Gravitational differentiation due to initial chemical stratification: origin of lunar asymmetry by the creep of dense KREEP?

Models for the evolution of the Moon must explain a number of important characteristics of its magmatic history. Primary among them is the emplacement of mare basalts primarily on one side of the Moon, and in a region of the surface roughly correlating with that containing high subsurface concentrations of KREEP. It is thus important to explore mechanisms which will regionally concentrate a KREEP layer early in lunar evolution and which ∼400–500 Myr later give rise to mare basalt generation beneath the same region. Gravitational instability of a chemically dense, ilmenite-rich cumulate created during the fractionation of an anorthositic crust is one candidate model. If so, gravitational instability must occur at long wavelength to explain the hemispheric asymmetry. We examine the Rayleigh–Taylor instability of a dense layer near the surface of a sphere, representing the lunar mantle, in which the viscosity varies with depth. Our results show that spherical harmonic degree 1 is the fastest growing wavelength of instability if the viscosity of the dense layer is sufficiently low relative to that of the deeper mantle or if the viscosity of the mantle increases with depth. However, the viscosity increase cannot be distributed over a depth that is too large. We explore whether a residual KREEP-like layer near the base of anorthositic crust might give rise to these needed conditions.