High-resolution reconstruction of the Hoei eruption (AD 1707) of Fuji volcano, Japan

Abstract

The Hoei eruption of Mount Fuji in 1707 caused the worst ashfall disaster in Japanese history. Despite the availability of numerous historical documents describing the eruption, the detailed eruption sequence has not been verified because the correlation between these descriptions and geological sequences remains unclear. In this study, we reconstruct the sequential change in the column height using newly established stratigraphy and a detailed timeline obtained from historical documents. The eruptive deposit was subdivided into 17 units on the basis of their facies, with the mass of each unit established using isopach maps. The eruption column height is a function of the magma discharge rate; hence, the column height of each unit was estimated from its erupted mass and duration, which were inferred from the historical documents. The correlation of each unit with the actual time was based on the tephra color and grain size, an unconformity caused by rainfall, and ashfall distribution. While reconstructing the unit boundaries, we found that not all of them represented a hiatus or relenting phase of ashfall. We detected only six obvious quiet intervals from the historical documents. Therefore, many of the unit boundaries may represent a hiatus or relenting phase that was too subtle to have been recorded in the historical documents. We define an eruptive pulse as a period of continuous ashfall followed by an obvious quiet interval. We divided the 17 units of the Hoei eruption into 6 pulses, and into 3 stages on the basis of the patterns of the eruptive pulses. The characteristics of the three stages are described as follows. Stage 1 had two energetic eruptive pulses (pulses 1 and 2; at least 20 km high column), each showing an intense initial outburst, followed by a decrease in intensity. The eruption sequence indicates ruptures of highly overpressured dacite and andesite magma chambers. The initial silicic eruption was followed by basaltic magma withdrawal from a deep and voluminous magma chamber. Stage II consisted of discrete subplinian pulses of relatively degassed basaltic magma. Although the eruption rate throughout the stage decreased, the magma supply from depth appears to have been sustained because extensive intrusion near the surface created the Mt. Hoei cryptodome near the vent. Stage III was principally characterized by sustained column activity without a clear repose time. During this stage, the column height appears to have been more than 13 km, and we recognized at least two distinct periods of increased activity in which the column height is presumed to have exceeded 16 km. The Cu-rich vesicular scoria and continuous eruption in stage III indicate a stable supply of volatile-rich magma from depth. No significant decay was observed in the magma discharge rate; hence, the eruption could have been halted by a sudden process such as conduit collapse, rather than decompression of the magma chamber.