc' yield stress which is lower than any strength measurement of natural samples in the brittle or plastic regimes. It is commonly believed that grain size evolution, in particular grains size reduction due to dynamic recrystallization, may decrease the strength of the lithosphere and therefore aid the onset and persistence of the mobile-lid regime. In this work, we carry out an investigation of 2D whole-mantle annulus models with varying yield stress. We compare cases with different grain growth and grain reduction parameters to cases with constant grain size to make inferences on the feasibility of a plate-like convective regime as a function of the yield strength of the lithosphere.

Our results show that viscosity profiles of models with dynamic grain-size evolution are inherently different to those with constant grain size, and that those profiles vary little when changing grain-size evolution parameters. In this context, the lower mantle shows greater variations in viscosity than the upper mantle: with viscosity contrasts between upper and lower mantle and plume widths comparable to those of the Earth, in particular in models with enhanced grain growth. Furthermore, our models show that, while enhancing grain size reduction reduces episodicity and increases mobility up to some point, increasing grain growth favors mobile-lid convection even more. This is at odds with previous conceptions of the grain-size-evolution-induced mobile-lid regime, where grain groth should promote healing of the lithosphere and therefore inhibit subduction. We hypothesize that increased stiffness of the bottom of the lithosphere, together with a more viscous lower mantle, are the main reasons for the grain-grouth-favored mobile-lid regime.