A review of carbonatitic magmatism in the Paran\ue1-Angola-Namibia (PAN) system.

Mesozoic to Cenozoic alkaline-carbonatitic complexes from southern Brazil, Angola and Namibia occur along main tectonic lineaments. In general, the alkaline-carbonatite complexes show intrusive/subintrusive, subcircular or oval shaped structures and are indicative of high upwelling energy. Processes of liquid immiscibility from trachytic-phonolitic liquids, starting from parental alkaline mafic magmas are believed to have generated carbonatitic liquids, as suggested by field relationships and geochemical characteristics. Ca-, Mg- and Fe-carbonatites are widespread even in the same complex. The occurrences comprise three main chronogroups, i.e. 1) Early Cretaceous (Eastern Paraguay; Brazil, Ponta Grossa Arch and Anitapolis; Angola and Namibia); 2) Late Cretaceous (Brazil , Ponta Grossa Arch, Lages and Alto Paranaiba. Namibia); 3) Paleogene, Brazil and Namibia Two principal types of associated alkaline rocks are represented, i.e. plagioleucitites l.s. (Eastern Paraguay; Brazil: Ponta Grossa Arch- Angola and Namibia) and kamafugites l.s. (Brazil: alto Paranaiba and Lages; Namibia). Significant variations in O-C isotope compositions are found in primary carbonates , the variations being mainly due to isotope exchange between carbonates and H2O-CO2-rich hydrothermal fluids, whereas magmatic processes, i.e. fractional crystallization or liquid immiscibility, probably affect the delta O-18 and delta C-13 values by not more than 2 delta%.. The isotope exchange model implies that the most significant isotopic variations took place in a hydrothermal environment, e.g. in the range 400-80 degrees C, involving fluids with CO2/H2O ratio ranging from 0.8 to 1. Sr-Nd-Pb isotope systematics highlight heterogeneous mixtures between HIMU and EMI mantle components, similar to the associated alkaline rocks and the flood tholeiites of the Parana-Angola-Etendeka (Namibia) system. This is also consistent with Re-Os systematics on selected mafic samples from the Alto Paranaiba alkaline-carbonatite province. The data relative to the noble gases suggest that the source(s) are similar to other mantle derived magmas (e.g. HIMU and MORB) and that the carbon of carbonatites is unlikely to be subduction-related carbon, and support a C-O fractionation model starting from mantle-derived sources. In spite of the strong variation shown by C-O isotopes, Sr-Nd-Pb-Os isotopic systematics could be related to an isotopically enriched source where the chemical heterogeneities reflect a depleted mantle "metasomatized" by small-volume melts and fluids rich in incompatible elements. These fluids are expected to have promoted crystallization in the mantle of K-rich phases that gave rise to a veined network variously enriched in LILE and LREE (cf. Foley, 1992b). The newly formed veins (enriched component) and peridotite matrix (depleted component) underwent a different isotopic evolution with time as reflected by the carbonatitic rocks. These conclusions may be extended to the whole Parana-Angola-Etendeka system, where isotopically distinct parent magmas were generated following two main enrichment events of the subcontinental lithospheric mantle at 2.0-1.4 and 1.0-0.5 Ga, respectively, as also supported by Re-Os systematics. The mantle sources preserved the isotopic heterogeneities over a long time, suggesting a non-convective lithospheric mantle beneath different cratons or intercratonic regions. Overall the data indicate that the alkaline-carbonatitic magmatism originated from a significant, but small scale heterogeneous subcontinental mantle. In this scenario, the Tristan da Cunha, Walvis Ridge-Rio Grande Rise and Vitoria-Trindade hotspot tracks might reflect the accomodation of stresses in the lithosphere during rifting, rather than continuous magmatic activity induced by mantle plumes beneath the moving lithosphere