Search

The rheological properties of Earth's lower mantle are key for mantle dynamics and planetary evolution. The main rock-forming minerals in the lower mantle are bridgmanite (Br) and smaller amounts of ferropericlase (Fp). Previous work has suggested that the large differences in viscosity between these minerals greatly affect the bulk rock rheology. The resulting effective rheology becomes highly strain-dependent as weaker Fp minerals become elongated and eventually interconnected. This implies that strain localization may occur in Earth's lower mantle. So far, there have been no studies on global-scale mantle convection in the presence of such strain-weakening (SW) rheology. Here, we present 2D numerical models of thermo-chemical convection in spherical annulus geometry including a new strain-dependent rheology formulation for lower mantle materials, combining rheological weakening and healing terms. We find that SW rheology has several direct and indirect effects on mantle convection. The most notable direct effect is the changing dynamics of weakened plume channels as well as the formation of larger thermochemical piles at the base of the mantle. The weakened plume conduits act as lubrication channels in the mantle and exhibit a lower thermal anomaly. SW rheology also reduces the overall viscosity, notable in terms of increasing convective vigor and core-mantle boundary heat flux. Finally, we put our results into context with existing hypotheses on the style of mantle convection and mixing. Most importantly, we suggest that the new kind of plume dynamics may explain the discrepancy between expected and observed thermal anomalies of deep-seated mantle plumes on Earth., Geochemistry, Geophysics, Geosystems, 23 (10), ISSN:1525-2027

Over the last decades, mantle convection models have become more and more advanced. Now, these numerical models are able to address compressibility, various compositions, phase transitions and depth- and temperature-dependent viscosity. However, grain size-dependent viscosity is still not widely used, although nature teaches us that grain size varies throughout the Earth. Experimentalists, field geologists and geodynamic modellers employing small-scale models all present growing evidence for the impact of grain size evolution on material weakening, deformation rate and on its coupling with other geophysical processes. Therefore, a deeper comprehension of how grain size is distributed in the Earth, whether geodynamic convection models can successfully reproduce natural observations, and how grain size evolution feeds back with other geophysical mechanisms are of high interest. As a varying grain size is thought to help explain observations that up to now have not been completely understood, high expectations towards integrating grain size evolution in geodynamic convection codes exist. With this Doctoral thesis I wish to shed light on the e ffect of grain size evolution and a composite viscosity on several geological phenomena. To study those, I employ regional and global mantle convection models. I test the influence of grain size evolution on the development and evolution of Large Low Shear Velocity Provinces, seismically observed regions at the core-mantle boundary. These regions are thought to be thermally and chemically distinct from the ambient mantle, and might have a di erent density and viscosity. This study demonstrates, that grain size-dependent viscosity does not increase the resistance of thermo-chemical piles towards downgoing slabs. Mostly, piles deform in grain size-sensitive di usion creep but they are not sti enough to counteract the force of downwellings. The results also show, that, in the ambient mantle, viscosity variations with time are much smaller when using a grain size-dependent viscosity compared to when employing a temperature-dependent viscosity. In a second chapter, I examine whether grain size reduction within shear zones can lead to strain localisation and transform fault initiation. The results show that the ductile weakening induced by grain size reduction results in sufficient localisation to initiate a transform fault which is stable for 2 Myr. The modelled transform faults initiate in the brittle crust and propagate downwards, thus indicating a top-down controlled process. Grain size, rheology and strain rate inside the shear zone of our models agree well with observations in nature. This chapter is extended by a semi-analytical study on transform fault development (chapter 4) in which the viscously deforming region below the brittle transform fault is examined. The preliminary results presented in this chapter agree with observations from the numerical study. Lastly, I investigate whether a grain size-dependent viscosity changes the amount of work required for subduction in a subduction system by using a semi-analytical approach for which I developed a new model. The general reduction of complexity in analytical models simplifies the observation of direct correlations between variables, and the influence of specific parameters on the modelled system. The approach employed in this thesis combines observations by mineral physicists with the advantage of numerical modelling and marks an important step towards more interdisciplinary research in the field of geoscience.