Causes for the Onset and Stability of Flat Slabs and Associated Overriding Plate Deformation Inferred from Numerical Thermo-Mechanical Models

Abstract:

Flat slab subduction comprises only ~10% of the present-day subduction systems. We systematically explore flat-slab subduction and over-riding plate deformation due to previously recognized factors (including a weak mantle wedge, a buoyant oceanic plateau/ridge, slab-age, presence of thick cratonic upper plate), as well as additional factors not considered before (transient plate-velocity variations, and the presence of stagnant slabs at the mantle transition zone). We use a visco-elasto-plastic code, SULEC (Buiter & Ellis, 2012), restrict ourselves to 2D models, and ignore phase transformations. We model a 6000 km by 1500 km domain with 1-10 km variable grid spacing using up-to-date laboratory-derived constitutive laws, and explore the effects of: (a) lateral/vertical structure of the crust and lithosphere (e.g., continental lithosphere with/without a craton; oceanic lithosphere with thickness governed by plate cooling, and a harzburgite layer); (b) a true free upper-surface to predict surface topography; and (c) only far-field (boundary) velocities driving plates, with trench-velocity depending on over-riding plate deformation and slab pull.

As in previous studies, we find that positive trench-retreat velocity and weak integrated slab-strength (as measured by differential stress) are necessary, but not sufficient, conditions for flat-slab subduction. While mantle-wedge suction associated with the presence of a craton does promote flat slab subduction, models without a craton also produce flat-slabs when (a) trench-retreat velocities are large and (b) either the slab is positively (structurally or thermally) buoyant or there is an underlying slab at the mantle transition zone. In our models, the duration of flat slab subduction depends on overall slab buoyancy. Younger oceanic lithosphere, esp. if it contains a plateau/ridge, provides a longer-lasting flat-slab (~ 10 Myr or longer, e.g., central Mexico), while older slabs with oceanic ridge/plateau (e.g., Peru) flatten for only a few Myr except in the presence of dynamic support from the transition zone (previously subducted slabs or a high viscosity contrast). Finally, the age of the subducting plate and its overall buoyancy seem to determine the deformation style within the over-riding plate.