Mantle convection structure beneath stationary continents V. P. Trubitsyn and A. M. Bobrov Abstract We made numerical tests to elucidate the impact of continents on the structure of thermal convection in the mantle. The mantle was modeled by a viscous fluid occupying a horizontally extended rectangular region having the aspect ratio 1\,:\,10. Continents were treated as thick solid heat-conducting plates placed on the mantle. Calculations were carried out for convection at Rayleigh numbers ${\rm Ra}=10^5$, $5\times 10^5$, and $10^6$. Plate thickness $d$ varied from 0.05 to 0.3, and their length $L$ changed from 0.3 to 2, in mantle depth units. Continents prevent heat release from the underlying mantle; the mantle material heats up and becomes lighter; as a result, a hot upwelling flow replaces downwelling. We calculated characteristic time $\tau$ for various values of model parameters and obtained analytical approximations for $\tau({\rm Ra})$, $\tau(d)$, and $\tau(L)$. This time was estimated as $2\times 10^8$ years for the parameters commonly assumed: plate thickness and length of about 300 km and 6000 km, respectively. Continental drift can be significant in restructuring of convective patterns. Upwelling flow can evolve beneath a continent of length $L$ if it moves a distance of less than $L/2$ in time $\tau$. Thus a hot upwelling flow must exist under a continent drifting with a velocity well below $1\:{\rm cm\:yr}^{-1}$, as in the case of Africa, whereas such a flow has no time to evolve beneath continents that drift fast enough (like Australia). Back to Computational Seismology, Vol. 4.