Oligocene High-MgO Alkali Basalts in Central Tibet: Implications for Magma–Mush Mixing and Mantle Processes

Abstract High-MgO (>9 wt %) basaltic rocks can be primary magmas and are used to constrain the geochemistry and temperature of the mantle. However, high MgO contents can also result from mixing between evolved melts and antecrysts or xenocrysts, and thus, the whole-rock composition might not represent the solidified equivalents of primary magma. Whether such mixing with crystals can result in erroneous interpretations of mantle processes remains unclear. This study presents a petrological and geochemical investigation of the post-collision high-MgO (>9 wt %) Lugu volcanic rocks in the southern Qiangtang terrane, central Tibet. The Lugu volcanic rocks comprise porphyritic and intersertal alkali basalts. Zircon U–Pb ages and 40Ar/39Ar dating suggest that the two types of alkali basalts were erupted at c. 29 Ma. Based on detailed petrographic observations and geochemical analysis, the porphyritic alkali basalts may represent near-primary melts, which are characterised by low SiO2 contents (40.9–45.1 wt %), high CaO/Al2O3 ratios (1.1–1.5), and arc-like trace element patterns. We suggest these basalts were derived by partial melting of enriched garnet peridotite (>3 GPa) in the presence of H2O and CO2. These geochemical features are different from those of the c. 30-Ma (ultra)-potassic rocks in the Qiangtang terrane, indicating that a heterogeneous lithospheric mantle existed beneath the Qiangtang terrane during the Oligocene. In contrast, although the intersertal alkali basalts have high MgO contents (>9 wt %), evidence from mineral chemistry indicates that the whole-rock compositions of the intersertal alkali basalts represent mixtures of evolved residual melts and cumulate crystals. They were the product of polybaric fractional crystallisation and the subsequent mixing of crystals and residual melts in a magmatic plumbing system. Furthermore, when intersertal alkali basalts are assumed to be primary melts, they would have been derived by partial melting of shallow (~2.5 GPa) CO2-poor pyroxenite or peridotite. These conditions are different from interpretations of the nature of the mantle source and melting conditions for porphyritic alkali basalts. Our results highlight that the interpretation of petrogenetic processes should be preceded by detailed mineralogical investigations.