On the effects of the lithosphere on mantle convection and evolution

Abstract

An idealized model is discussed for plate motion and mantle convection. It consists of a “numerical” square box filled with a Newtonian-Boussinesq fluid heated from below. At the surface, the fluid has a high viscosity, i.e. it contains a stiff lid which may or may not be laterally decoupled by weakness zones. Lid stiffness and coupling were varied systematically. Only if the lid is effectively decoupled laterally, does it move as a quasi-rigid plate. It is then driven by the lateral pressure differential from the rising and descending plumes in the fluid and it moves against viscous shear at its bottom. Gravitational sliding and sinking may also drive the plate if allowed for in the model. A decoupled freely-moving stiff plate effectively removes heat from the interior. A strongly-coupled stiff lid, on the other hand, or one with a very large aspect ratio inhibits heat transfer through the convecting layer. If heating occurs by a given flux through the bottom or internal heat generation the internal cell temperatures become very high, because the “stagnant” surface lid can only conduct the heat by assuming the required temperature gradient throughout its thickness. Since the Earth is partly covered by conducting continental lids which in addition contain much of the heat sources, the thermal history may be affected; caution must be exercised in parameterized convection modelling in this respect. The overall effect cannot easily be predicted because mantle flow and continental plates will interact. The importance of mechanical decoupling for plate motion strongly calls for thorough studies of the decoupling mechanisms, particularly at convergent plate boundaries, bending, stick-slip processes, and the role of volatiles.