抄録
Mantle melting and production of magmas in NE Japan may be controlled by locally developed hot regions within the mantle wedge that form inclined, 50 km-wide fingers. In this case, are these hot fingers chemically and/or isotopically different from the host mantle wedge? Forty-four Quaternary volcanoes in NE Japan have been reviewed to evaluate twodimensional strontium isotopic variations, and to infer 87Sr/86Sr contours of the source mantle. The isotopic composition of magma source materials at depth is found to have little relationship with slab depth, suggesting that mantle heterogeneity was established before the flux of fluid released from the subducting slab reached the magma source regions. On the other hand, Miocene Japan Sea back-arc Yamato basin basalts have the same isotopic variation as the Quaternary volcanic arc. Cousens et al. (1994) suggested the possibility that partial melts of sediments, forming at a depth of >200 km may mix with mantle wedge material (87Sr/86Sr0.703), resulting in a magma source component with enriched 87Sr/86Sr of0.705. I suggest that after the cessation of Yamato basin rifting, a MORB-like mantle source (87Sr/86Sr0.703) in the mantle wedge below the Quaternary NE Japan arc was replenished by a fertile mantle material (87Sr/86Sr0.705) through convection induced by the subducting lithosphere. On its way to the shallower mantle wedge (<150 km), the fertile mantle material changes shape from a hot sheet to hot fingers, for reasons not yet fully understood. Thus, the hot fingers, with 87Sr/86Sr of0.705, extend from150 km below the back-arc region towards the shallower mantle (50 km) beneath the volcanic front. A conveyor-like return flow is interpreted to carry the remnants of these fingers to depth, resulting in greater amounts of fertile material being incorporated in diapirs beneath the volcanic front, and smaller amounts incorporated in areas behind the front.
