Miyakejima, a volcano island, resting on the Izu-Mariana arc is an active basaltic stratovolcano. The eruptive activity is divided into five stages; (1) Ofunato stage (10,000 y. B. P.-4,000 y. B. P.), (2) Tsubota stage (4,000 y. B. P.-2,500 y. B. P.), (3) Oyama stage I (2,500 y. B. P.-1,300 y. B. P.), (4) Oyama stage II (1,300 y. B. P-A. D.1469), (5) Shinmio stage (A. D. 1469-1983 eruption). Evolution of the magma plumbing system during the last 10,000 years is discussed on the basis of petrological investigations of systematically collected rocks from various stratigraphic levels. All five stages consist of a number of lava flows and scoria fall deposits. SiO₂ content shows a range from 48.6 wt.% to 63.2 wt.%. Since 4,000 years ago, whole-rock Mg# (=Mg/(Mg+Fe)×100) systematically decreases at the initial eruption of each stage. These variations can be explained basically by fractional crystallization. On the other hand, mineral compositions and petrographical characters indicate that most of the eruptive products during the last 10,000 years contain three different types of phenocrysts (Type 1, Type 2 and Type 3), which can not equilibrate with each other. Different types of phenocrysts often exist within a single rock sample. Type 1 phenocrysts consist of calcic plagioclase (An90) and magnesian olivine (Fo80). Considering Fe-Mg partitioning between liquid and olivine, olivine of Type 1 phenocrysts can not equilibrate with any whole-rock compositions. Type 1 phenocrysts occur in Ofunato stage and in the earlier period of each stage. Abrupt increase of whole-rock Mg# strongly suggests that relatively undifferentiated magma carrying Type 1 phenocrysts was supplied into the preexisting magma plumbing system. Phenocrysts of Type 2 (plagioclase, olivine, and clinopyroxene) and Type 3 (plagioclase, clinopyroxene, orthopyroxene, and magnetite) are able to equilibrium with basaltic magma and andesitic magma, respectively. Type 2 phenocrysts occur since Hatchodaira eruption (2,500 y. B. P.). Type 3 phenocrysts occur in Tsubota and the later stage. Subsequently, whole-rock Mg# decreased by fractionation of Type 2 phenocrysts in Oyama and Shinmio stage and Type 3 phenocrysts in Tsubota stage.

View full abstract

A new scheme for hypocenter calculation is proposed for volcanic regions. The scheme finds the global minimum of the travel time residual, and the resulting solution is more stable than that from the conventional Geiger's method, especially when the number of observations is small or the station distribution is unsuited. In the first step (Step 1) of this scheme, the target area js parameterized by the node. The node distance depends on the heterogeneity of the velocity structure; usually about 0.3-1 km is sufficient. Travel times between nodes and stations are stored. In the first half of the second step (Step 2A), a node is sought that minimizes the sum of squares of the travel time residuals for the event. By changing the depth of this node, a set of initial hypocenter candidates is generated. In the following step (Step 2B), with these candidates for the initial hypocenter, precise locations of the event are obtained by nonlinear calculations using the simplex method. The hypocenter with minimum travel time residual is then selected as the most probable hypocenter. For the travel time calculation, the 3-D robust seismic ray tracer known as Fermat (Nishi, 2001) is used. If the velocity structure and the locations of the observation stations do not change, the results of Step 1 are effective for all events. Consequently, repetitions of only Step 2A and 2B are sufficient for every event. Successful outcomes of calculation using synthetic and actual data are obtained with practicable CPU times.

View full abstract