Can slabs melt beneath forearcs in hot subduction zones?

Abstract:

At subduction zones, thermal modeling predict that the shallow part of the downgoing oceanic crust (< 80 - 100 km depth to the slab) is usually too cold to cross the water-rich solidus and melts beneath the forearc. Yet, the occasional occurrence of adakites, commonly considered as slab melts, in the forearc region challenges our understanding of the shallow subduction processes. Adakites are unusual felsic rocks commonly associated with asthenospheric slab window opening or fast subduction of young (< 25 Ma) oceanic plate that enable slab melting at shallow depths; but their genesis has remained controversial. Here, we present a new approach that provides new constraints on adakite petrogenesis in hot subduction zones (the Philippines) and above an asthenospheric window (Baja California, Mexico). We use amphibole compositions to estimate the magma storage depths and the composition of the parental melts to test the hypothesis that adakites are pristine slab melts. We find that adakites from Baja California and Philippines formed by two distinct petrogenetic scenarios. In Baja California, hydrous mantle melts mixed/mingled with high-pressure (HP) adakite-type, slab melts within a lower crustal (~30 km depth) magma storage region before stalling into the upper arc crust (~7-15 km depth). In contrast, in the Philippines, primitive mantle melts stalled and crystallized within lower and upper crustal magma storage regions to produce silica-rich melts with an adakitic signature. Thereby, slab melting is not required to produce an adakitic geochemical fingerprint in hot subduction zones. However, our results also suggest that the downgoing crust potentially melted beneath Baja California.