HR: 10:35h
AN: T31H-02    [PDF]
TI: The Causes of Melt Differentiation at the Izu Arc Volcanic Front
AU: * Straub, S M
EM: smstraub@ldeo.columbia.edu
AF: Lamont Doherty Earth Observatory, 61 Route 9W, Palisades, NY 10964 United States
AB: While it is common consensus that mafic arc magmas (basalts to magnesian andesites) are partial melts of the upper mantle, the provenance of the evolved arc magmas (dacites and rhyolites) remains much contested. Classic models suggest either fractional crystallization or partial melting of the upper plate crust, or a combination of both, as causes of melt differentiation. Unfortunately, it is difficult to test such models in natural systems since most arc volcanic rocks are fully crystallized, and liquid compositions cannot be directly compared to their cogenetic phenocrysts. Such possibilities, however, arise from the Cenozoic fallout tephras from the intraoceanic Izu Bonin arc. The tephra melts originate from similar mantle sources as the low-K Quaternary Izu arc front (Izu VF) volcanic rocks. Individual fallout layers from single eruptions are commonly zoned and frequently contain a range of basalt to rhyolite glass shards together with plagioclase (An$_{42-96}$), clinopyroxene (En$_{34-75}$), orthopyroxene (En$_{41-73}$) and titanomagnetite (Usp$_{14-50}$). Trace amounts of Cl-apatite (Cl=0.8-2.4 wt%) are confined to high-silica tephras, whereas olivine is absent despite its presence in the Izu VF basalt lavas. The Cenozoic tephra glasses (approximately 1500 individual glasses from 43 layers) display coherent elemental systematics through time. At any given age, the tephra glasses display a distinct bimodal distribution, with maxima at ~53-54 wt% SiO$_{2}$ (basaltic andesitic) and ~70-72 wt% SiO$_{2}$ (rhyolitic), respectively. Basaltic andesitic glasses overlap widely with the Izu VF lavas, whereas the dacitic-rhyolitic pole is almost exclusively represented by the tephra. Chemical and petrographic evidence of melt mixing is ubiquitous in all tephras, indicating melt mixing as important generic process. The incompatible element K varies by a factor of two in abundance at any given SiO$_{2}$. The linear mixing trends of K$_{2}$O vs. SiO$_{2}$ in individual fallout tephras, however, always have similar slopes with K$_{2}$O being always more enriched in the more siliceous glasses. It is this uniformity of the K$_{2}$O zoning, that effectively rules out that the zoned tephra melts formed by mixing of mafic and siliceous component melts that originate from either mantle and upper crustal sources (i.e. mixing of mantle melts with crustal partial melts), or by mixing of derivate melts that stem from different batches of mantle melts. Therefore, the only viable process of upper crustal differentiation appears to be mixing of cogenetic basaltic-andesite and dacitic-rhyolitic component melts that evolved by fractional crystallization from a single batch of mantle melt, and that became mixed during eruption. However, quantitative models cannot reproduce the intra-layer tephra zonation by fractional crystallization processes. Therefore, I suggest that the subducting slab may possibly play a role in the generation of the siliceous Izu VF melts. A scenario in where mafic mantle melts mix incompletely with hydrous, K$_{2}$O-bearing siliceous fluids from the slab appears be able to explain many of the chemical and petrographic features observed in both the IzuVF tephra and lavas.
DE: 1010 Chemical evolution
DE: 1020 Composition of the crust
DE: 1030 Geochemical cycles (0330)
DE: 8404 Ash deposits
SC: Tectonophysics [T]
MN: 2003 Fall Meeting