Transportation of H2O beneath the Japan arcs and its implications for global water circulation

In order to understand the transportation and circulation of H2O in subduction zones, geological and seismological observations and the corresponding quantitative modeling are reviewed and consolidated, with a special emphasis on the Japan arcs. The Japan arcs are ideal for such a study in that (1) they are very active magmatically and seismically with abundant high-quality observations, and (2) subduction parameters change significantly along the Japan arcs allowing one to investigate how the variations of ‘input’ parameters affect the resultant structures and processes. First, numerical models for the transportation of H2O and melting are introduced. The generation and migration of aqueous fluid, its interaction with the convecting solid, and melting are considered, based on the realistic phase relationships of the peridotitic and basaltic systems. Application of the models to the Japan arcs and comparison between the model results and the observations (distribution of volcanoes and seismic structures) along the Japan arcs are then discussed. The three cases from the NE, central and SW Japan arcs clarify the following points. First, an aqueous fluid released from the subducting oceanic crust forms a serpentinite layer in the mantle wedge just above the subducting slab, maintaining a condition close to equilibrium in terms of hydration–dehydration reactions during aqueous fluid migration. Second, most of the H2O is subducted to a depth where serpentine and chlorite in the serpentinite layer break down. This depth depends on the thermal structure of the slab, and is greater for the older plate beneath the Japan arcs. As a consequence, aqueous fluid and melt are, in general, not supplied straight upward to the volcanic front. This result is rather different from the conventional view. Finally, based on the understanding obtained from the Japan arcs, the global budget and circulation of H2O are discussed. Even after the completion of dehydration of major hydrous mineral phases in the subducting materials, especially at the base of the mantle wedge just above the subducting slab, nominally anhydrous phases of the upper mantle assemblage (i.e., olivine, pyroxenes, garnet) can carry a significant amount of H2O (1.1 × 1011 to 7.8 × 1011 kg yr− 1) into the deep mantle. The minimum estimate is comparable to the outflow of H2O from the mantle at mid-ocean ridges and hotspots, while the maximum estimate greatly exceeds it. The large estimated range is mainly due to the uncertainty of the maximum H2O contents in the nominally anhydrous minerals of the upper mantle, which controls the net influx of H2O into the deep mantle through subduction zones.Water; Subduction; Magmatism; Global circulation