Three-dimensional numerical simulations of crustal deformation and subcontinental mantle convection

3-D simulations of mantle convection allowing for continental crust are explored to study the effects of crustal thickening on lithosphere stability and of continents on large-scale mantle flow. Simulations begin with a crustal layer within the upper thermal boundary layer of a mantle convection roll in a 1 × 1 × 1 Cartesian domain. Convective stresses cause crust to thicken above a sheet-like mantle downwelling. For mild convective vigor an initial crustal thickness variation is required to induce 3-D lithospheric instability below the zone of crustal convergence. The amplitude of the required variation decreases with increasing convective vigor. Morphologically, instability is manifest in formation of drip-like thermals that exist within the large-scale roll associated with initial crustal thickening. A strong surface signature of the drips is their ability to cause deviations from local Airy compensation of topography. After the initial thickening phase, the crustal accumulation that forms serves as a model analog to a continent. Its presence leads to mantle flow patterns distinctly different from the steady-state roll that results in its absence. Large lateral thermal gradients are generated at its edge allowing this region to be the initiation site for continued small-scale thermal instabilities. Eventually these instabilities induce a restructuring of large-scale mantle flow, with the roll pattern being replaced by a square cell. Although preliminary and idealized, the simulations do show the fluid dynamical plausibility behind the idea that significant mantle variations can be generated along the strike of a largely 2-D mountain chain by the formation of the chain itself. The ability of a model continent to cause a change in fundamental convective planform also suggests that the effects of continental crust on mantle convection may be low-order despite the seemingly trivial volume of crust relative to mantle.