Evolution of the long-wavelength, subduction-driven topography of South America since 150 Ma

Subduction to the west of South America spans 6000 km along strike and has been active for over 250 Myr. The influence of the history of subduction on the geodynamics of South America has been profound, driving mountain building and arc volcanism in the Andean Cordillera. Here, we investigate the long-wavelength changes in the topography of South America associated with subduction and plate motion and their interplay with the lithospheric deformation associated with the opening of the South Atlantic. We pay particular attention to the topographic expression of flat-lying subduction zones.

We develop time-dependent geodynamic models of mantle flow and lithosphere deformation to investigate the evolution of South American dynamic and total topography since the late Jurassic (150 Ma). Our models are semi-empirical because the computational cost of fully dynamic, evolutionary models is still prohibitive. We impose the kinematics of global plate reconstructions with deforming continents in forward global mantle convection models with compositionally distinct crust and continental lithosphere embedded within the thermal lithosphere. The shallow thermal structure of subducting slabs is imposed, allowing us to investigate the evolution of dynamic topography around flat slab segments in time-dependent models.

Multiple cases are used to investigate how the evolution of South American dynamic topography is influenced by mantle viscosity, the kinematics of the opening of the South Atlantic and alternative scenarios for recent and past flat-slab subduction. We predict that the migration of South America over sinking oceanic lithosphere resulted in continental tilt to the west until ~ 45 Ma, inverting to an eastward tilt thereafter. This first-order result is consistent with the reversal of the drainage of the Amazon River system. We investigate which scenarios of flat-slab subduction since the Eocene are compatible with geological constraints on the evolution of the Solimoes Basin, the Chaco Basin, the Sierras Pampeanas and the Central Patagonian Basin. To broadly constrain mantle viscosity, we compare models to the total subsidence inferred from well data offshore Argentina and Brazil, and to mantle tomography, since the initial and boundary conditions are based on independent plate reconstructions.