Parametric study for self-sustained Andean-type subduction speed

The relation between subduction dynamics, plate rheology and geometry is still not well understood. To numerically assess how subduction convergence velocity develops, a wide range of simulations is required to quantify any correlation between physical parameters and kinematic behavior of a subduction system. In this study, we performed a set of 2D numerical simulations designed to better constrain the range of rheological and geometrical conditions necessary to model subduction dynamics. We used the parallel numerical code LaMEM to simulate thermo-mechanical convection. In addition, we coupled these numerical simulations with MAGEMin to compute a self-consistent mineral assemblage of the asthenospheric mantle and the plates, and we parameterized the lower mantle using a linear equation, following the Clapeyron slope from Faccenda and Zilio (2017). The modeled region is 9300 km wide and accounts for both the upper and whole lower mantle. We consider an Andean type subduction system where our baseline scenarios are defined by a partially subducting oceanic plate beneath a continental plate. Once the simulation starts, the subducted portion of the oceanic plate triggers the subduction thanks to a weak zone between the lower and upper plate. The subduction is sustained by the negative buoyancy of the lower plate with respect to the surrounding mantle. We aim to simulate subduction dynamics that exhibit convergence velocities and long term behavior, including lower mantle penetration, similar to what is observed in nature. We investigate the role of the length of the subducting plate, the geometry of its composing lithological units, and the viscosity of the asthenosphere and the lower mantle. We find that the subduction velocity is inversely correlated with the non subduction length of the oceanic plate, and that a less viscous asthenospheric mantle increases the subduction speed.