Investigating ice shell convection with a lower boundary defined by changes in phase and composition: Implications for Europa

Models of solid-state convection typically employ a spatially-fixed lower boundary with a velocity boundary condition that impedes material flow across it. Oftentimes, such a boundary condition is justified due to the large density contrast that the boundary represents; for example, a fixed, no-flow lower boundary condition is an appropriate approximation for the Earth's core-mantle boundary due to the large density difference between silicate rock and iron. Less extreme density contrasts are not well approximated by a fixed lower boundary. For example, although the boundary between Earth's upper and lower mantles represents a sharp density contrast due to a phase change between ringwoodite and perovskite, it is not strong enough to restrict flow across it. It is unclear what type of lower boundary is most appropriate for modeling convection within an ice shell floating on a saline water ocean such as hypothesized for Europa. It is important to determine whether material is expected to advect across the ice-water boundary and to determine whether significant vertical topography exists along the boundary (i.e., variation in ice shell thickness) because both will affect the vigor and wavelength of convection and, more importantly, the ability of the convective ice shell to produce plumes. To investigate this, we perform two sets of numerical convection calculations, each identical except for the treatment of the lower boundary. The control set employs a fixed, no-flow lower boundary, and the test set utilizes thermochemical convection to allow for the self-consistent generation of a lower boundary. We examine various density contrasts between ice and water to reflect uncertainty in ocean salinity. We report how the treatment of the lower boundary influences mass flux, vigor and wavelength of convection, and the formation or inhibition of plumes.