Differentiation of continent crust by cumulate remelting during continental slab tearing: Evidence from Miocene high-silica potassic rocks in southern Tibet

The role played by felsic igneous magmatism in crustal differentiation is of importance for the evolution and habitability of Earth system. However, the petrogenesis of felsic rocks and their genetic links to mafic magmas remain controversial, hindering our understanding of crustal differentiation processes. Cenozoic post-collisional felsic potassic rocks occur widely in the well-known Himalaya-Tibet orogen and are intimately associated, in space and time, with mantle-derived ultrapotassic rocks, which thus provide a clear case study to investigate the genesis of felsic rocks and their genetic relationship with mafic rocks. Here we report zircon U-Pb ages and Hf-O isotopes, mineral major elements, and whole-rock elemental and Sr-Nd-Pb isotopic compositions for the TangraYumco high-silica potassic rocks (trachytes) in central Lhasa block of southern Tibet. The results, combined with literature data of the TangraYumco ultrapotassic rocks, can help better illuminate their genesis and the role of ultrapotassic rocks in their generation. The trachytes erupted in the Miocene (~13 Ma), coeval with ultrapotassic lavas in this area. They have very enriched Sr-Nd-Pb-Hf-O isotope compositions similar to those of ultrapotassic rocks. However, they display more fractionated LREEs (light rare earth elements) than ultrapotassic rocks. The trachytes also show nearly constant Y and Yb contents with increasing SiO2. This indicates that they were not derived by fractional crystallization from ultrapotassic magmas or by mixing between ultrapotassic and ancient crust-derived felsic magmas. Instead, given their lower LREE contents than ultrapotassic rocks and occurrence of clinopyroxene antecrysts, we suggest that the TangraYumco trachytes were likely produced by remelting of newly-formed, Sr-Nd-Pb isotopically enriched, and K-rich cumulates that crystallized from ultrapotassic magmas. Moreover, taking into account regional magmatism and tectonics, we propose that the TangraYumco-Xuruco post-collisional coeval magmatism and rift system represent a response to tearing of subducted Indian continental slab. Finally, in combination with the nature of post-collisional magmatism in the Lhasa block, we suggest that in addition to reworking of juvenile and ancient crust, underplating of postcollisional mafic magmas and remelting of their cumulates could be an important way to differentiate continental crust in collisional orogeny zones.