Re-analysis of 1986 eruption of Izu-Ohshima by tremor source location

One of the difficult problems in the predictions of volcanic eruption is that eruption evolves with time,so that adaptive grasp of the phenomena is needed. This is essentially different from the situations in the earthquake prediction. During the eruption of 1986 Izu-Ohshima eruption style changed from strombolian to vulcanian-subplinian in association with migration of the vent from the central crater to NW-SE trending fissures. This unexpected shift of the eruption caused strong fears among the people and all the residents in the island were forced to leave the island for almost one month. Almost 30 years has passed since the last eruption and Izu-Ohshima is expected to be approaching the next eruption. Motivated by this bitter lesson to prospect future eruption we reanalysed seismic records of 1986 eruption to characterize the tremor activity. Volcanic tremors have been recognized as a good indicator for the adaptive monitoring of eruption and new methods in the analysis have been proposed since the time of eruption. The first stage of the eruption is characterized by strombolian type central vent eruption, where continuous excitation of the tremor was observed. The tremor sources are determined by the Amplitude Inversion Method(AIM) proposed by Battaglia and Aki 2003 and they mostly concentrate into narrow region beneath the central vent. This activity started from 15/Nov. to 19/Nov. The amplitude of the tremor monotonously increased towards the end while the magma effusion rate monotonously declined. After the cessation of the eruption for 40 hours the fissure eruption started to continue for 2~3 days. In this second stage sporadic wave-train type tremors (duration of several minutes) were observed. The locations are along the line of NW-SE trend, which overlaps with the surface fissures although extending in much wider region than the fissure. We carefully separated jaggy parts in the continuous tremor of the first stage and we found the locations are also along the similar trend. This indicates magma supply to the fissure preceded at least by several days before the eruption. Since this kind of intriguing clustering is not observed in the seismicity we suggest an important role of volcanic tremor in the adaptive monitoring of eruption.