10:45 AM - 12:15 PM
[SVC29-P08] Timescales from mixing to eruptions: the case of historical Plinian eruptions at Sakurajima volcano
Keywords:Eruption trigger, Elemental diffusion, Plinian eruption, Sakurajima Volcano
Mafic magma recharge of a crustal reservoir and subsequent mixing have been considered to trigger silicic explosive eruptions. However, in many active volcanoes, recharges have been frequently monitored geophysically but rarely led to large-scale eruptions, which makes their role as triggers questionable. A detailed reconstruction of the magmatic processes leading to past volcanic eruptions is essential to deepen our understanding of the triggering mechanisms. Sakurajima Volcano, Japan, is a suitable subject for this purpose. The three Plinian eruptions since the 15th century (AD 1471, 1779, and 1914 eruptions) were fed by mixed andesitic-dacitic magmas, and a comprehensive geophysical network has accurately uncovered the structure of the current sub-volcanic system. Araya et al. (2019, JpGU meeting) reported that magnetite phenocrysts in the 1471, 1779, and 1914 Plinian pumices were chemically homogeneous within each crystal, but their chemistry varied among crystals in each eruption, indicating no magmatic perturbation within the timescale of diffusive homogenization of magnetite (approximately a few months). However, this is a minimum estimate; thus, it remains unclear when pre-eruptive magma mixing occurs. This study investigated the zoning in orthopyroxene, which has a longer diffusion timescale than magnetite, to constrain the timing of magma mixing prior to the Sakurajima historical Plinian eruptions.
The orthopyroxene phenocryst cores have Mg# (=Mg/(Fe+Mg) in mol%) values of 53–71, 54–71, and 61–73 for the 1471, 1779, and 1914 eruptions, respectively. Almost all the orthopyroxene phenocrysts were zoning-less or exhibited blurred zoning. Sharp reverse zoning was rare. For the 1914 eruption, only one grain showed sharp zoning among the 160 orthopyroxene phenocrysts observed. Core-to-rim zoning profiles were obtained for 20 reversely zoned orthopyroxene phenocrysts from the 1914 eruption, and the diffusion time of Mg–Fe was calculated. Nineteen of the 20 orthopyroxene grains exhibited a diffusion time of 2.6–231 years. One orthopyroxene grain, which have sharp reverse zoning previously mentioned, showed a distinctly shorter diffusion time of 0.13 years. The rim of this orthopyroxene has a Mg# of 68 mol%, which is lower than the equilibrium with a mafic endmember magma (~74 mol%; Yanagi et al., 1991, Geochemical Journal), and therefore the sharp zoning can be attributed to the mixing of local heterogeneities within the magma reservoirs. Consequently, it was inferred that there was no mixing with mafic magma just before the 1914 eruption. Although the Mg–Fe diffusion time was not calculated for the 1471 and 1779 eruptions, the same probably holds true for the two eruptions because the orthopyroxene and magnetite in the 1471 and 1779 Plinian pumices have characteristics similar to those of the 1914 Plinian pumice. Thus, the mafic magma recharge and subsequent mixing may not have been the immediate causes of the historical Plinian eruptions. It has been shown that the Plinian magmas were emplaced in a shallow conduit (1–3 in depth) prior to the eruptions (Araya et al., 2019, Scientific Reports). Crystallization-induced vesiculation within the conduit could have finally triggered the historical Plinian eruptions.
The orthopyroxene phenocryst cores have Mg# (=Mg/(Fe+Mg) in mol%) values of 53–71, 54–71, and 61–73 for the 1471, 1779, and 1914 eruptions, respectively. Almost all the orthopyroxene phenocrysts were zoning-less or exhibited blurred zoning. Sharp reverse zoning was rare. For the 1914 eruption, only one grain showed sharp zoning among the 160 orthopyroxene phenocrysts observed. Core-to-rim zoning profiles were obtained for 20 reversely zoned orthopyroxene phenocrysts from the 1914 eruption, and the diffusion time of Mg–Fe was calculated. Nineteen of the 20 orthopyroxene grains exhibited a diffusion time of 2.6–231 years. One orthopyroxene grain, which have sharp reverse zoning previously mentioned, showed a distinctly shorter diffusion time of 0.13 years. The rim of this orthopyroxene has a Mg# of 68 mol%, which is lower than the equilibrium with a mafic endmember magma (~74 mol%; Yanagi et al., 1991, Geochemical Journal), and therefore the sharp zoning can be attributed to the mixing of local heterogeneities within the magma reservoirs. Consequently, it was inferred that there was no mixing with mafic magma just before the 1914 eruption. Although the Mg–Fe diffusion time was not calculated for the 1471 and 1779 eruptions, the same probably holds true for the two eruptions because the orthopyroxene and magnetite in the 1471 and 1779 Plinian pumices have characteristics similar to those of the 1914 Plinian pumice. Thus, the mafic magma recharge and subsequent mixing may not have been the immediate causes of the historical Plinian eruptions. It has been shown that the Plinian magmas were emplaced in a shallow conduit (1–3 in depth) prior to the eruptions (Araya et al., 2019, Scientific Reports). Crystallization-induced vesiculation within the conduit could have finally triggered the historical Plinian eruptions.