An amphibolitic source for “adakitic” I-type plutons in continental collision zones

Rocks with “adakitic” affinities are widespread in continental collision zones, recording important information about crustal thickening. Residual garnet during partial melting of the lower crust is traditionally considered to be responsible for heavy rare earth element and Y depletion, which is a key characteristic feature of these rocks. However, magmatic garnet is rarely observed in such adakitic igneous rocks and associated protolithic eclogite is also rarely reported. Instead, amphibole is a proposed abundant phase in lower arc crust. Whilst both garnet- and amphibole-rich assemblages can act as potential sources for intracrustal intermediate-silicic melts, their distinctive fractionation of fSm/Nd and fLu/Hf should, with time, lead to different Nd-Hf isotopic compositions in high Sr/Y rocks. Here we report whole-rock radiogenic Sr-Nd-Hf and stable Fe isotope compositions of well-characterized Oligocene to Miocene (33-14 Ma), post-collisional high-K adakitic intrusions from the Gangdese belt, South Tibet. The plutons have a low alumina saturation index and depleted Dy/Yb, pointing to negligible amounts of sediment contribution. They are characterized by variable initial 87Sr/86Sr ratios (0.7048 to 0.7078), mostly negative εNd (−4.9 to +0.3), positive εHf values (+1.5 to +8.3), and variable but heavy δ57Fe values ranging from +0.04 ± 0.08‰ to +0.35 ± 0.08‰. Modeling of Fe isotopes and fractionation factors of fSm/Nd and fLu/Hf indicate an amphibolite source. A Nd-Hf isotope correlation that is quasi-parallel to the mantle array can be best explained by the evolution of amphibolitic lower crust with a residence time of ca. 400 Myrs. Although these adakitic rocks have depleted, mantle-like Hf isotope compositions, they are part of an existing, older crustal foundation and thus represent a process of crustal reworking, instead of crustal growth. Therefore, proportions of mantle-derived materials that contributed to the Phanerozoic crustal growth based on Hf-in-zircon signatures alone may be overestimated. Combined mineral and whole-rock radiogenic and stable isotopes are required to address this issue in the future.