Seismic Tomography of South America: Continental Lithosphere, Subduction-Zone Structures, and the Origins of Intraplate Magmatism

Abstract

The lithospheric structure of South America records its long and complex tectonic evolution, during which it has influenced the distribution and composition of local magmatism, the geometry of the subduction at its western margin, and the behaviour of the Andean orogeny. Throughout the years, seismic tomography models have provided increasingly accurate images of the Earth’s structure beneath South America. However, our understanding of the upper mantle underlying the continent and its surroundings is still limited by the sparse and uneven distribution of seismic data. In this work, we present a new seismic tomography model of South America and surrounding oceans, including the crust and upper mantle. To address the heterogeneous data coverage in South America, our model SACI-24 (South America Continental Imaging 2024) is based on the inversion of more than 970,000 waveforms from ~300,000 earthquakes, recorded by more than 9000 seismic stations globally and locally. The combined regional and global dataset ensures the densest possible data sampling of the South American continent, its margins, and the surrounding oceans. We invert the waveforms using the Automated Multimode Inversion (AMI) of S-, multiple S-, and surface waves. AMI generates a set of independent linear equations with uncorrelated uncertainties for each source-receiver path. These equations describe the path-average S-wave velocity structure within approximate sensitivity kernels. We assemble these equations into a large linear system and solve it to obtain the 3D distribution of S-wave velocities in the crust, upper mantle, and transition zone. Our model is calculated globally; however, the parametrization and regularisation values are optimized for the South American region, and the data coverage is maximized only in the hemisphere centred on South America. Within the lithosphere, our model reveals a more complex structure of the cratonic lithosphere than previously proposed. Internal heterogeneities within cratonic boundaries include regions of thinner lithosphere, which correspond to areas of proposed rifting in previous tectonic cycles. Inside the boundaries of the Amazon Craton, two distinct cratonic blocks are identified, separated by the Amazon basin. In the São Francisco Craton, newly imaged thin lithosphere underlies the Paleoproterozoic Paramirim Aulacogen area. Our work further reveals separate, high-velocity lithospheric blocks beneath the Parnaíba and Paraná basins. South of the Paraná Basin, our model maps for the first time evidence of the cratonic root of the Rio de la Plata Craton. The detailed mapping of the lithospheric structure reveals its control on several geodynamic processes. Comparison of cratonic lithosphere with the reconstructed position of hotspots reveals that lithospheric thickness is the main control on the distribution of volcanism in large igneous provinces. Under the roots of the Andean Cordillera, our model finds multiple regions of underthrusting of continental lithosphere from South America. Our high-resolution images of the active margin reveal highly heterogeneous subduction, indicating several slab segments from the Nazca and Caribbean plates. East of the Peruvian flat slab, we find areas of thick lithosphere that we attribute to the Amazon Craton, indicating the role of cratonic hydrodynamic suction in flat slab formation. Furthermore, we identify regions under the Altiplano and Puna Plateaus undergoing lithospheric root growth and delamination, indicative of varying stages of Cordillera cyclicity. Within the mantle transition zone, we detect segments of colder mantle from the subduction of the Nazca plate. East of these subduction segments, we identify fast seismic anomalies we interpret as remnants of past lithospheric delamination events. Underlying the thick continental lithosphere, our model reveals prominent low-seismic velocity anomalies extending as deep as the transition zone, with the strongest anomaly found beneath southeast Brazil. Whole-mantle tomography shows that this anomaly extends to a depth of at least 2000 km depth, suggesting a deeply rooted mantle plume. We postulate that interaction with past slab fragments from the long-lived Nazca subduction modifies the plume ascension path, causing it to branch out and travel horizontally to areas of thin oceanic lithosphere, where we find the present-day location of the many hotspots in the Atlantic Ocean.