Reconstructing the Cenozoic evolution of the mantle: Implications for mantle plume dynamics under the Pacific and Indian plates

The lack of knowledge of the initial thermal state of the mantle in the geological past is an outstanding problem in mantle convection. The resolution of this problem also requires the modelling of 3-D mantle evolution that yields maximum consistency with a wide suite of geophysical constraints. Quantifying the robustness of the reconstructed thermal evolution is another major concern. To solve and estimate the robustness of the time-reversed (inverse) problem of mantle convection, we analyse two different numerical techniques; the quasi-reversible (QRV) and the backward advection (BAD) methods. Our investigation extends over the 65 Myr interval encompassing the Cenozoic era using a pseudo-spectral solution for compressible-flow thermal convection in 3-D spherical geometry. We find that the two dominant issues for solving the inverse problem of mantle convection are the choice of horizontally-averaged temperature (i.e., geotherm) and mechanical surface boundary conditions. We find, in particular, that the inclusion of thermal boundary layers that yield Earth-like heat flux at the top and bottom of the mantle has a critical impact on the reconstruction of mantle evolution. We have developed a new regularisation scheme for the QRV method using a time-dependent regularisation function. This revised implementation of the QRV method delivers time-dependent reconstructions of mantle heterogeneity that reveal: (1) the stability of Pacific and African 'large low shear velocity provinces' (LLSVP) over the last 65 Myr; (2) strong upward deflections of the CMB topography at 65 Ma beneath: the North Atlantic, the south-central Pacific, the East Pacific Rise (EPR) and the eastern Antarctica; (3) an anchored deep-mantle plume ascending directly under the EPR (Easter and Pitcairn hotspots) throughout the Cenozoic era; and (4) the appearance of the transient Reunion plume head beneath the western edge of the Deccan Plateau at 65 Ma. Our reconstructions of Cenozoic mantle evolution thus suggest that mantle plumes play a key role in driving surface tectonic processes and large-scale volcanism. (C) 2014 Elsevier B.V. All rights reserved.