Mantle melting beneath the Tibetan Plateau: Experimental constraints on ultrapotassic magmatism

Phase equilibrium experiments on primitive Miocene olivine leucitite (Bb‐107) from the Qiangtang terrane of the Tibetan Plateau were performed from 1.0 to 2.2 GPa and 1270 to 1440°C. The composition is multiply saturated with olivine and clinopyroxene from 1.2 to 2.2 GPa and 1340°C under nominally anhydrous conditions. Phase assemblages in the experiments have been used to model the effects of high‐pressure fractional crystallization. The results are consistent with an origin of Bb‐107 as a modified mantle melt produced by fractional crystallization in the lithospheric mantle or lower continental crust. Liquids from spinel + garnet peridotite melting experiments are compositionally similar to the primitive fractionation corrected melt. If metasomatized mantle contained H 2 O when melting began, the depth of melting approaches the base of the Moho and is 1300°C at 2 GPa for 5 wt % H 2 O in the melt and 1360°C at 2.2 GPa for 2 wt % H 2 O, and 1420°C at 2.4 GPa for a dry melt. Major and trace element evidence from Bb‐107 and other primitive Tibetan shoshonitic lavas indicates that these magmas may be derived as low‐extent melts (1–3 wt %) of a metasomatized mantle in the spinel and garnet stability fields. The depth of melting and the geochemical characteristics of the Tibetan lavas are correlated. Increased extents of melting correlate with increasing depth of melting. This correlation is to be expected if melting occurred during lithospheric thinning during downward convective flow at the margin of the lithosphere–asthenosphere boundary.