Olivine Melt Inclusion Constraints on Some Intensive Properties of Subvolcanic Crystal Mushes and Their Evolution through Boundary Layer Fractionation in Northern Japan

Abstract Magma differentiation in arc settings has usually been attributed to an interplay of processes (fractional crystallization, assimilation, and magma mixing). Homogeneous fractional crystallization has been widely used to model the magmatic evolution of volcanic systems in arc settings due to its simplicity, even though boundary layer fractionation (BLF) has been proposed as a preponderant process of differentiation in hydrous magmatic systems. Both models produce distinct compositional paths and the application of the wrong model yields erroneous estimates of parameters like pressure–temperature-H2O conditions and primary melt compositions. Melt inclusion (MI) populations corrected for post-entrapment processes have the potential to help discriminate between these two types of fractional crystallization, as their compositions are not affected by crystal accumulation and should capture the magmatic evolution as crystallization occurs. In this study, olivine-hosted MIs are used to assess the differentiation trends of basic arc magmas in northern Japan. Differentiation trends from five arc volcanic systems in northern Japan show that BLF is ubiquitous. Homogeneous fractionation models are unable to explain the liquid lines of descent of minor elements, like TiO2 and P2O5. To reproduce these differentiation trends, the presence of accessory phases like titanomagnetite or apatite are required, which in many cases are not equilibrated by the melt or need to be fractionated in amounts that are incompatible with homogeneous fractionation. The prevalence of BLF in all studied arc magmas of northern Japan indicates that solidification fronts are key environments in the crustal evolution of some hydrous subduction zone magmas.