Noble gas and halogen records of subduction fluids along the paleo subduction boundary at the base of the shallow mantle wedge: example of the Cretaceous Sanbagawa belt, Japan

Deep slow earthquakes are generally recognized where subducting oceanic plates are in contact with the shallow mantle. High fluid pressure is commonly invoked as an important factor in the generation of slow earthquakes and the associated fluids are thought to be derived from the breakdown of hydrous minerals such as chlorite and serpentine. When considering how such fluids may be related to earthquake generation it is important to consider both the quantities and fluid paths ways. Buoyancy will tend to drive the fluids vertically upwards. But serpentinite developed along the base of the mantle wedge may act as an effective seal and cause flow to be channelled along the subduction boundary. Numerous serpentinite bodies are exposed throughout the Cretaceous subduction-type Sanbagawa belt of SW Japan. These serpentinite bodies are derived from the wedge mantle and the adjacent metamorphic units, consisting primarily of mafic, quartz, and  pelitic schists are derived from basalt, chert and mudstone of the crust of the subducting slab. The boundary between the serpentinite bodies and the schists therefore represents the paleo subduction boundary and the rocks along this boundary are a potentially important record of the way in which subduction fluids move. We highlight the characteristics of two separate kilometer-scale bodies, the Shiragama Yama body (SY) which rose from depths of ~35 km and the Kamabuse Yama (KY) body which rose from depths of ~25 km. The mineralogy of SY suggests SiO2 transported by hydrous fluids is restricted to a ~70 m thick shear zone at the base consisting of high-T serpentine, antigorite. In contrast, KY shows evidence for pervasive SiO2 enrichment with a more limited zone sheared zone consisting dominantly of low-T serpentine, chrysotile. Analyses of noble gas and halogen of the boundary domain lithologies were performed to help identify the source of fluids related to hydration and material transport. In SY these data support the idea that far-travelled fluids were channeled along the subduction boundary. The results for KY are more complex with distinct fluids responsible for serpentinization (1) and metasomatism (2). The metasomatism can be further divided into a chlorite-forming stage (2-1) and a later talc-forming stage (2-2). The sources for the fluids involved in each of these stages were likely derived from (1) altered oceanic crust, (2-1) altered oceanic crust + sedimentary porosity, and (2-2) sedimentary porosity + serpentinized slab.

Combining the results from SY and KY suggests that channeling of subduction fluids in the Sanbagawa subduction zone was important at depths of around 35 km but was less effective at shallower levels. The change in the source region of the fluids with time shown in KY, suggests fluid flow may become more channelized as a shear zone develops along the subuction interface.