The effect of plate motion history on the longevity of deep mantle heterogeneities

Understanding the first-order dynamical structure and evolution of Earth's mantle is a fundamental goal in solid-earth geophysics. Tomographic observations reveal a lower mantle characterised by higher-than-average shear-wave speeds beneath Asia and encircling the Pacific, consistent with cold slabs beneath regions of ancient subduction, and lower-than-average shear-wave speeds in broad regional areas beneath Africa and the Central Pacific (termed LLSVPs). The LLSVPs are not well understood from a dynamical perspective and their origin and evolution remain enigmatic. Some numerical studies propose that the LLSVP beneath Africa is post-Pangean in origin, formed as a result of return flow in the mantle due to circum-Pangean subduction, countered by an older Pacific LLSVP, suggested to have formed during the break up of Rodinia. This propounds that, prior to the formation of Pangea, the lower mantle was dominated by a degree-1 convection pattern with a major upwelling centred close to the present-day Pacific LLSVP and subduction concentrated mainly in the antipodal hemisphere. In contrast, palaeomagnetic observations which proffer a link between the reconstructed eruption sites of Phanerozoic kimberlites and Large Igneous Provinces with regions on the margins of the present-day LLSVPs suggest that the anomalies may have remained stationary for at least the last 540 Myr and further that the anomalies were largely insensitive to the formation and subsequent break-up of Pangea. Here we investigate the evolution and long-term stability of LLSVP-like structures in Earth's mantle by integrating plate tectonics and numerical models of global thermochemical mantle dynamics. We explore the possibility that either one or both LLSVPs existed prior to the formation of Pangea and improve upon previous studies by using a new, true polar wander-corrected global plate model to impose surface velocity boundary conditions for a time interval that spans the amalgamation and subsequent break-up of the supercontinent. We find that, were only the Pacific LLSVP to exist prior to the formation of Pangea, the African LLSVP would not have been created within the lifetime of the supercontinent. We also find that, were the mantle to be dominated by two antipodal LLSVP-like structures prior to the formation of Pangea, the structures would remain relatively unchanged to the present day and would be insensitive to the formation and break-up of the supercontinent. Our results suggest that both the African and Pacific LLSVPs have remained close to their present-day positions for at least the past 410 Myr.