Small-scale convection in the D layer -: art. no. 2016

[1] Small-scale convection has been suggested as a possible explanation for seismic heterogeneities in the D layer of the Earth's mantle. Recently developed scaling laws for convection with realistic viscosities allow quantitative assessment of this hypothesis. Large temperature and viscosity contrasts across the thermal boundary layer at the core-mantle boundary suggest that small-scale convection starts at the bottom of the thermal boundary layer and propagates upward before the thermal boundary layer as a whole becomes unstable. The convection boundary is likely to become a chemical boundary as a result of mixing within the convective layer. This implies that the D discontinuity can represent both convective and chemical boundary and that the presence or absence of small-scale convection can be responsible for the observed intermittent nature of the D discontinuity. Most lateral heterogeneities in the D layer are concentrated near the convection boundary, consistent with seismic data. The length scale of lateral temperature variations, the topography of the core-mantle boundary, and the viscosity of the D layer are in agreement with observational constraints. The thickness of the secondary thermal boundary layer formed at the bottom of the convective layer is similar to the thickness of the ultralow-velocity zone. The magnitude of lateral variations in temperature can only marginally be responsible for lateral variations in seismic velocities. Variations in the topography of the convection boundary and chemical heterogeneities are likely to be more important factors. A large seismic velocity drop in the ultralow-velocity zone cannot be explained by thermal effects alone and must be caused by other factors such as partial melting.