Crustal thicknesses in SE Brazilian Shield by receiver function analysis: Implications for isostatic compensation
 The Brazilian Lithosphere Seismic Project (BLSP, a joint project by University of Sao Paulo and Carnegie Institution, 1992-1999) operated more than 20 temporary broadband stations in the southeastern Brazilian shield. The area, a transect 1000 km long and 300 km wide, covers different geological provinces: the Precambrian Sao Francisco craton, the adjacent Brasiliano (700-500 Ma) fold belts, and the Parana basin of Paleozoic origin. Crustal thicknesses were estimated for 23 sites using receiver functions. For each station, receiver functions were stacked for different sets of earthquakes according to azimuth and distance. The P-to-S Moho converted phase was clearly identified at most sites. Crustal thicknesses were estimated using an average crustal P wave velocity of 6.5 km/s. Poisson's ratio of 0.23 (Vp/Vs = 1.70) was used for the Sao Francisco craton and adjacent fold belt (based on travel times from small, local earthquakes) and 0.25 was used for the Parana basin and coastal belt. Crustal thicknesses ranged from 35-47 km. Although there is a clear inverse correlation between topography and Bouguer gravity anomalies in the study area, Moho depths show the opposite pattern from that expected: areas of low topography and less negative Bouguer anomalies, such as the Parana basin, have thicker crust (40-47 km) compared with the high elevation areas of the craton and fold belt (37-43 km). Two hypothesis are proposed to explain the data: (1) A lower density, by 30-40 kg/m(3), in the lithospheric mantle under the Archean block of the Sao Francisco craton relative to the Proterozoic lithosphere is responsible for maintaining the high elevations in the plateau area. Relatively low density and high P wave velocity are compatible with a depleted (low FeO) composition for the Archean lithosphere. (2) Alternatively, if the density contrasts between Archean and Proterozoic lithospheres are smaller than the values above, then the crust beneath the Parana basin must be more dense than that of the craton. Higher crustal density and high Poisson's ratio would be consistent with magmatic underplating in the lower crust beneath the Parana basin, as inferred from other studies.