T23A-4635:
Lithosphere-Mantle Interactions Associated with Flat-Slab Subduction

Tuesday, 16 December 2014
Melanie Gerault1, Thorsten W Becker1, Laurent Husson2 and Eugene Humphreys3, (1)University of Southern California, Department of Earth Sciences, Los Angeles, CA, United States, (2)Universite Joseph Fourier, ISTerre UMR CNRS 5275, Grenoble, France, (3)Univ Oregon, Eugene, OR, United States
Abstract:
Episodes of flat-slab subduction along the western margin of the Americas may have lead to the formation of intra-continental basins and seas, as well as mountain belts and continental plateaux. Here, we explore some of the consequences of a flat slab morphology, linking dynamic topography and stress patterns in continents to slab and mantle dynamics. Using a 2-D cylindrical code, we develop general models and apply them to the North and South America plates. The results are primarily controlled by the coupling along the slab-continent interface (due to geometry and viscosity), the viscosity of the mantle wedge, and the buoyancy of the subducted lithosphere. All models predict broad subsidence, large deviatoric stresses, and horizontal compression above the tip of the flat slab and the deep slab hinge. In models where the slab lays horizontally for hundreds of kilometers, overriding plate compression focuses on both ends of the flat segment, where normal-dip subduction exerts a direct downward pull. In between, a broad low-stress region gets uplifted proportionally to the amount of coupling between the slab and the continent. Anomalously buoyant seafloor enhances this effect but is not required. The downward bending of the flat slab extremities causes its upper part to undergo extension and the lower part to compress. These results have potential for explaining the existence of relatively undeformed, uplifted regions surrounded by mountain belts, such as in the western U.S. and parts of the Andes. Adequately modeling topography and stress in the unusual setting of southwestern Mexico requires a low-viscosity subduction interface and mantle wedge. Our results are only partially controlled by the buoyancy of the subducting plate, suggesting that the viscosity and the morphology of the slab are important, and that the often-used low resolution and "Stokeslet" models may be missing substantial effects.