Partial melting of mantle wedge and evolution of island arc crust

Abstract

Evolution of the crust in the northeast Honshu arc is discussed on the basis of the melting experiments of a peridotite under hydrous conditions and the crustal compositions estimated from deep crustal xenoliths and seismic velocity data. The chemical compositions of melts formed by partial melting of a hydrous peridotite within the pressure‐temperature conditions for shallow parts of the mantle wedge (1.2–2.0 GPa) have been experimentally determined with the “sandwich experiment” technique. The melts range from olivine tholeiitic to magnesian andesitic compositions (SiO 2 48–56 wt %; MgO 9–14 wt %) beneath the volcanic front in the northeast Honshu arc and are slightly less silicic beneath the coast of the Sea of Japan. The crustal compositions of the volcanic zone in the northeast Honshu arc estimated from deep‐seated crustal inclusions in volcanic rocks and the velocity structures of the crust are andesitic in the central zone and high‐alumina basaltic near the coast of the Sea of Japan. The compositions of possible primary melts formed in the mantle wedge are significantly more mafic than these crustal compositions. Consequently, if the major part of the crustal materials in the volcanic zone was formed by magmas generated in the mantle wedge, thick (up to 30 km) ultramafic cumulates including olivine pyroxenite must have been formed beneath Moho during fractional crystallization and would constitute the upper part of the mantle wedge. The volume of the mantle wedge consisting entirely of even fertile peridotite was not sufficient to produce crustal materials and cumulates in the volcanic zone by such processes in the northeast Honshu arc. A static mantle wedge model is, therefore, not satisfactory for the growth of the crust in northeast Honshu arc. A dynamic model is suggested in which the mantle materials have been supplied into the mantle wedge from the back arc side by a convective flow.