Résumé
The northern part of South America has undergone several major changes in its tectonic setting since the Triassic, evolving from rifting to the development of a retroarc-foreland basin on its Andean margin, leading to margin extension and continental uplift following the breakup between Africa and South America. So far, it is unclear when and how these geodynamic events affected the paleo-Amazonian drainage. In this study, we investigate the provenance of Triassic–Eocene sedimentary rocks deposited in the northern Peruvian Amazonian basin, based on their Sr-Nd isotopic compositions and U-Pb zircon dating. The Triassic–Jurassic samples have εNd(0) values ranging from −7.9 to −10.0 and main U-Pb zircon peaks at 0.9–1.3 Ga (31–33%) and 0.5–0.7 Ga (21–28%) that suggest a mixed clastic supply from the Western and Eastern Cordillera or craton within a rift to post rift setting. Samples from the Albian–Maastrichtian interval yield much lower εNd(0) values (−16.8 to −18.6) and a dominance of zircon grains derived from terranes in the easternmost Brazilian Shield (Ventuari-Tapajos (2.0–1.82 Ga), Rio Negro-Jurena (1.82–1.54 Ga) and Rondonia San Ignacio (1.54–1.3 Ga), thus indicating a cratonic source for these sedimentary rocks. Finally, the early Paleocene–Eocene sedimentary rocks record the first arrival of Andean detritus in the Amazonian retroarc foreland, with εNd(0) values ranging between −5.6 to −12.0 and up to 16% of the zircon grains yielding ages younger than 120 Ma. Together with recently published studies, these provenance data document the existence of a long-lived Aptian–Maastrichtian continent-wide cratonic drainage in the northern part of South America. This cratonic drainage developed in response to the late Cretaceous uplift of the northeastern part of the South America craton likely driven by geodynamic processes related to post rift opening of the Atlantic Equatorial Ocean. The formation of a late Maastrichtian–early Paleocene mountain chain in the Peruvian Andes associated with the onset of provenance from the Andean orogenic belt led to a shift of the Amazon drainage pattern. Hence, the earliest Andean-Amazonian rivers are no younger than late Maastrichtian-early Paleocene.