The effects of supercontinent size and thermal insulation on the formation of mantle plumes

Geophysical evidence suggests the present day thermal field of the Earth is characterised by anomalously warm material beneath the African and Pacific plates. Continental insulation during the Mesozoic offers a possible explanation for why the mantle below the African plate, a former site of continental aggregation, is warmer than expected. We investigate the effect of continental insulation in 2D and 3D mantle convection models featuring mechanically and thermally distinct continental and oceanic plates to determine the significance of continental insulation. Supercontinental insulation is modelled by limiting continental surface heat flux relative to the heat flux through the isothermal surface of the oceanic plates. For 2D models, we vary continental insulation and width to assess the thermal response of the mantle to each parameter. Our findings indicate that subduction patterns determined by continental width play the dominant role in enabling the formation of subcontinental mantle upwellings. Extending our study to 3D calculations with supercontinental coverage of the mantle comparable to Pangea's, we again find that subcontinental plumes develop as a consequence of subduction patterns rather than continental thermal insulation properties. Moreover, we find that despite the presence of an overlying supercontinent with insulating properties appropriate for modelling terrestrial dynamics, averaged subcontinental mantle temperatures do not significantly exceed sub-oceanic temperatures on timescales relevant to super-continent assembly.

► We examine the effect of supercontinent area and insulation on mantle flow reversals.
► We model distinct oceanic and continental plates.
► We find that continental insulation does not affect subcontinental plume growth.
► Our findings show that subcontinental plumes develop due to subduction patterns.
► Subcontinental temperatures do not exceed the sub-oceanic over relevant time-scales.