Formation of a Compositionally Reverse Zoned Magma Chamber: Petrology of the ad 1640 and 1694 Eruptions of Hokkaido-Komagatake Volcano, Japan

Historical eruptive activity commenced at Hokkaido-Komagatake volcano in ad 1640 after about 5000 years of dormancy. The ad 1640 event was followed by another eruption in ad 1694. Petrological investigation of these eruptions provides useful information for understanding the initial state of the magma system and its early evolution. Both events were plinian eruptions in which the ejecta were primarily white pumice (silicic andesite) with lesser amounts of mafic clasts (mafic to intermediate andesite). In both cases the mafic material erupted prior to the silicic. The ad 1640 ejecta are of three types: white pumice, scoria and banded pumice. The whole-rock chemistry of the banded pumice shows a linear chemical trend connecting the white pumice and scoria, which is interpreted to indicate that magma mingling occurred between two end-members. Highly porphyritic white pumice shows no evidence of magma mixing, indicating that this material is directly representative of the silicic end-member magma. Scoria, which is interpreted to represent the mafic end-member magma, is nearly aphyric but the phenocrysts that are present are similar to those contained in the white pumice. In the ad 1640 eruption, the injected mafic magma is inferred to have ascended as a dike into a silicic chamber while mingling; the mafic material was erupted first. In the ad 1694 eruption, gray pumice occurs in addition to the other three types of ejecta and this has whole-rock compositions intermediate between those of the white pumice and scoria. In addition, the gray pumice contains phenocrysts with reverse zonation, suggesting that it is a product of mixing between the end-member magmas. These pumices suggest that a zoned magma body was formed during the ad 1640 eruption and that the two contrasting magmas mixed until the ad 1694 eruption. The chamber was compositionally reverse zoned and consisted of upper mafic and lower silicic zones with hybrid layers between them; the silicic magma was strongly porphyritic and thus denser than the aphyric mafic magma. The sequence of the ad 1694 eruption can be explained by sequential ejection from the upper part of the zoned chamber without a new injection of mafic magma. This is consistent with the view that the ad 1694 banded pumice was the product of mingling between the silicic and hybrid magmas and that mafic magma had not interacted with the silicic magma.