Post-collisional ultrapotassic volcanism in the Tangra Yumco-Xuruco graben, south Tibet: Constraints from geochemistry and Sr–Nd–Pb isotope

Major-trace elements and Sr–Nd–Pb isotopic data are presented for newly discovered felsic ultrapotassic lavas in the Tangra Yumco-Xuruco graben, south Tibet. The felsic ultrapotassic lavas are characterized by high SiO2 and K2O, low MgO and CaO contents. They have also high abundances of incompatible trace elements such as large ion lithophile elements (LILEs) and light rare earth elements (LREEs), especially Rb (562–893 ppm), Th (148–182 ppm) and La (101–131 ppm). One of the most striking features of these rocks is their high Th/La (1.15–1.53) and Rb/Sr (1.0–1.26) ratios with pronounced Eu and Ce anomalies (Eu/Eu⁎ = 0.56–0.60, Ce/Ce⁎ = 0.73–0.90), while the high field strength elements (HFSEs) and Sr are depleted relative to LILEs and HREEs. The felsic ultrapotassic lavas show highly radiogenic 87Sr/86Sr (0.7200–0.7282) and unradiogenic 143Nd/144Nd (0.51191–0.51197), radiogenic 206Pb/204Pb (18.794–19.063), 207Pb/204Pb (15.697–15.824) and 208Pb/204Pb (39.455–39.954). The similarity of trace element patterns and Sr–Nd–Pb isotopic components with the high-MgO ultrapotassic rocks in the province suggests that both the felsic ultrapotassic and mafic ultrapotassic lavas were derived from partial melting of the south Tibetan lithospheric mantle metasomatized by sediment-derived melts. However, the remarkable differences in some key trace element pairs between the felsic ultrapotassic lavas and mafic ultrapotassic rocks in the Tangra Yumco-Xuruco graben indicate that the felsic ultrapotassic lavas were not the highly-differentiated products from the mafic ultrapotassic magmas. Our new data suggest that these felsic ultrapotassic lavas have a distinct magma source beneath the south Tibetan crust, which is different from that of the mafic ultrapotassic lavas. Their high light-REE enrichment (La/Yb(N) = 41.98–51.41) but low heavy-REE fractionation (Dy/Yb(N) = 1.99–2.11) suggest an amphibole and phlogopite-rich mantle source in the genesis of the potassic lavas. In contrast, the mafic ultrapotassic lavas in the graben have variable elevated Dy/Yb(N) (2.1–2.8) and La/Yb(N) (39–128) ratios, reflecting their sources with residual garnet. By implication, the discrepancy in their source assemblages suggests that the depths of the felsic ultrapotassic sources could be shallower than that of the mafic ultrapotassic sources in the south Tibetan lithosphere.In addition, proposed mantle metasomatism by subducted sediment-derived melts for the genesis of the ultrapotassic lavas might also be applicable for along-strike variation in the post-collisional volcanism in south Tibet. Isotopic mixing calculation illustrates that the subducted sediments represented by the Himalayan basement are an effective reservoir for bulk assimilation of the south Tibetan lithospheric mantle, and hence sediment input of various proportions into the mantle sources could be responsible for the isotopic variation seen in the post-collisional calc-alkaline and potassic volcanics in the eastern part of the southern Tibet. Along-strike variation in the post-collisional volcanism was most probably caused by a westward increase in sediment input in the southern Lhasa block during the subduction of the Tethyan oceanic lithosphere.