The thermal influence of the subducting slab beneath South America from 410 and 660 km discontinuity observations

Regional seismic network data from deep South American earthquakes to western United States and western European seismic arrays is slant stacked to detect weak near-source interactions with upper mantle discontinuities. These observations are complemented by an analysis of earlier work by Sacks & Snoke (1977) who observed S to P conversions from deep events to stations in South America, and similar observations from 1994-95 events using the BANJO and SEDA networks. Observations of the depth of the 410 km discontinuity are made beneath central South America in the vicinity of the aseismic region of the subducting Nazca Plate. These results image the 410 km discontinuity over a lateral extent of up to 850 km perpendicular to the slab and over a distance of 2700 km along the length of the slab. Away from the subducting slab the discontinuity is mainly seen near its global average depth, whilst inside the slab there is evidence for its elevation by up to around 60 km but with significant scatter in the data. These results are consistent with the presence of a continuous slab through the aseismic region with a thermal anomaly of 900 degreesC at 350 km depth. This value is in good agreement with simple thermal models though our data are too sparse to accurately constrain them. Sparse observations of the 660 km discontinuity agree with tomographic models suggesting penetration of the lower mantle by the slab in the north but stagnation at the base of the transition zone in the south. The geographical distribution of the data, however, does not allow us to rule out the possibility of slab stagnation at the base of the transition zone in the north. These observations, together with the presence of deep earthquakes, require more complicated thermal models than previously used to explain them, possibly including changes in slab dip and age with depth.