Numerical Models of Subduction Beneath Non-Uniform Overriding Plates. Implications for Subduction Velocity and Seismic Anisotropy

Abstract:

Subduction zones show significant along-strike variations of different subduction features. In particular there is a good correlation between variations of overriding plate thickness and variations of slab dip, seismic anisotropy and subduction velocity. Previous numerical models have shown that overriding plate thermal state influences the slab dip and variations in slab dip can cause trench-parallel flow above the slab. This suggests a causal link between overriding plate structure, slab geometry and mantle flow in subduction zones. Models also show that interplate coupling is stronger for colder overriding plates, which might lead to lower subduction velocities.

We implement generic numerical models to study the effect of the overriding plate structure on the evolution of slab geometry, induced mantle flow and subduction velocity. We solve the 3D, time dependent thermo-mechanical equations with a non-linear rheology to simulate buoyancy driven subduction processes. We find that along-strike variations in thermal thickness of the overriding plate cause increased hydrodynamic suction and shallower slab dip beneath the colder portion of the overriding plate; the variation in slab geometry drives strong trench-parallel flow beneath the slab and a complex flow pattern above the slab. Induced mantle flow varies significantly with time, suggesting that the global variability in seismic anisotropy observations in subduction zones is in part due to the non-steady-state behavior of these systems. This new mechanism for driving trench-parallel flow provides a good explanation for seismic anisotropy observations from the Middle and South America subduction zones. Our results also show that increased interplate coupling beneath a colder portion of the overriding plate leads to a reduction of the subduction velocity in the region, leading to along-strike variations subduction velocity. These results provide a good explanation of the observed present-day variations in subduction velocity in Middle and South America, as well as the Phillipine Sea Plate. Therefore, in addition to mechanisms such as variations in subduction plate age or slab length in the upper mantle, the possible role of overriding plate structure should be taken into account in interpreting mechanisms for variation subduction velocity.