Laboratory and theoretical modeling of self-consistent plate tectonics – PTECTO

According to the theory of plate tectonics, Earth's surface comprises a network of quasi-rigid plates in relative motion, each of which « lives » for up to 200 MY from its generation at a mid-ocean ridge to its disappearance in a subduction zone. Clearly, the motions observed on the surface must be related in some way to the thermal and/or thermochemical convection that occurs deeper in the mantle. Yet forty years after the formulation of plate tectonics theory, the question of how plate tectonics arises in a self-consistent manner from mantle convection remains one of the most important unresolved problems in Earth science. Answering this question would have broad implications for many aspects of Earth and planetary dynamics, such as how plate tectonics first appeared from a primitive magma ocean, why it has lasted so long on Earth, and why it does not occur (at least today) on other planets. We propose a integrated laboratory, theoretical, and numerical study of how plate tectonics is generated by mantle convection. The essential impulse for the study is the recent discovery in our laboratory of a class of experimental fluids (colloidal suspensions) that exhibit earthlike plate tectonics when convecting. On the experimental side, we will carry out a systematic study of plate tectonics in convecting colloidal suspensions, with an emphasis on understanding the origin of one-sided subduction and episodic vs. continuous subduction modes. The experience gained in these experiments will then be applied in a second series of experiments on the effect of rheology on the dynamics of rifted colloidal suspensions, an analog model for accreting plate boundaries. On the theoretical side, we will focus on understanding two important aspects of the subduction process as seen in our preliminary experiments: its initiation, and the « fully developed » stage where the descending slab deforms as a thin viscous sheet embedded in an ambient fluid of lower viscosity. The former problem will be approached using asymptotic thin-layer theory that accounts for both viscous and brittle rheology, while the second will be attacked using boundary-element numerical simulations.