Mantle Convection with Active Chemical Heterogeneities

Abstract

The main source of mantle heterogeneity is the subduction of basaltic oceanic crust. Whether or not coherent lumps of continental crust are recycled in significant amounts is unkown. These heterogeneities will — at least in parts — be torn into progessively longer and thinner streaks (or schlieren, tendrils, bands, lamellae), giving rise to a “marble cake mantle” (Allègre and Turcotte, 1986). The question of how fast the thickness reduction proceeds to the point of complete mixing has been studied by means of numerical modelling (Olson et al., 1984; Hoffman and McKenzie, 1985; Gurnis 1986b). The various results are in disagreement and seem to depend on model assumptions. In all these models (with one exception: Gurnis, 1986a) the chemical heterogeneity has been taken as passive, which means that differences in physical properties, like density or viscosity, between the injected former crust and the host rock are neglected. The purpose of this study is to construct models of active chemical heterogeneity. It splits into two parts. In the first part the consequence of a viscosity difference between the streaks and their host rocks is considered; it leads to anisotropic rheological behavior. In the second part density differences are considered to determine the feasibility of segregation of light and heavy material.