10:45 〜 11:00
[SVC36-06] Geochemical variations of volcanic rocks in Mt.Fuji: Magmatic processes identified by Independent Component Analysis
キーワード：沈み込み帯火成作用、化学組成、統計解析、富士山
Mt. Fuji is a Quaternary volcano, and it has erupted approximately 400 km3 of lava flows and scoria over 100,000 years (Tsukui et al., 1986). It is located near the northern end of the Izu-Bonin volcanic arc, where the Izu-Bonin volcanic arc on the Philippine Sea Plate (PHS) and the Pacific Plate (PA) subducts beneath the Honshu arc. A previous study showed that the mantle wedge in central Japan (Chubu district) is relatively cold because of overlapping subduction of the PHS and PA (Nakamura and Iwamori, 2013), which may suggest a relatively cold environment beneath Mt. Fuji. On the contrary, Mt. Fuji is one of the volcanoes which have the highest level of eruption rates in Japan. The reason for the high eruption rate may be related to the source and evolution of magma in Mt. Fuji. The lavas in Mt. Fuji exhibit a relatively narrow range in chemical composition (e.g., SiO2 content of 49-52 wt%), although tFeO/MgO ratio (1.4-3.1) and incompatible elements concentrations exhibit somewhat wider ranges. These characteristics are not common to many other volcanoes along the Honshu arc and cause difficulties in capturing the magma processes in Mt. Fuji. In this study, we constructed a new geochemical dataset consisting of 946 lava samples based on the data from the previous study (Takahashi et al., 2003). The purpose of this study is to capture the geochemical characteristics of lavas in Mt. Fuji by using a statistical method, independent component analysis (ICA), and whitened data-based k-means cluster analysis (KCA).
Using ICA, we found four independent components (IC 1 to 4) which can be directly related to magma processes. Three of the four can be explained by the forward models of crystal fractionation under different conditions. These three processes reflected the differences in mineral combinations depending on physical-chemical conditions such as the depth and temperature of the magma reservoirs, and the water content of the magma. IC 1 indicated crystallization with increasing SiO2 content, which is commonly seen in arc lavas. IC 3 captured crystal fractionation occurred at a depth greater than ~5 kbar. IC 4 represents the crystal fractionation of mafic minerals with a relatively small amount of plagioclase crystals. This can occur by several mechanisms, e.g., the delay in plagioclase crystallization in a melt with a high-water content or plagioclase phenocrysts floating in a relatively dense melt in a large magma reservoir. IC 2, the independent component remaining to be explained, is characterized by a relatively wide variation in incompatible element concentrations with nearly constant SiO2 and MgO content, and IC2 discriminates Hoshiyama-Age (Older-Fuji) from the younger stages. Plausible crystal fractionation of a constant set of mineral assemblages and compositions cannot account for IC 2 based on the variation of major elements. Other mechanisms and factors such as variable crystallization and source materials/conditions would be required to explain the difference between Older-Fuji and younger lavas.
Using ICA, we found four independent components (IC 1 to 4) which can be directly related to magma processes. Three of the four can be explained by the forward models of crystal fractionation under different conditions. These three processes reflected the differences in mineral combinations depending on physical-chemical conditions such as the depth and temperature of the magma reservoirs, and the water content of the magma. IC 1 indicated crystallization with increasing SiO2 content, which is commonly seen in arc lavas. IC 3 captured crystal fractionation occurred at a depth greater than ~5 kbar. IC 4 represents the crystal fractionation of mafic minerals with a relatively small amount of plagioclase crystals. This can occur by several mechanisms, e.g., the delay in plagioclase crystallization in a melt with a high-water content or plagioclase phenocrysts floating in a relatively dense melt in a large magma reservoir. IC 2, the independent component remaining to be explained, is characterized by a relatively wide variation in incompatible element concentrations with nearly constant SiO2 and MgO content, and IC2 discriminates Hoshiyama-Age (Older-Fuji) from the younger stages. Plausible crystal fractionation of a constant set of mineral assemblages and compositions cannot account for IC 2 based on the variation of major elements. Other mechanisms and factors such as variable crystallization and source materials/conditions would be required to explain the difference between Older-Fuji and younger lavas.