Thermal Evolution of a compositionally stratified

For subduction to occur, plates must bend and slide past overriding plates along fault zones. This deformation is associated with significant energy dissipation, which changes the energy balance of mantle convection and influences the thermal history of the Earth. To parameterize these effects, a subduction zone was included in a small region of a finite element model for the mantle, which also features an asthenosphere and a mid-oceanic ridge. Velocity boundary conditions were imposed in the vicinity of the subduction zone and a rate for subduction was specified, that balances the energy budget for convection. This balance includes an expression for the energy needed to bend the oceanic lithosphere as it subducts. Four different modes of energy dissipation were considered: viscous bending, brittle bending, viscous simple shear, and brittle simple shear. We present theoretical arguments for, and numerical illustrations of the fact that for most modes of deformation, the simple powerlaw relationship of parameterized convection Nu ~ RaO is not valid anymore, although it is still a good first order approximation. In the case of viscous bending dissipation and non-depth dependent brittle simple shear however, Nu ~ Ra does hold. # is less than the value of 1/3 predicted by standard boundary layer theory. For viscous energy dissipation, two different regimes of mantle convection can be considered, depending on the effective viscosity of the lithosphere: the "mobile lid" regime, and the "stagnant lid" regime. For brittle dissipation, the lithosphere strength is a function of yield stress which, when nearing a certain critical value, introduces a third regime, that of the "episodic overturning". Within the "mobile lid" regime, the plate velocities for models with a subduction zone governed by brittle behavior are far less dependent on the plate stress than those models with viscous deformation. This suggests that the plate motion is resisted by viscous stresses in the mantle. The "mobile lid" would be representative for mantle convection associated with plate tectonics, as we observe on Earth. A "stagnant lid" would be the case for the Moon or Mars, while Venus could experience the "episodic overturn" regime featuring cyclic and catastrophic brittle mobilization of a lithosphere with high friction coefficient. Thesis Supervisor: Bradford H. Hager Title: Cecil and Ida Green Professor of Earth Sciences Acknowledgments In the first place, I would like to acknowledge Brad Hager, my thesis advisor and Geosystems coordinator, who suggested to me this topic which I enjoyed studying very much. I cannot thank enough Clint Conrad, whose Ph.D. thesis work "Effects of Lithospheric Strength on Convection in the Earth's Mantle" was the firm basis for this Master's thesis. He helped me out whenever I had a problem with ConMan, or with anything else. His respose from Caltech always came very promptly, thanks to the marvel of e-mail. Maybe I should hence also thank the people that made the Internet! Many thanks to my Geosystems colleagues Lorraine, Russ, Chris, and Victoria for the fantastic year I had with them @ MIT. Good luck to you all! This work was performed while being a recipient of a Francqui Fellowship of the Belgian American Educational Foundation. Partial financial support was provided by NSF grant EAR-9905779.