Recent Progress of Geophysical and Geological Studies of Mt. Fuji

Earth-Science Reviews 194 (2019) 264–282

Contents lists available at ScienceDirect

Earth-Science Reviews

journal homepage: www.elsevier.com/locate/earscirev

Recent progress of geophysical and geological studies of Mt. Fuji Volcano, Japan T ⁎ Yosuke Aokia, , Kae Tsunematsub,c, Mitsuhiro Yoshimotob a Earthquake Research Institute, The University of Tokyo, Tokyo, Japan b Mount Fuji Research Institute, Yamanashi Prefectural Government, Fujiyoshida, Yamanashi, Japan c Now at Department of Earth and Environmental Sciences, Yamagata University, Yamagata, Japan

ARTICLE INFO ABSTRACT

Keywords: Mt. Fuji, located in central Japan near the triple junction of the Philippine Sea, Eurasia (or Amurian), and North Mt. Fuji American (or Okhotsk) plates, is one of the arc volcanoes associated with the subduction of the Pacific plate. Mt. Volcanic eruption Fuji has unique features including an average eruption rate in the last 100,000 years of 4–6km3/ka, which is Magma plumbing system much higher than that of other volcanoes along the same arc (0.01 and 0.1 km3/ka). Basaltic rocks dominate in Volcano monitoring Mt. Fuji, while other arc volcanoes are dominated by intermediate and felsic compositions, and Mt. Fuji has Forecasting an eruption erupted both explosively and effusively, with the two largest eruptions in the last 2000 years having different styles; the 864–866 CE Jogan eruption was effusive, while the 1707 Hoei eruption, the most recent eruption, was explosive. Mt. Fuji is not only interesting from a scientific point of view, but also important from a societal point of view because it is only 100 km from the Tokyo metropolitan region, one of the biggest cities in the world hosting more than 30 million people. The Hoei eruption resulted in more than a few tens of millimeters of ashfall in Tokyo by more than 150 mm. With this background, Mt. Fuji has been studied intensively using geological and geophysical methods to understand the evolution, magma plumbing system, and current activity of the volcano. This article synthesizes the current knowledge about Mt. Fuji. In particular, we provide background information required to understand the Jogan and Hoei eruptions, and discuss how to assess the size, timing, and style of the next eruption. Statistical insights suggest that the next eruption of Mt. Fuji is more likely to be voluminous with a long precursor, given a repose time of more than 300 years.

1. Introduction tectonic setting of Mt. Fuji (Fig. 1) might play a role, but the prevalence of basaltic magmas has not been explained. Mt. Fuji is one of the arc volcanoes associated with the subduction Third, Mt. Fuji has erupted in various styles from effusive to ex- of the Pacific plate (Fig. 1). Mt. Fuji (3776 m) stands out from other plosive eruptions. While the Jogan eruption was mainly effusive, the volcanoes in Japan by at least five aspects: volume, composition, 1707 Hoei eruption, another one of the largest eruptions in the last eruption style, vent location, and hazard. It has a large volume of about 2000 years, was an explosive eruption which ejected a total of about 400–500 km3 despite being active only in the last 1.6 km3 of basaltic and dacitic materials (Miyaji et al., 2011). 80,000–100,000 years, corresponding to the average ejection rate of Fourth, Mt. Fuji has erupted both from the summit and the flank. 4–6km3/ka. This ejection rate is significantly larger than that of other Since the ambient stress field around Mt. Fuji is dominated by north- arc volcanoes along the same arc, between 0.01 and 0.1 km3 (Tsukui west-southeast compression due to the collision with the Philippine Sea et al., 1986; Yoshimoto et al., 2010). plate (e.g., Ukawa, 1991; Araragi et al., 2015), the vents are mainly Second, the chemical composition of Mt. Fuji is distinct from other distributed on the northwestern and southeastern flanks. In fact, all the arc volcanoes. Intermediate and felsic magmas such as andesite, dacite, eruptions of Mt. Fuji in the last 2200 years, including the Jogan and or rhyolite dominate in arc volcanoes, whereas Mt. Fuji has mainly Hoei eruptions, are from the flank rather than the summit. This eruptive ejected basaltic rocks. For example, the 864–866 CE Jogan eruption, history indicates that large areas are under potential threat from lava one of the two largest eruptions in the last 2000 years, ejected a total of flows. about ~1.4 km3 of basaltic materials (Miyaji et al., 2006). The complex Finally, Mt. Fuji is only about 100 km from the Tokyo metropolitan

⁎ Corresponding author. E-mail address: [email protected] (Y. Aoki). https://doi.org/10.1016/j.earscirev.2019.05.003 Received 2 April 2018; Received in revised form 11 April 2019; Accepted 2 May 2019 Available online 11 May 2019 0012-8252/ © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/). Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 1. Tectonic setting around Mt. Fuji. PAC, PHS, EUR, AMR, NA, and OKH denote Pacific, Philippine Sea, Eurasian, Amurian, North American, and Okhotsk plates, respectively. Red triangles represent active volcanoes, in which Mt. Fuji is represented by the biggest triangle. Gray dash lines represent depth contours of the subducting Philippine Sea plate with an interval of 10 km given by Hirose et al. (2008) and Nakajima et al. (2009). Rupture areas of the 1854 Ansei and 1923 Kanto earthquakes are also shown. Yellow rectangles denote the areas shown as close-up views in Fig. 2. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) area with a population of more than 30 million. The area is thus po- rates of about 25 and 15 mm/yr, respectively (e.g., Argus et al., 2010). tentially at risk from volcanic hazards. Indeed, the most recent Hoei The Philippine Sea plate collides at the northern tip near Mt. Fuji eruption was so explosive that the Tokyo metropolitan region received (Fig. 1). While the Philippine Sea plate is an oceanic plate, the colliding up to 100 mm of ash (Miyaji et al., 2011). The same or similar eruption section in it is part of an island arc with lower average density than now would cause severe volcanic hazards in Tokyo and the surrounding oceanic plates. The island arc on the Philippine Sea plate thus cannot areas. The Hoei eruption started 49 days after a nearby M ∼ 8.5 subduct as smoothly as oceanic plates, resulting in a collision at the earthquake, indicating that the region was affected by volcanic hazards plate boundary. less than two months after the seismic activity. Seismicity at the subducting and colliding plate boundaries around Therefore, understanding the potential location, style, size, and the northern tip of the Philippine Sea plate exhibit a marked difference timing of the next eruption of Mt. Fuji is important not only from a (Ishida, 1992). The subducting plate boundary exhibits abundant seis- scientific standpoint but also for the hazard assessment. This article micity including large earthquakes of M∼8 such as the 1854 Ansei and reviews insights into understanding how Mt. Fuji was formed and po- 1923 Kanto earthquakes in the Suruga and Sagami troughs, respec- tentially assesses its next eruption of Mt. Fuji. Section 2 summarizes the tively. On the other hand, the colliding part of the plate boundary lacks tectonic setting and stress field around Mt. Fuji. We have reviewed the major seismicity. Although seismic reflectors were detected to the north evolution of Mt. Fuji during the last 400,000 years in Sections 3. of Mt. Fuji (Iidaka et al., 1990), it lacks widespread seismic reflectors Sections 4 and 5 describe insights into the internal structure and current associated with plate subduction. The subducting plate boundary hosts activity of Mt. Fuji. Based on the discussion between Sections 2 and 5, obvious seismic reflectors corresponding to the Moho of the subducting Sections 6 and 7 summarize how we can address the research questions Philippine Sea plate (Arai and Iwasaki, 2015). posed in this section and assess the next eruption. 2.2. Regional and local stress field 2. Tectonic background The plate configuration described above results in a regional stress 2.1. Plate motions around Mt. Fuji field with northwest-southeast compression and northeast-southwest extension around Mt. Fuji (Ukawa, 1991). Mt. Fuji and Izu-Oshima Mt. Fuji is located near the triple junction of the Philippine Sea, volcanoes (Fig. 1) are elongated in a northwest-southeast direction (Ida, Eurasian (or Amurian), and North American (or Okhotsk) plates, under 2009; Fig. 2), with dominant strikes of volcanic fissures (Takada et al., which the Pacific plate subducts at a convergence rate of ~85 mm/yr 2007), and focal mechanisms of shallow earthquakes (Ukawa, 1991; (Argus et al., 2010; Fig. 1). From the south, the Philippine Sea plate Araragi et al., 2015) confirming the stress field. The shape of volcanic subducts beneath the North American (or Okhotsk) and Eurasian (or islands becomes more rounded (Fig. 2), and the orientation of volcanic Amurian) plates from the Suruga and Sagami troughs with convergence fissures becomes radial with distance from the plate boundary (Ida,

265 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 2. Close-up views of (a) Mt. Fuji, (b) Izu-Oshima, and (c) Miyakejima volcanoes, whose locations are shown in Fig. 1. The shape of Mt. Fuji and Izu-Oshima volcanoes are more elliptical than Miyakejima because the ambient stress fields around them are more deviatoric than the stress field around Miyakejima.

2009). These observations confirm the collision and subduction of the anisotropy is likely the differential stress. Their study suggests that the Philippine Sea plate control the regional stress field. stress field around Mt. Fuji at shallow (≤4 km) depths is, at least, a The local stress field on the edifice of Mt. Fuji is not only controlled combination of the regional stress field and the stress from volcano by the collision and subduction of the Philippine Sea plate but also by loading. the stress field exerted by the loading of the volcano itself. Araragi et al. (2015) investigated seismic anisotropy around Mt. Fuji at depths shal- lower than ∼4 km, and found that while the fast direction of the 3. Evolution of Mt. Fuji and recent eruptions seismic wavespeed is radially distributed near the summit of Mt. Fuji, it roughly strikes northwest-southeast (Fig. 3). The differential stress, The evolution and geology of Mt. Fuji have been extensively in- orientation of cracks, or lattice preferred orientation of rocks generate vestigated since pioneering studies by Prof. Hiromichi Tsuya in the anisotropy of seismic velocity. In this case, the origin of seismic 1930s (e.g., Tsuya, 1955, 1962), leading to the most recent version of the geological map (Takada et al., 2016). Thus, we do not attempt to

Fig. 3. Rose diagrams of fast polarization directions obtained from a shear-wave splitting analysis for each seismic site. Gray lines indicate directions of the maximum compression at a depth of 2.0 km that fits with the obtained fast polarization directions; they are obtained by a combination of the point load at the summit of Mt. Fuji and an appropriate ambient stress field (see Araragi et al., 2015, for details). Modified from Araragi et al. (2015).

266 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 4. A schematic north-south cross section through Mt. Fuji. After Yoshimoto et al. (2010).

volcano to the south of the current Mt. Fuji started to grow about 400 ka ago. Pre-Komitake volcano, located to the north of current Mt. Fuji, started to grow with effusions of basaltic lava flows at around 270 ka, ending up with explosive eruptions of basaltic andesite and dacite magmas at around 160 ka. Komitake volcano grew, by effusions of basaltic lavas, on top of Pre-Komitake volcano between around 160 ka and 100 ka, when the youngest part of Mt. Fuji started to grow. Yoshimoto et al. (2010) proposed that a change in regional tectonics around 150 ka ago might have been responsible for a change in the rock chemistry from Pre-Komitake volcano to subsequent volcanoes, but this conjecture has not been proved yet.

3.2. Evolution after 100 ka

Basaltic volcanism dominates in the current Fuji volcano, and its activity can be divided into two periods, the older Fuji between 100 ka and 10 ka and the younger Fuji since 10 ka. Eruptions of the older Fuji may have been explosive considering that extensive deposits from old Fuji eruptions cover an area of ∼250 km2 mainly to the east of Mt. Fuji (Miyaji et al., 1992; Yoshimoto et al., 2010). Basaltic lava flows dom- inate the eruptions of younger Fuji, but explosive eruptions of dacitic and andesitic magmas such as the Zunasawa eruption ∼3 ka ago and the 1707 Hoei eruptions also exist. After a dominance of sub-Plinian explosive eruptions of basaltic magma, widespread scoria falls, and lahar deposits between 100 and 17 ka (Uesugi, 1990; Yamamoto et al., 2007) made volcanic fans on the lowlands of the volcano. Voluminous lava flows dominated the volcanic activity between 17 ka and 8 ka, but the insufficient number of measurements does not permit an evaluation of the ejected volume during this period. Explosive eruptions and lava flows from the summit and flank until 1500 BCE (stage 3: see Fig. 5) followed a relatively mild activity be- tween 10,000 BCE, and 3500 BCE. It is during this period that the main growth of the volcano occurred due to voluminous ejections of magma; the size of the volcano was already as large as that in the present by 3500 BCE. The current summit of Mt. Fuji is composed of rocks ejected during this period. Mt. Fuji erupted explosively between 1500 BCE and 1300 BCE from the summit with ejections of scoria and pyroclastic flows (e.g., Fig. 5. (top) Cumulative volume of eruptive products of Mt. Fuji over the last Yamamoto et al., 2005a). One of the most notable ejecta is the Zuna- 11,000 years. (bottom) Cumulative volume of eruptive products over the last sawa deposit, ejected around 1300 BCE from the eastern flank, where it 2200 years. After Miyaji (2007). is distributed (e.g., Machida, 1977; Yamamoto et al., 2005b). The Zu-

nasawa deposit contains rocks with SiO2 content of up to 70%, similar give a comprehensive review of the evolution and previous eruptions of to the scoria ejected by the Hoei eruption. fl Mt. Fuji. Instead, here we brie y summarize the eruptions in chron- The debris deposit associated with a flank collapse in the eastern ological order, focusing on the recent Jogan and Hoei eruptions, which flank around 900 BCE cover 53 km2 with a volume of 1.05 km3 (Miyaji represent diverse eruptive styles of Mt. Fuji. et al., 2004; Miyaji, 2007). Whether a phreatic eruption or the shaking from an earthquake triggered the collapse is unknown. 3.1. Evolution before 100 ka The flank collapse 2.9 ka ago was followed by a series of summit eruptions, one of the most significant of which occurred 2.2 ka ago, a Scientific drilling revealed that the evolution of Mt. Fuji is divided sub-Plinian eruption with an eruptive volume of ∼0.5 km3 (Suzuki and into several stages (Yoshimoto et al., 2010; Fig. 4). First, Ashitake Fujii, 2010). The eruption 2.2 ka ago was the latest summit eruption of

267 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 6. Isopach of the tephra ejected from the 1707 Hoei eruption. a) Location of Fuji Volcano. b) Reginal isopach of the total Hoei tephra. c) Local isopach. Solid dots and stars indicate outcrops and locations where historical documents are available. Thickness is shown in centimeters. Squares indicate cities and villages; N, Narita; S, Sawara; E, Edo or present Tokyo; Y, Yokohama, H; Hiratsuka; Od, Odawara; O, Oyama; G, Gotemba; Sb, Subashiri; Sn, Susono; Sy, Suyama. After Miyaji et al. (2011).

Mt. Fuji. 3.2.2. Eruptions between the Jogan and Hoei eruptions After this summit eruption in 2.2 ka, small-scale flank eruptions The 937 eruption was originated from three fissures aligned and occurred until the 8 th century, after which Mt. Fuji erupted at least ten located on the northern flank between 1800 and 2900 m above sea times in 781, 800–802, 864–866, 937, 999, 1033, 1083, 1435–1436, level. Some lavas reached up to 17 km to the northeast from the vent 1511, and 1707 (Fig. 5). Dating of these eruptions is difficult, and thus (Fujita et al., 2002). Yamamoto et al. (2005b) suggested that the 937 historical accounts and mapping of eruptive products are the only ways eruption was from the southern flank as well as the northern flank, for constraining the exact dates of these eruptions. although the volume of the eruptive product from the southern flank is The 800–802 eruption ejected massive tephras from the north- estimated to be only 0.01 km3 compared with 0.06 km3 from the western and northeastern flanks. Historical accounts suggest that Mt. northern flank. Fuji might have erupted in 834–848, 853, 858, and 859 as well The 1033 eruption also originated from the northern and southern (Koyama, 1998a) but a more careful examination of the documents is flanks (Yamamoto et al., 2005b). On the northern flank, a 3 km-long necessary to confirm the dates. fissure between 2150 and 3450 m above sea level and a vent at 2050 m above sea level were responsible for the eruption. Lavas flowed northeastward for up to 7 km from the vent (Fujita et al., 2002). 3.2.1. The 864–866 Jogan eruption Details of the 1083 eruption are not well known, but Koyama The 864–866 Jogan eruption is one of the largest eruptions in last (1998b) suggested that the eruption was explosive based on a historical 2200 years with more than 1.4 km3 of lavas from vents aligned on the account that the eruption was heard in Kyoto, ∼270 km from Mt. Fuji. northwestern flank with a horizontal extent of ∼6 km. The lavas buried Details of the 1435–1436 and 1511 eruptions are not also well known most of a lake that existed at that time. The interaction between the but historical accounts suggest that lava effusion might have dominated lava and lake water triggered a secondary phreatic eruption, generating the 1435–1436 eruption (Koyama, 1998b). The 1511 eruption might pillow lavas (Miyaji et al., 2006).

268 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 7. (a) Distribution of hypocenters around Mt. Fuji between 1 January 1995 and 31 December 2018 with depths shallower than 30 km and M = 1.0. Focal mechanisms are also shown. Focal mechanisms in red correspond to earthquakes with significant number aftershocks in 2011 and 2012. Large and small green triangles indicate the location of the summit of Mt. Fuji and Hakone Volcano, respec- tively. (b) The cumulative number of earthquakes of M≥ 1 between 1 January 1995 and 31 December 2018. The earthquake rate is stationary except when aftershocks of the East Shizuoka earthquake on 15 March 2011 and a M = 5.4 earthquake on 27 January 2012 occurred. A significant increase of seismicity in 2015 is due to elevated volcanic activity in Hakone Volcano (e.g., Mannen et al., 2018). (For interpretation of the references to colour in this figure legend, the reader is referred to the web ver- sion of this article.)

have originated from the northern flank (Koyama, 1998b). column heights of more than 16 km. A sudden process such as conduit collapse, rather than a gradual process such as decompression of the magma chamber, halted the eruption. 3.2.3. The 1707 Hoei eruption The Hoei eruption accumulated more than a meter of ash within a The 1707 Hoei eruption, the most recent one, is one of the most few kilometers from the vent, and more than 100 mm of ash accumu- explosive eruptions in Mt. Fuji in historical time. The eruption started lated 100 km away from the vent in the downwind direction (Fig. 6). on 16 December 1707, 49 days after the 1707 Hoei earthquake The Japanese government predicts an economic loss of ∼2.5 trillion (M = 8.7) occurred in the Nankai trough, southwest Japan. This tem- yen (or 25 billion US dollars) if a Hoei-type eruption would occur red poral coincidence and static stress changes in Mt. Fuji by the Hoei today (Aramaki, 2007). earthquake led Chesley et al. (2012) to suggest that the extensional stress changes of about 0.1 MPa by the Hoei earthquake might have triggered the eruption. The Hoei eruption lasted for 16 days until 1 3.2.4. Modern activity January 1708. While Mt. Fuji has not erupted since 1707, deep low-frequency The eruption has been reconstructed in detail by Miyaji et al. earthquakes have been persisting at least since the first seismic ob- (2011). The total ejected volume is estimated to be 1.6 km3 with a servations in the 1980s (Kanjo et al., 1984; Ukawa, 2005). The number Volcano Explosivity Index (Newhall and Self, 1982) of 5. The eruption of deep low-frequency earthquakes increased in 2000–2001 but this started by two energetic eruptive pulses with an ash column higher than unrest did not result in surficial expressions (Ukawa, 2005). The East

20 km. The energetic pulses gave way to discrete sub-Plinian basaltic Shizuoka earthquake (Mw 6.0) four days after the 2011 Tohoku-oki eruptions before sustained activity with column heights of more than earthquake was located right beneath Mt. Fuji but did not trigger the 13 km, during which there were at least two heightened phases with volcanic activity of Mt. Fuji. Also, the earthquake did not trigger any

269 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

on 27 January 2012 occurred and when the Hakone Volcano was active in 2015 (e.g., Mannen et al., 2018).

4. Geophysical view of the dmagma plumbing system

The magma plumbing system of an active volcano can be inferred from both petrological and geophysical insights. Petrology can provide the information about the depth of magma reservoirs as well as the decompression rate, but not about the horizontal extent of the magma reservoir. On the other hand, seismic and electromagnetic imaging can detect magma reservoirs as bodies of low seismic velocity and low re- sistivity, but the images provide no information on how the magma migrated toward the surface.

4.1. Petrological and geochemical insights

The FeO*/MgO ratio of the volcanic products from Mt. Fuji is stable over time, leading Togashi and Takahashi (2007) to suggest that the magma reservoir of Mt. Fuji has been receiving magma more or less constantly over the last ∼100,000 years. Considering a high eruption rate during that time (Tsukui et al., 1986), the rate of magma supply to the magma reservoir is high. Togashi and Takahashi (2007) suggested that this magma supply does not allow magma to reside in the magma Fig. 8. A schematic model of the magma plumbing system of Mt. Fuji inferred ff from melt inclusions. The magma plumbing system consists of relatively deep reservoir long enough to be di erentiated, resulting in a dominance of basaltic rocks. (∼20 km) basaltic and shallow SiO2-rich magma reservoirs (∼8–9 km). After Kaneko et al. (2010). An investigation of volcanic products by Yoshimoto et al. (2004) and a melt inclusion study by Kaneko et al. (2010) of the Hoei eruption deposits suggests the existence of two magma reservoirs beneath Mt. Depth Fuji Izu arc Fuji, the deeper one being at ∼20 km and the shallower one at 8–9km EU below sea level (Fig. 8). Kaneko et al. (2010) proposed that magmas in PHS the deeper and shallower reservoir would be basaltic and felsic, re- 10 Granitic rock spectively. An accumulation of magma in the conduit left behind by previous eruptions forms the shallower magma reservoir. Differentia- tion over time makes the accumulated magma more felsic. This model Gabbro with two magma reservoirs can explain why felsic eruptions sometimes 20 occur in a volcano dominated by basaltic rocks like Mt. Fuji, and the chronology of the 1707 Hoei eruption which started by silicic eruptions followed by basaltic eruptions (Miyaji et al., 2011). The deeper magma 30 reservoir is deeper than that of other volcanoes along the Izu-Bonin arc such as Izu-Oshima whose magma reservoir is at around 10 km (e.g., Mikada et al., 1997). Fujii (2007) suggested that this deep magma re- 40 Peridotite servoir is produced because of the underplating of the Philippine Sea plate, making the crust thicker. (Fig. 9). Togashi and Takahashi (2007) also suggested the presence of a deep magma reservoir to explain the Hoei explosive eruption. They suggested that, during the Hoei eruption, 50 (km) the magma rose up from the magma reservoir quickly enough not to allow time to degas. The Togashi and Takahashi (2007) model is Fig. 9. A schematic cross-section beneath Mt. Fuji and a volcano along the Izu somewhat di fferent from Kaneko et al. (2010) because it does not re- arc, indicating that the location of granitic mid-crust controls the depth of the quire the presence of a shallow magma reservoir. magma reservoir, shown in gray. The granitic mid-crust beneath Mt.Fuji is Shibata et al. (2015) suggested from Sr and Nd isotope ratio that the deeper than the Izu arc because the crustal materials of the Izu arc underplate magmas of Pre-Komitake (≥270–160 ka; Fig. 4) share a similar mantle the Eurasian plate, denoted as EU in the figure, through subduction of the source with those of subsequent volcanoes, Komitake and Fuji (Fig. 4). Philippine Sea plate, denoted as PHS in the figure. Modified from Fujii (2007). They also argued, using the major element content data, that the Pre- Komitake magmas could have experienced fractionation at a pressure of deformation or increase in seismic activity around Mt. Fuji. (Fujita ≥1.2 GPa or ≥35 km, greater than that of subsequent magmas, ∼500 et al., 2013). Fig. 7(a) denotes the distribution of hypocenters of MPa or ∼15 km. It may imply the existence of a fossil magma reservoir earthquakes shallower than 30 km between 1995 and 2018 with focal that has been active during the Pre-Komitake era at ∼35 km beneath mechanisms. Seismic activity is not spatially uniform but concentrated Mt. Fuji. to the east of Mt. Fuji. Aftershocks of the East Shizuoka earthquake and Migration of magma during the eruption 2.2 ka ago, the latest seismicity related to volcanic activity in Hakone Volcano, shown by a summit eruption of Mt. Fuji, has been reconstructed from groundmass small green triangle in Fig. 7(a), are also seen. Focal mechanisms are microlite textures by Suzuki and Fujii (2010) who inferred that the rather diverse, but they are dominated by thrust and strike-slip faulting eruption was preceded by a magma decompression at a rate of with compressional axes striking roughly northwest-southeast. Fig. 7(b) 0.005–0.012 MPa/s (Stage 1 of Fig. 10), which can be roughly trans- depicts the cumulative number of earthquakes exceeding M = 1.0. It lated as a magma ascent rate of 0.17–0.40 m/s when the density of the demonstrates that the seismicity is temporally stationary except when host rock is assumed to be 3000 kg/m3. The earliest stage of the erup- aftershocks of the East Shizuoka earthquake and a M = 5.4 earthquake tion involves widening of the vent, conduit, or both with an increasing

270 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 10. A schematic view of the evolution of the eruption 2.2 ka ago, the last summit eruption of Mt. Fuji as reconstructed from groundmass microlite textures. The earlier and later stages are shown in the left and right panels, respectively. After Suzuki and Fujii (2010). rate of magma decompression (Stage 2 in Fig. 10). The decompression Intrinsic attenuation in the crust beneath Mt. Fuji, however, is not rate further increased up to 0.13–0.22 MPa/s (Stage 3 in Fig. 10) with significantly different from that in surrounding regions (Chung et al., an increased eruption rate and decreasing outgassing rate. Variable 2009), indicating that partial melt beneath Mt. Fuji is not as abundant decompression rates of magma characterize the later stage (Stages 4 as that in hot spot volcanoes. and 5 in Fig. 10) because of a variable flow rate of magma between the Kinoshita et al. (2015) used receiver functions to image the seismic central and marginal parts of the conduit. The eruption 2.2ka ago is also structure down to about 80 km. They found a low- velocity zone at characterized by a decrease of overpressure at the reservoir after the 13–26 km below Mt. Fuji, the bottom of which is characterized by a major eruption in Stage 3. In the last stage, the magma velocity has a sharp velocity contrast (Fig. 12). The low velocity likely marks a spatial gradient, with a low velocity at the margin and high at the magma reservoir beneath Mt. Fuji. They also found that a velocity center of the conduit. An overall increase of outgassing from the con- contrast at 40–50 km associated with underplating of basaltic materials, duit margin is inferred in this stage (Stage 5 in Fig. 10). which volcanic arcs ubiquitously host (e.g., Takahashi et al., 2009), disappears beneath Mt. Fuji. It might reflect a pathway of magma from the subducting Pacific plate. Also, a geochemical insight by Nakamura 4.2. Seismic imaging et al. (2008) suggested that volcanic fluids in Mt. Fuji are dominantly fed from the Pacific plate instead of the overriding Philippine Sea plate. Seismic imaging has contributed to the delineation of the magma These studies indicate that the magma of Mt. Fuji is fed from the Pacific plumbing system of Mt. Fuji. Lees and Ukawa (1992) found an extended plate through the region marked by a lack of velocity contrast at depths region of low seismic velocities at depths of 20 km or deeper beneath of 40–50 km that Kinoshita et al. (2015) found. Mt. Fuji. Since the depth of low seismic velocities is consistent with the A seismic exploration experiment with five active sources and 469 depths of the magma reservoir inferred from petrology as described temporary seismometers was conducted in 2003 (Oikawa et al., 2006; above, they were interpreted as evidence of a magma reservoir beneath Tsutsui et al., 2007) to infer the seismic structure beneath Mt. Fuji in Mt. Fuji. However, the spatial resolution of their imagery is not good the upper crust. The active sources and seismometers were linearly fi enough to con rm the interpretation. Recent studies (e.g., Kamiya and aligned to delineate a two-dimensional seismic velocity structure across Kobayashi, 2007; Matsubara et al., 2008; Nakajima et al., 2009) also Mt. Fuji. Tsutsui et al. (2007) found a reflector dipping to the southwest found similar low-velocity anomalies in their datasets, but the spatial at depths of around 15 km to the northeast and 25 km to the southwest resolution is still not good enough to delineate the magma plumbing (Fig. 13). This reflector corresponds to the top of subducting Philippine system in detail. Sea plate. They also identified a series of short and horizontal reflectors Nakamichi et al. (2007) applied local seismic tomography to image with lengths of a few kilometers at depths 8–15 km (Fig. 13). One of the a seismic structure around Mt. Fuji from a densely deployed temporary reflectors is located near the shallowest part of low-frequency earth- seismic network (Fig. 11). They found that low P-wave velocities and quakes (Nakamichi et al., 2004), so these reflectors probably reflect low Vp/Vs ratios characterize the area around hypocenters of low-fre- velocity discontinuities due to the presence of volcanic materials. quency earthquakes, interpreting the low-velocity zone as indicative of the presence of volatiles such as H2OorCO2. However, they failed to image 20 km or deeper, where the magma reservoir is likely to exist. 4.3. Electromagnetic imaging Scattering attenuation investigated by Chung et al. (2009) found that the crust beneath Mt. Fuji is more inhomogeneous than that in The seismic structure discussed so far can be compared with re- surrounding regions; the mean free path of seismic waves around Mt. sistivity structure. Aizawa et al. (2004) derived the resistivity structure Fuji is about 1 km at 2–8 Hz, a typical value seen in active volcanoes. beneath Mt. Fuji down to 50 km from a campaign-mode

271 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 11. (a) Distribution of epicenters (gray dots) and grid locations (open red squares) of the inversion by Nakamichi et al. (2007). Spacing of the open red squares serves as a proxy of the spatial resolution of the seismic tomography. Blue lines A–A' and B–B′ show the cross sections in (b)–(g). (b) P-wave velocities, (c) S-wave velocities, and (d) Vp/Vs ratio along the profile A–A′ in (a). The red arrow in (b) indicates the location of Mt. Fuji. The contour intervals for P and S wave velocities are 0.5 km/s and 0.25 km/s, respectively. Black and red dots represent hypocenters of tectonic and deep low-frequency earthquakes, respectively. Note that velocities are given only in areas where the resolution is sufficiently high. (e)(f)(g) Similar to (b)(c)(d) but along the B–B′ profile in (a). The thick white line shows the location of a reflector inferred by Tsumura et al. (1993). Modified from Nakamichi et al. (2007). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

electromagnetic survey. A prominent feature derived from their study is hypocenter of the East Shizuoka earthquake (Mw6.0) on 15 March 2011, a high resistivity body corresponding to the subduction of the Phi- four days after the 2011 Tohoku-oki earthquake. Aizawa et al. (2016) lippine Sea plate (R1 in Fig. 14) which is ubiquitous in subduction combined this resistivity structure and isotopic properties of ground- zones. Another prominent feature is a low resistivity body below depths water in the area to interpret that the Mw 6.0 earthquake was triggered where low-frequency earthquakes occur. Although the spatial resolu- within a fracture zone through which magmatic gases penetrate. tion is not good enough, the low resistivity they found might corre- Another electromagnetic survey revealed that there is an area re- spond to the magma reservoir or magma pathway discussed by presented by low resistivity at the shallowest level, 500–2000 m above Kinoshita et al. (2015). A more sophisticated analysis by Aizawa et al. sea level (Fig. 16; Aizawa et al., 2005). This low resistivity is inter- (2016) shows that the low resistivity region extends up to a depth of preted as a confined hydrothermal system with a spatial extent of a few about 7 km with a slight offset from the summit to the southwest kilometers rather than spreading over the whole edifice. Aizawa et al. (Fig. 15). This low resistivity body roughly corresponds to the (2005) suggested that the confined hydrothermal system resulted from

272 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 12. Cross sections of receiver function amplitude with a cut off frequency of 1.0 Hz along eleven lines in the upper left panel. Lines A, B, C, and D are denoted in blue and lines a, b, c, d, e, f, and g are denoted in red, respectively. Red indicates a velocity boundary, the bottom of which has higher velocity, and blue indicates vice versa. Gray dots and white circles in each panel are hypocenters with a magnitude larger than 0.1 between 2002 and 2005. White circles represent low-frequency earthquakes. Pink circles represent hypocenters of low-frequency earthquakes relocated by Nakamichi et al. (2007). Green triangles show the location of the summit of Mt. Fuji. Black solid lines denote velocity boundaries from receiver functions inferred by Kinoshita et al. (2015). White arrows indicate the gap in the velocity boundary inferred by Kinoshita et al. (2015). Modified from Kinoshita et al. (2015). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) the limited circulation of hydrothermal fluids beneath Mt. Fuji, con- earthquakes between 1980 and 2004 to indicate that both number and sistent with a lack of fumaroles, shallowest seismicity, and natural hot radiated energy of low-frequency earthquakes are more or less constant spring in and around Mt. Fuji. over time except when a swarm of low-frequency earthquakes was observed between 2000 and 2001 (Fig. 17). Ukawa (2005) also relocated low-frequency earthquakes since the 5. Geophysical monitoring of Mt. Fuji 1980s to show that they are concentrated a few kilometers to the northeast of the summit with depths around 15 km (Fig. 18). Nakamichi Mt. Fuji has been monitored most extensively by seismic methods et al. (2004) focused on more recent low-frequency earthquakes be- since the 1980s when low-frequency earthquakes were observed for the tween 1998 and 2003, leading to a similar conclusion (Fig. 19). Fur- fi rst time there (e.g., Kanjo et al., 1984). Enhancement of seismic net- thermore, they found that the swarm seismicity aligns along a line work in the 1990s allowed us to investigate details of the low-frequency striking from northwest to southeast, consistent with regional stress earthquakes. Ukawa (2005) compiled the cumulative number of field. This seismicity suggests that the generation of low-frequency earthquakes and wave energy generated by the low-frequency

273 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 13. Spatial distribution of seismic reflectors identified by Tsutsui et al. (2007). Labels j, g, and m correspond to the upper surface of the subducting Philippine Sea plate. Labels, a, b, c, d, e, and f correspond to reflectors that are probably due to the presence of volcanic fluids. Solid ellipsoids represent the location of mid-crustal low-frequency earthquakes (MLF; Nakamichi et al., 2004). S1, S2, S3, S4, and K1 represent the location of active sources. K and M denote layers bounded by solid and broken lines and defined by Oikawa et al. (2006).FJF, SKF, and SGF denote the location of Fujigawa, Shibakawa, and Agoyama faults. After Tsutsui et al. (2007).

they changed seismic velocities. Brenguier et al. (2014) inferred that the Tohoku-oki earthquake reduced the seismic velocity of Mt. Fuji at depths shallower than ∼10 km by about 0.1%, which is larger than surrounding areas. They also found that velocity changes of Mt. Fuji are particularly susceptible to either static or dynamic stress changes as in other volcanic regions. The cause of this heterogeneous susceptibility is not understood yet. Araragi et al. (personal communication) recently found that the East Shizuoka earthquake also resulted in a velocity reduction. Comparing the spatial distribution of the velocity drops due to the Tohoku-oki and East Shizuoka earthquakes will gain insights into the mechanics of seismic velocity changes induced by static and dy- namic stress changes. Mt. Fuji has not been only monitored by seismometers, but also by geodetic instruments such as Global Navigation Satellite System (GNSS) and tiltmeters. Fig. 21 depicts the distribution of GNSS sites. The GNSS measurements revealed an expansion of the volcano from late 2005 to − 2010 with a total extensional strain of about 2 × 10 6 (Harada et al., Fig. 14. A two-dimensional resistivity structure beneath Mt. Fuji. Black dots 2010). The observed deformation, however, did not accompany any represent hypocenters between 1998 and 2002. Starts indicate hypocenters of seismic activity. Tiltmeters did not record any signals associated with low-frequency earthquakes. After Aizawa et al. (2004). this expansion because it is too slow to be detected by tiltmeters. However, they would be able to record signals if the deformation would earthquakes beneath Mt. Fuji might be related to dike intrusion at depth has a shorter timescale. (Nakamichi et al., 2004). Comparing these earthquake locations with seismic velocities de- 6. Recommendations for further understanding rived from seismic tomography indicates that the hypocenters of low- frequency earthquakes are characterized by low P-wave velocities and Previous geological and geophysical studies have revealed much high Vp/Vs ratios (Nakamichi et al., 2007; Fig. 11). This led Nakamichi about how Mt. Fuji works, but our understanding is not comprehensive et al. (2007) to suggest that volatiles such as supercritical H2OorCO2, enough to forecast the time and style of the next eruption. Here we rather than partial melt of magmas, are responsible for generating the make some suggestion that might lead to better understanding of Mt. low-frequency earthquakes. Fuji. It starts from how to forecast the location of the future vents Mt. Fuji is now monitored by more than 20 seismic stations, many of (Section 6.1). We then discuss how to forecast the timing, style, and size which are broadband sensors, with five and 18 seismic stations within of the next eruption (Sections 6.2 and 6.3). Finally, we try to suggest five and 15 km from the summit, respectively. The most notable future directions in Section 6.4 to address research questions specificto earthquakes recorded with the current seismic network are probably understand the volcanism of Mt. Fuji. the 2011 Tohoku-oki earthquake (Mw = 9.0) and the East Shizuoka earthquake (Mw = 6.0) that occurred near Mt. Fuji 4 days after the 6.1. Location of future vents Tohoku-oki earthquake. Geodetic data and aftershock distribution show that the East Shizuoka earthquake is an earthquake with a left-lateral Since Mt. Fuji has exclusively erupted from its flanks over the last slip with some thrust component on a fault striking NNE-SSW (Fujita 2200 years, it is important to assess the location of the next eruption not et al., 2013; Fig. 20). The static stress change induced by the earthquake only from a scientific but also from a practical viewpoint. A reasonable – is on the order of 0.1 1 MPa at the magma reservoir of Mt. Fuji, larger way to assess the location is to assume that since the stress state of a than that due to the Tohoku-oki earthquake, which was volcano does not change much over time, the next eruption will occur – 0.001 0.01 MPa (Fujita et al., 2013). more likely at a previous vent. Spatial distribution of previous fissure Although these earthquakes did not trigger any volcanic activity, eruptions compiled by Takada et al. (2007) indicates that the fissures

274 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 15. (a) Plan view of the resistivity distribution around Mt. Fuji at 1, 3, 6, and 10 km below sea level. Areas where reliable resistivity was not obtained are masked by black. Black dots represent where the magnetotelluric measurements were made. (b) Vertical view of resistivity across a profile A–A′ shown (a). Black circles denote hypocenters after the 2011 Tohoku-oki earthquake. Inverted triangles represent where the magnetotelluric measurements were made. After Aizawa et al. (2016).

Fig. 16. Two-dimensional shallow resistivity structure beneath Mt. Fuji. After Aizawa et al. (2005).

Fig. 17. Cumulative number of the DLF events (dotted curve) and cumulative are distributed radially around the summit and distributed more on the wave energy (solid line) between 1980 and 2004. After Ukawa (2005). northwestern and southeastern flanks with fissures striking northwest- southeast (Fig. 22). This distribution is consistent with the notion that stress field generated by the loading of Mt. Fuji itself (e.g., Ukawa, the stress field around Mt. Fuji can be represented by a combination of a 1991; Araragi et al., 2015). It qualitatively suggests that the location of regional stress field with northwest-southeast compression and a local the next eruption will more likely be on the northwestern or

275 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 18. Distribution of seismic stations (as of 2001). Numbers in parentheses represent station corrections for P-wave and S-wave arri- vals, where positive numbers indicate that the observed arrivals are systematically late. (b) Distribution of hypocenters of low-frequency earthquakes since 1987. Only earthquakes with reliable locations were chosen. Hypocenters in three intervals are shown in different colors; green for those between 1987 and March 1995; blue for those between April 1995 and July 2000; red for those between August 2000 and May 2001. (c) Distribution of hypocenters of deep low-frequency earth- quakes considering station corrections. As in (b), only hypocenters with good quality were chosen. Dashed lines indicate boundaries of prefectures. The topographic contour interval in (b) and (c) is 300 m. After Ukawa (2005). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

southeastern flank. 13.5 km from the summit in about 1500 BCE. These flank eruptions Then how can we assess the vent opening probability more quan- were followed by eruptions around the summit by 1000 BCE. Vent lo- titatively? A number of previous studies have proposed methods to cations became far removed from the summit once again, with vents as investigate spatial distribution of the vent opening probability in a far as 13.5 km in 800–900 CE. They became contained within ∼6kmin quantitative way (e.g., Marti and Felpeto, 2010; Connor et al., 2012; around 1000 CE (Takada et al., 2007). Takada (1997, 1999) explained Cappello et al., 2012; Selva et al., 2012) with a number of applications this variability as resulting from stress perturbation due to a series of (e.g, Del Negro et al., 2013; Cappello et al., 2013, 2015; Bevilacqua dike intrusions during the episodes. Taking this temporal variability of et al., 2015; Tadini et al., 2017a, 2017b). While these methods are vent locations into account will further refine the forecast of the loca- variable in mathematical details, most of the studies assess the vent tion of the next eruption, and so a more quantitative assessment is re- opening probability by assuming that the next vent opening is most quired to precisely understand the current state of the stress beneath likely located around preexisting geological structures such as vents, Mt. Fuji, leading to a more sophisticated forecast of the location of the faults, or fractures. In the case of Mt. Fuji, the spatial distribution of next eruption. vent opening probability can be assessed from the spatial distribution of vents of previous eruptions (Takada et al., 2007; Fig. 22). As the am- 6.2. Style of the next eruption bient stress field around Mt. Fuji is highly anisotropic (Ukawa, 1991: Araragi et al., 2015), an improvement of the methodology to take the It is important to assess whether the next eruption is effusive or ambient stress field into account might be necessary to better assess the explosive not only from a scientific but also a practical point of view vent opening probability. because effusive and explosive eruptions cause different hazards (e.g., Vent locations in Mt. Fuji are, in fact, temporally variable (e.g., Schmincke, 2004; Parfitt and Wilson, 2008). While volcanic products Takada et al., 2007). They started to be removed from the summit since from Mt. Fujiare dominated by basaltic rocks, previous records show about 2000 BCE with the farthest vent located at approximately that the volcano has experienced both effusive and explosive eruptions

276 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 19. Distribution of deep low-frequency earthquakes between 1998 and 2003. (a) Map view of relocated hypocenters. The source mechanism of the largest earthquake with M 2.3 is also shown. (b) Cross-section A-A′ in (a). (c) Cross-section B-B′ in (a). After Nakamichi et al. (2004).

Fig. 20. A fault model of the 2011 East Shizuoka earthquake. A gray rectangle represents the modeled fault. Gray circles depict the after- shock distribution. Black and red vectors denote observed and calcu- lated displacements, respectively. After Fujita et al. (2013). (For in- terpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

with significant hazards from both types of eruptions, as discussed in eruptions. There are a number of sophisticated simulations of the lava Section 3.Itisdifficult to forecast the style of the next eruption because flows which take topography and temperature-dependent rheology of various factors such as degassing rate, discharge rate, composition, or lava into account (e.g., Crisci et al., 2004; Vicari et al., 2007; Connor volatile content of magma control the eruption style (e.g., Kozono et al., et al., 2012; Kelfoun and Vargas, 2016; Dietterich et al., 2017 and re- 2013). Volcano monitoring, however, could help forecast the style be- ferences therein), pyroclastic density currents (e.g., Dufek, 2016; cause whether an eruption is explosive or effusive depends on magma Gueugneau et al., 2017; Dioguardi and Mele, 2018, and references discharge rate, thus potentially on the supply rate of magma (e.g., therein) and the dispersion of volcanic ash (e.g., Costa et al., 2016; Kozono et al., 2013) that can be assessed by geodetic measurements. Suzuki et al., 2016, and references therein). These simulations indicate It is also important to assess the possible volcanic hazards from the that a probabilistic assessment of the location, style, or magnitude of an

277 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

compilation by Passarelli and Brodsky (2012) shows that the run-up time and repose time roughly have a linear relation, in which the run- up time after a repose time of 300 years is expected to be around 100 days. Furthermore, the next eruption could be preceded by crises without an eruption well before the actual eruption. The next eruption of Mt. Fuji should be then preceded by detectable unrest. Indeed, the Hoei eruption was preceded by two weeks of intense seismicity (Tsuya, 1955; Koyama, 1998b; Hayashi and Koyama, 2002; Miyashita et al., 2007). A compilation of seismic activity preceding eruptions in Mt. St. Helens, Bezymianny and Alsakan volcanoes show that more intense seismic activity precedes an eruption after a longer repose time. Thelen et al. (2010) found the repose time and cumulative seismic moment roughly has a linear relation. They found that the 1980 Mt. St. Helens, western United States, eruption, after a dormancy of 123 years, and the 1956 Bezymianny, Kamchatka, eruption, after a dormancy of about 1000 years, were preceded by earthquakes with a cumulative moment magnitude exceeding 6.0. Sizable earthquake will likely precede the next eruption of Mt. Fuji. Indeed, many people felt earthquakes, some of which exceeded magnitude 4.0, before the Hoei eruption (Hayashi and Koyama, 2002).

6.4. Why is Mt. Fuji different from other volcanoes?

The recommendations described above are more or less applicable Fig. 21. Distribution of GNSS sites. Squares and circles denote those operated to volcanoes in general. In other words, vent location, time, size, and by Geospatial Information Authority and National Research Institute for Earth style of the next eruption need to be well assessed to better understand Science and Disaster Resilience, respectively. the volcanism and forecast the next eruption of any volcanoes. Here we make suggestion to better address research questions specifictoMt. Fuji, which are, as described in the Introduction, 1) why is Mt. Fuji eruption can lead to a probabilistic hazard assessment. larger than other volcanoes along the same arc?, 2) why has Mt. Fuji dominantly ejected basaltic rocks?, 3) why has Mt. Fuji erupted in 6.3. Timing and size of the next eruption various styles from effusive to explosive eruptions?, and 4) why has Mt. Fuji erupted both from the summit and flank? Mt. Fuji has been dormant for more than 300 years. This repose time Questions 1) and 2) are, in fact, interrelated. A high rate of magma is anomalously long compared with the previous repose time in Mt. supply from depth makes Mt. Fuji large. Also, high rate of magma Fuji. We thus expect that the next eruption will occur soon, but it is supply does not allow magma to reside in the magma reservoir long impossible to predict the precise time of it; previous records of erup- enough to differentiate and make the ejected magma a more evolved tions of Mt. Fuji suggest that the eruptions are neither time-predictable chemical composition like andesite or dacite (e.g., Togashi and nor size-predictable (Fig. 5). Then, how does this long repose time af- Takahashi, 2007). Then why is the rate of magma supply of Mt. Fuji fect the next eruption? A compilation of eruptions in Indonesia after much higher than other volcanoes along the same arc? Ichihara et al. 1800 suggests that a longer repose time tends to lead to a more ex- (2017) hypothesized that a volcano along the Izu-Bonin arc on the plosive eruption (Bebbington, 2014). Although the maximum repose Philippine Sea plate (we hereby call it “Volcano X”) subducted to reach time investigated by Bebbington (2014) is a little more than 100 years, beneath Mt. Fuji ∼100,000 years ago, after which the eruption rate of in contrast to the current repose time of Mt. Fuji of more than Mt. Fuji is 50–500 times larger than that of volcanoes along the same 300 years, extrapolation suggests that the next eruption of Mt. Fuji will arc (Tsukui et al., 1986). Here we explain why this hypothesis can solve more likely be an explosive one. In addition to the statistical assess- the mysteries of Mt. Fuji as described above following Ichihara et al. ments, the current lack of degassing from the surface leads Togashi and (2017). Takahashi (2007) to suggest that the current magma of Mt. Fuji con- When Volcano X starts to subduct, the magma reservoir of Volcano tains a large amount of volatiles. The exsolution of the volatiles would X is expanded because the vent of Volcano X becomes plugged. When make the next eruption explosive. Volcano X reaches the magma conduit of the original Mt. Fuji, then Seismic imaging seems to lack a large magma reservoir with hun- magma reservoirs of the two volcanoes merge to increase the volume of dreds of cubic kilometers in the crust and upper mantle (Nakamichi magma beneath Mt. Fuji. Both magmas originate from the Pacific plate, et al., 2007; Kinoshita et al., 2015; Figs. 11, 12) as seen in caldera- consistent with Nakamura et al. (2008) who showed that the fluid in the forming volcanoes like Yellowstone, western United States (e.g., Huang magma of Mt. Fuji is dominated by that from the Pacific plate. The et al., 2015) or Toba, Indonesia (Jaxybulatov et al., 2014). The lack of magma reservoir of Volcano X, is originally at a depth of ∼10 km be- such large magma reservoirs is consistent with a general lack of large fore subduction, but it gets deeper when it reaches beneath Mt. Fuji volcanoes in stratovolcanoes (e.g., Lin et al., 2014; Kiser et al., 2016). because of subduction, consistent with the depth of the magma re- Therefore it is reasonable to conclude that Mt. Fuji is not capable of servoir inferred from petrological (Fujii, 2007; Kaneko et al., 2010) and generating a catastrophic eruption with a volume of hundreds or seismological (e.g., Lees and Ukawa, 1992; Kinoshita et al., 2015) thousands of cubic kilometers, consistent with a lack previous records evidence. of such eruptions there. However, considering that seismic imaging is A hypothesis proposed by Ichihara et al. (2017), as described above, not capable of imaging smaller magma reservoirs, Mt. Fuji is capable of seems to be consistent with many features Mt. Fuji has qualitatively. A generating smaller but still hazardous eruptions with a volume on the quantitative assessment of whether the hypothesis is plausible or not is order of up to cubic kilometers like the Jogan and Hoei eruptions. required from thermo- and fluid-dynamic point of view. Also, it is de- An eruption of a longer repose time tends to be preceded by a longer sirable to detect the subducted Volcano X from seismic or electro- run-up time between the onset of activity and the eruption. A global magnetic imaging to confirm the hypothesis.

278 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

Fig. 22. Location of fissure eruptions after 15000 BCE. Red, yellow, green, blue lines denote eruptive fissures dated after 300 BCE, between 1500 BCE and 300, between the Kikai-Akahoha eruption (K-Ah; 7300 BCE) and 1700 BCE, and between 15,000 BCE and the Kikai-Akahoya eruption, respectively. After Takada et al. (2007). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The mechanics of the Hoei eruption give insights into how to ad- same location as a previous intrusion possible after the stress change dress Question 3). Kaneko et al. (2010) proposed that a mixture of due to the previous intrusion is relaxed. basaltic magma intruded from a depth and differentiated silicic magma Takada (1997) suggested that the duration of flank and summit at a shallower depth made the Hoei eruption explosive (Fig. 8). Ex- eruption periods are controlled by the strain rate and the supply rate of plosive eruptions of Mt. Fuji in the past (see, for example, Section 3.2) magma, however, the factors controlling the durations are not clearly might have been triggered similarly. To confirm this hypothesis, melt understood. Previous records show that Mt. Fuji has repeated flank and inclusion studies to investigate the magma mixing during past eruptions summit eruptions every ∼1000 years, where summit eruptions domi- are required. nated between 1000 BCE and 1 CE and flank eruptions were dominated A transition between the summit and flank eruptions (Question 4) between 1 CE and 1000 CE (see Fig. 7 of Takada, 1997). An extra- can be discussed in a general context. Takada (1997) found that many polation of this trend suggests that the summit eruptions dominated volcanoes change the location of their eruptions over time in which the between 1000 CE and 2000 CE and is now in a transition period from location of flank eruptions converges toward the summit or the central the summit to flank eruptions. However, this interpolation cannot be volcano before the period of summit eruptions. The mechanics for this justified because of the relative inactivity of Mt. Fuji after 1000 CE. cyclic pattern can be understood by considering the stress perturbation by each dike intrusion (Takada, 1997). A dike intrusion makes the 7. Summary surrounding rock more compressive. This stress field leads to the next intrusion taking place away from the previous intrusion. In other Mt. Fuji is a young volcano which started to grow only around words, if a dike intrusion takes place away from the summit, the next 400 ka ago. Rock chemistry suggests that it has been divided into a few intrusion takes place closer to the summit. The location of dike intru- stages, the most recent of which started around 10 ka. The eruption rate sions converges toward the summit after a series of intrusions. Takada was high between 10 ka and 8 ka when the eruptive volume was more (1997) suggested that viscoelastic relaxation of the host rock relaxes the than 30 km3 DRE only in ∼2000 years, compared with a total of a little stress field perturbed by dike intrusions, making an intrusion at the more than 40 km3 in the last 10,000 years. Mt. Fuji was relatively

279 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282 dormant between 8 ka and 5 ka after which it was reactivated. Some figures are created with Generic Mapping Tools (Wessel and The eruptions of Mt. Fuji over the last 2000 years have exclusively Smith, 1998; Wessel et al., 2013). been from northwestern and southeastern flanks, consistent with the regional stress of northwest-southeast compression. Some vents are References located radially with respect to the summit, consistent with an isotropic stress field near the summit due to the loading of the volcano. The Aizawa, K., Yoshimura, R., Oshiman, N., 2004. Splitting of the Philippine Sea Plate and a largest eruptions during the last 2000 years include the 864 Jogan and magma chamber beneath Mt. Fuji. Geophys. Res. Lett. 31, L09603. https://doi.org/ 3 3 10.1029/2004GL019477. 1707 Hoei eruptions which ejected 2.1 km and 1.6 km of magma, Aizawa, K., Yoshimura, R., Oshiman, N., Yamazaki, K., Uto, T., Ogawa, Y., Tank, S.B., respectively. The Jogan eruption mainly emitted basaltic lavas, while Kanda, W., Sakanaka, S., Furukawa, Y., Hashimoto, T., Uyeshima, M., Ogawa, T., the Hoei eruption was an explosive eruption triggered by the intrusion Shiozaki, I., Hurst, A.W., 2005. Hydrothermal system beneath Mt. Fuji volcano in- ferred from magnetotelurics and electric self-potential. Earth Planet. Sci. Lett. 235, of a basaltic magma into a dacitic magma chamber. The Hoei eruption 343–355. https://doi.org/10.1016/j.epsl.2005.03.023. started 49 days after a nearby M∼8.5 earthquake so that the eruption Aizawa, K., Sumino, H., Uyeshima, M., Yamaya, Y., Hase, H., Takahashi, H.A., Takahashi, was likely triggered by the change of static stress, dynamic stress, or M., Kazahaya, K., Ohno, M., Rung-Arunwan, T., Ogawa, Y., 2016. Gas pathways and – both, due to the M∼8.5 earthquake. Mt. Fuji has been dormant since remotely triggered earthquakes beneath Mount Fuji, Japan. Geology 44, 127 130. https://doi.org/10.1130/G37313.1. 1707, but low-frequency earthquakes have been observed at depths Arai, R., Iwasaki, T., 2015. Transition from collision to subduction and its relation to slab around 15 km. The activity of the low-frequency earthquakes became seismicity and plate coupling. Earth Planets Space 67, 76. https://doi.org/10.1186/ high in 2000, but no associated surficial anomalies were observed. The s40623-015-0247-6. Aramaki, S., 2007. Volcanic disaster mitigation maps of Fuji Volcano and overview of East Shizuoka earthquake (Mw 6.0) right beneath Mt. Fuji on 15 March mitigation programs. In: Aramaki, S. (Ed.), Fuji Volcano, Yamanashi Institute of 2011, four days after the 2011 Tohoku-oki earthquake, did not trigger Environmental Sciences, pp. 451–475 (in Japanese with English abstract). any volcanic activity. Araragi, K.R., Savage, M.K., Ohminato, T., Aoki, Y., 2015. Seismic anisotropy of the upper crust around Mount Fuji, Japan. J. Geophys. Res. 120, 2739–2751. https://doi.org/ To understand the characteristics of Mt. Fuji, the magma plumbing 10.1002/2014JB011554. system of Mt. Fuji has been studied intensively by geological and geo- Argus, D.F., Gordon, R.G., Heflin, M.B., Ma, C., Eanes, R.J., Willis, P., Peltier, W.R., Owen, physical methods. Geological considerations suggest that the magma S.E., 2010. The angular velocities of the plates and the velocity of Earth's Centre from – – space geodesy. Geophys. J. Int. 180, 913 960. https://doi.org/10.1111/j.1365-246X. reservoirs of Mt. Fuji are located at 20 km and 8 9 km beneath the 2009.04463.x. volcano (Kaneko et al., 2010). The depth of the magma reservoir at Bebbington, M.S., 2014. Long-term forecasting of volcanic explosivity. Geophys. J. Int. 20 km is deeper than that in other volcanoes along the same arc because 197, 1500–1515. https://doi.org/10.1093/gji/ggu078. Bevilacqua, A., Isaia, R., Neri, A., Vitale, S., Aspinall, W.P., Bisson, M., Fiandoli, F., of the underplating of the Philippine Sea plate, resulting in an appar- Baxter, P.J., Bertagnini, A., Esposti Ongaro, T., Iannuzzi, E., Pistolesi, M., Rosi, M., ently thicker crust (Fujii, 2007). The shallower reservoir consists of 2015. Quantifying volcanic hazard at Campi Flegrei caldera (Italy) with uncertainty more evolved magma, to which basaltic magma was injected, and assessment: 1. Vent opening maps. J. Geophys. Res. Solid Earth 120, 2309–2329. subsequently triggered the 1707 Hoei eruption (Kaneko et al., 2010). https://doi.org/10.1002/2014JB011775. Brenguier, F., Campillo, M., Takeda, T., Aoki, Y., Shapiro, N.M., Briand, X., Emoto, K., Seismic studies also suggest a magma reservoir at around 20 km, right Miyake, H., 2014. Mapping pressurized volcanic fluids from induced crustal seismic beneath where low-frequency earthquakes occur (Nakamichi et al., velocity drops. Science 345, 80–82. https://doi.org/10.1126/science.1254073. 2007; Kinoshita et al., 2015). Electromagnetic studies suggest a con- Cappello, A., Neri, M., Acocella, A., Gallo, G., Vicari, A., Del Negro, C., 2012. Spatial vent opening probability map of Etna volcano (Sicily, Italy). Bull. Volcanol. 74, ductive body at depths more than 20 km, which is likely to represent a 2083–2094. https://doi.org/10.1007/s00445-012-0647-4. magma reservoir (Aizawa et al., 2004). Seismic observations revealed Cappello, A., Bilotta, G., Neri, M., Del Negro, C., 2013. Probabilistic modeling of future – that low-frequency earthquakes in 2000 were aligned with a strike of volcanic eruptions at Mount Etna. J. Geophys. Res. Solid Earth 118, 1925 1935. https://doi.org/10.1002/jgrb.50190. northwest-southeast a few kilometers to the northeast of the summit Cappello, A., Geshi, N., Neri, M., Del Negro, C., 2015. Lava flow hazards – an impeding (Nakamichi et al., 2004; Ukawa, 2005). threat at Miyakejima volcano, Japan. J. Volcanol. Geotherm. Res. 308, 1–9. https:// For a better understanding of the Mt. Fuji magma plumbing system doi.org/10.1016/j.jvolgeores.2015.10.005. Chesley, C., LaFemina, P.C., Puskas, C., Kobayashi, D., 2012. The 1707 Mw 8.7 Hoei and improved forecasting of the evolution of future eruptions, we need earthqukae triggered the largest historical eruption of Mt. Fuji. Geophys. Res. Lett. to combine geophysical and petrological insights about the location and 39, L24309. https://doi.org/10.1029/2012GL053868. properties of the magma reservoir of Mt. Fuji. Geophysical insights Chung, T.W., Lees, J.M., Yoshimoto, K., Fujita, E., Ukawa, M., 2009. Intrinsic and scat- tering attenuation of the Mt. Fuji Region, Japan. Geophys. J. Int. 177, 1366–1382. gained from seismic, geodetic, and electromagnetic measurements can https://doi.org/10.1111/j.1365-246X.2009.04121.x. determine the location of the magma reservoir. However, these esti- Connor, L.J., Connor, C.B., Meliksetian, K., Savov, I., 2012. Probabilistic approach to mations have uncertainties. Petrological measurements do not provide modeling lava flow inundation: a lava flow hazard assessment for a nuclear facility in any information on the horizontal location of the magma reservoir but Armenia. J. Appl. Volcanol. 1, 3. https://doi.org/10.1186/2191-5040-1-3. Costa, A., Suzuki, Y.J., Cerminara, M., Devenish, B.J., Esposti Ongaro, T., Herzog, M., Van can constrain the depth of the magma reservoir with some uncertainties Eaton, A.R., Denby, L.C., Bursik, M., de'Michieli Vitturi, M., Engwell, S., Neri, A., independently from geophysical measurements. Also what is unique Barsotti, S., Folch, A., Macedonio, G., Girault, F., Carazzo, G., Tait, S., Kaminski, E., about the petrological measurements is that they can assess the magma Mastin, L.G., Woodhouse, M.J., Phillips, J.C., Hogg, A.J., Degruyter, W., Bonadonna, C., 2016. Results of the eruptive column model inter-comparison study. J. Volcanol. ascent rate, which helps assess the magma movement preceding an Geotherm. Res. 325, 2–25. https://doi.org/10.1016/j.jvolgeores.2016.01.017. eruption. With this point of view, it is important to combine these Crisci, G.M., Rongo, R., Di Gregorio, S., Spataro, W., 2004. The simulation model SCIARA: fl complementary measurements to gain more insights into the magma the 1991 and 2001 lava ows at Mount Etna. J. Volcanol. Geotherm. Res. 132, 253–267. https://doi.org/10.1016/S0377-0273(03)00349-4. dynamics beneath Mt. Fuji during its activity to and better forecast an Del Negro, C., Cappello, A., Neri, M., Bilotta, G., Herault, A., Ganci, G., 2013. Lava flow eruption. hazards at Mount Etna: constraints imposed by eruptive history and numerical si- mulations. Sci. Rep. 3, 3493. https://doi.org/10.1038/srep03493. Dietterich, H.R., Lev, E., Chen, J., Richardson, J.A., Cashman, K.V., 2017. Benchmarking Acknowledgments computational fluid dynamics models of lava flow simulation for hazard assessment, forecasting, and risk management. J. Appl. Volcanol. 6, 9. https://doi.org/10.1186/ This project was initiated as a field trip guide for The 3rd s13617-017-0061-x. “ Dioguardi, F., Mele, D., 2018. PYFLOW_2.0: a computer program for calculating flow International Summer School of Earthquake Science Monitoring phy- properties and impact parameters of past dilute pyroclastic density currents based on sical properties associated with tectonic processes” in September 2015 field data. Bull. Volcanol. 80, 28. https://doi.org/10.1007/s00445-017-1191-z. organized by Earthquake Research Institute, The University of Tokyo Dufek, J., 2016. The fluid mechanics of pyroclastic density currents. Annu. Rev. Fluid – fl and Southern California Earthquake Center and The 1st field camp of Mech. 48, 459 485. https://doi.org/10.1146/annurev- uid-122414-034252. Fujii, T., 2007. Magmatology of Fuji volcano. In: Aramaki, S. (Ed.), Fuji Volcano. Asian Consortium of Volcanology in November 2015. Detailed and Yamanashi Institute of Environmental Sciences, pp. 233–244 (in Japanese with constructive reviews by Fidel Costa and three anonymous reviewers English abstract). substantially improved the manuscript. We thank Richard Jordan for Fujita, K., Suzuki, Y., Yoshino, N., Kitagawa, J., Koyama, M., Miyaji, N., Shimoyama, T., Anyoji, N., 2002. Stratigraphy of the Kenmarubi lavas and associated deposits in the language editing which also substantially improved the manuscript.

280 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

north slope of Fuji Volcacno. In: Japan. Japan Earth and Planetary Science Joint beneath Izu-Oshima volcano inferred from a seismic scattering analysis and possible Meeting, Abstract V032–P021. interpretation of the magma plumbing system of the 1986 eruptive activity. Phys. Fujita, E., Kozono, T., Ueda, H., Kohno, Y., Yoshioka, S., Toda, N., Kikuchi, A., Ida, Y., Earth Planet. Inter. 104, 257–269. https://doi.org/10.1016/S0031-9201(97) 2013. Stress field change around the Mount Fuji volcano magma system caused by 00060-5. the Tohoku megathrust earthquake, Japan. Bull. Volcanol. 75, 679. https://doi.org/ Miyaji, N., 2007. Eruptive history, eruption rate and scale of eruptions for the Fuji 10.1007/s00445-012-0679-9. Volcano during the last 11,000 years. In: Aramaki, S. (Ed.), Fuji Volcano. Yamanashi Gueugneau, V., Kelfoun, K., Roche, O., Chupin, L., 2017. Effects of pore pressure in Institute of Environmental Sciences, pp. 79–95 (in Japanese with English abstract). pyroclastic flows: Numerical simulation and experimental validation. Geophys. Res. Miyaji, N., Endo, K., Togashi, S., Uesugi, Y., 1992. Tephrochronological history of Mt. Lett. 44, 2194–2202. https://doi.org/10.1002/2017GL072591. Fuji. In: Kato, H., Noro, H. (Eds.), 29 th IGC Field Trip Guide Book. vol. 4. Volcanoes Hayashi, Y., Koyama, M., 2002. Estimated magnitudes of the earthquakes preceding the and geothermal fields of Japan, Geol. Surv. Jpn, Tsukuba, pp. 75–109. 1707 Hoei eruption of Fuji Volcano. Historical Earthquakes 18, 127–132 (in Japanese Miyaji, N., Togashi, S., Chiba, T., 2004. A large-scale collapse event at the eastern slope of with English abstract). Fuji volcano about 2900 years ago. Bull. Volcanol. Soc. Jpn. 49, 237–248 (in Hirose, F., Nakajima, J., Hasegawa, A., 2008. Three-dimensional seismic velocity struc- Japanese with English abstract). ture and configuration of the Philippine Sea slab in southwestern Japan estimated by Miyaji, N., Kamiji, M., Suzuki, S., Chiba, T., Endo, K., Takahashi, M., Murata, T., double-difference tomography. J. Geophys. Res. 113, B09315. https://doi.org/10. Watanabe, Y., 2006. Phreatic explosion deposit formed by the inflow of Aokigahara 1029/2007JB005274. lava flows at AD 864 to the lake Shoji of Fuji volcano. In: Japan Geoscience Union Huang, H.-H., Lin, F.-C., Schmandt, B., Farrell, J., Smith, R.B., Tsai, V.C., 2015. The Meeting, Abstract V101-010. Yellowstone magmatic system from the mantle plume to the upper crust. Science 348, Miyaji, N., Kan'no, A., Kanamaru, T., Mannen, K., 2011. High-resolution reconstruction of 773–777. https://doi.org/10.1126/science.aaa5648. the Hoei eruption (AD 1707) of Fuji volcano. Japan. J. Volcanol. Geotherm. Res. 207, Ichihara, M., Adam, C., VIdal, V., Grosse, P., Mibe, K., Orihashi, Y., 2017. Dots-and-lines 113–129. https://doi.org/10.1016/j.jjvolgeores.2011.06.013. approach to subduction volcanism and tectonics. J. Geogr. 126, 181–193 (in Miyashita, M., Churei, M., Uhira, K., Hayashi, Y., Katayama, H., Fujii, T., Murakami, M., Japanese with English abstract). Ukawa, M., Shirato, S., Yamasato, H., Yokota, T., 2007. Surveillance of activity of Fuji Ida, Y., 2009. Dependence of volcanic systems on tectonic stress conditions as revealed by volcano, Japan – scenario of the 1707 Hoei Eruption and volcano information. In: features of volcanoes near Izu peninsula, Japan. J. Volcanol. Geotherm. Res. 181, Aramaki, S. (Ed.), Fuji Volcano. Yamanashi Institute of Environmental Sciences, pp. 35–46. https://doi.org/10.1016/j.jvolgeores.2008.12.006. 441–449 (in Japanese with English abstract). Iidaka, T., Mizoue, M., Nakamura, I., Tsukuda, T., Sakai, K., Kobayasi, M., Haneda, T., Nakajima, J., Hirose, F., Hasegawa, A., 2009. Seismotectonics beneath the Tokyo me- Hashimoto, S., 1990. The upper boundary of the Philippine Sea plate beneath the tropolitan area, Japan: effect of slab-slab contact and overlap of seismicity. J. western Kanto region estimated from S–P-converted wave. Tectonophysics 179, Geophys. Res. 114, B08309. https://doi.org/10.1029/2008JB00610. 321–326. https://doi.org/10.1016/0040-1951(90)90297-L. Nakamichi, H., Ukawa, M., Sakai, S., 2004. Precise hypocenter locations of midcrustal Ishida, M., 1992. Geometry and relative motion of the Philippine Sea plate and Pacific low-frequency earthquakes beneath Mt. Fuji, Japan. Earth Planets Space 56, e37–e40. plate beneath the Kanto-Tokai district, Japan. J. Geophys. Res. 97, 489–513. https:// https://doi.org/10.1186/BF03352542. doi.org/10.1029/91JB02567. Nakamichi, H., Watanabe, H., Ohminato, T., 2007. Three-dimensional velocity structures Jaxybulatov, K., Shapiro, N.M., Koulakov, I., Mordret, A., Landès, M., Sens-Schönfelder, of Mount Fuji and the South Fossa Magna, Central Japan. J. Geophys. Res. 112, C., 2014. A large magmatic sill complex beneath the Toba caldera. Science 346, B03310. https://doi.org/10.1029/2005JB004161. 617–620. https://doi.org/10.1126/science.1258582. Nakamura, H., Iwamori, H., Kimura, J.-I., 2008. Geochemical evidence for enhanced fluid Kamiya, S., Kobayashi, Y., 2007. Thickness variation of the descending Philippine Sea flux due to overlapping subducting plates. Nat. Geosci. 1, 380–384. https://doi.org/ slab and its relationship to volcanism beneath the Kanto-Tokai district, Central Japan. 10.1038/ngeo200. J. Geophys. Res. 112, B06302. https://doi.org/10.1029/2005JB004219. Newhall, C.G., Self, S., 1982. The Volcanic Explosivity Index (VEI): an estimate of ex- Kaneko, T., Yasuda, A., Fujii, T., Yoshimoto, M., 2010. Crypto-magma chambers beneath plosive magnitude for historical volcanism. J. Geophys. Res. 87, 1231–1238. https:// Mt. Fuji. J. Volcanol. Geotherm. Res. 193, 161–170. https://doi.org/10.1016/j. doi.org/10.1029/JC087iC02p01231. jvolgeores.2010.04.002. Oikawa, J., et al., 2006. Seismic exploration at Fuji volcano with active sources – the Kanjo, K., Karakama, I., Matsu'ura, R.S., 1984. Seismic activities of Mt. Fuji region de- outline of the experiment and the arrival time data. Bull. Earthq. Res. Inst., Univ. tected by continuous observation of micro-earthquakes. J. Phys. Earth 32, 463–468. Tokyo 81, 71–94. https://doi.org/10.4294/jpe1952.32.463. Parfitt, E.A., Wilson, L., 2008. Fundamentals of Physical Volcanology. Blackwell, pp. 230. Kelfoun, K., Vargas, S.V., 2016. VolcFlow capabilities and potential development for the Passarelli, L., Brodsky, E.E., 2012. The correlation between run-up and repose time of simulation of lava flows. Geol. Soc. Lond. Spec. Publ. 426, 337–343. https://doi.org/ volcanic eruptions. Geophys. J. Int. 188, 1025–1045. https://doi.org/10.1111/j. 10.1144/SP426.8. 1365-246X.2011.05298.x. Kinoshita, S.M., Igarashi, T., Aoki, Y., Takeo, M., 2015. Imaging crust and upper mantle Schmincke, H.-U., 2004. Volcanism. Springer, pp. 324. beneath Mount Fuji, Japan, by receiver functions. J. Geophys. Res. 120, 3240–3254. Selva, J., Orsi, G., Di Vito, M.A., Marzocchi, W., Sandri, L., 2012. Probability hazard map https://doi.org/10.1002/2014JB011522. for future vent opening at the Campi Flegrei caldera, Italy. Bull. Volcanol. 74, Kiser, E., Palomeras, I., Levander, A., Zelt, C., Harder, S., Schmandt, B., Hansen, S., 497–510. Creager, K., Ulberg, C., 2016. Magma reservoirs from the upper crust to the Moho Shibata, T., Yoshimoto, M., Fujii, T., Nakada, S., 2015. Geochemical and Sr-Nd isotopic inferred from high-resolution Vp and Vs models beneath Mount St. Helens, characteristics of Quaternary magmas from the Pre-Komitake volcano. J. Mineral. Washington State, USA. Geology 44, 411–414. https://doi.org/10.1130/G37591.1. Petrol. Sci. 110, 65–70. https://doi.org/10.2465/jmps.141022e. Koyama, M., 1998a. Reevaluation of the 800–802 A.D. eruption of Fuji Volcano, Japan, Suzuki, Y., Fujii, T., 2010. Effect of syneruptive decompression path on shifting intensity and its Influence on the ancient traffic network around the volcano, based on eruptive in basaltic sub-Plinian eruption: Implication of microlites in Yufune-2 scoria from Fuji deposits and historical records. Bull. Volcanol. Soc. Jpn. 43, 349–371 (in Japanese volcano, Japan. J. Volcanol. Geotherm. Res. 198, 158–176. https://doi.org/10.1016/ with English abstract). j.jvolgeores.2010.08.020. Koyama, M., 1998b. Reevaluation of the eruptive history of Fuji Volcano, Japan, mainly Suzuki, Y.J., Costa, A., Cerminara, M., Esposti Ongaro, T., Herzog, M., Van Eaton, A.R., based on historical documents. Bull. Volcanol. Soc. Jpn. 43, 323–347 (in Japanese Denby, L.C., 2016. Inter-comparison of three-dimensional models of volcanic plumes. with English abstract). J. Volcanol. Geotherm. Res. 326, 26–42. https://doi.org/10.1016/j.jvolgeores.2016. Kozono, T., Ueda, H., Ozawa, T., Koyaguchi, T., Fujita, E., Tomiya, A., Suzuki, Y.J., 2013. 06.011. Magma discharge variations during the 2011 eruptions of Shinmoe-dake volcano, Tadini, A., Bisson, M., Neri, A., Cioni, R., Bevilaqua, A., Aspinall, W.P., 2017a. Assessing Japan, revealed by geodetic and satellite observations. Bull. Volcanol. 75, 695. future vent opening locations at the Somma-Vesuvio volcanic complex: 1. A new https://doi.org/10.1007/s00445-013-0695-4. information geodetabase with uncertainty characterizations. J. Geophys. Res. Solid Lees, J.M., Ukawa, M., 1992. The South Fossa Magna, Japan, revealed by high-resolution Earth 122, 4336–4356. https://doi.org/10.1002/2016JB013858. P- and S-wave travel time tomography. Tectonophysics 208, 377–396. https://doi. Tadini, A., Bevilacqua, A., Neri, A., Cioni, R., Aspinall, W.P., Bisson, M., Isaia, R., org/10.1016/0040-1951(92)90436-A. Mazzarini, F., Valentine, G.A., Vitale, S., Baxter, P.J., Bertagnini, A., Cerminara, M., Lin, G., Shearer, P.M., Matoza, R.S., Okubo, P.G., Amelung, F., 2014. Three-dimensional de Michieli Vitturi, M., Di Roberto, A., Engwell, S., Esposti Ongaro, T., Fiandoli, F., seismic velocity structure of Mauna Loa and Kilauea volcanoes in Hawaii from local Pistolesi, M., 2017b. Assessing future vent opening locations at the Somma-Vesuvio seismic tomography. J. Geophys. Res. Solid Earth 119, 4377–4392. https://doi.org/ volcanic complex: 2. Probability maps of the caldera for a future Plinian/sub-Plinian 10.1002/2013JB010820. event with uncertainty quantification. J. Geophys. Res. Solid Earth 122, 4357–4376. Machida, H., 1977. Tephra Talks: Natural History of Volcanoes and Plains. Soju Shobo, https://doi.org/10.1002/2016JB013860. pp. 324 (in Japanese). Takada, A., 1997. Cyclic flank-vent and central-vent eruption patterns. Bull. Volcanol. 58, Mannen, K., Yukutake, Y., Kikugawa, G., Harada, M., Itadera, K., Takenaka, J., 2018. 539–556. https://doi.org/10.1007/s004450050161. Chronology of the 2015 eruption of Hakone volcano, Japan: geological background, Takada, A., 1999. Variations in magma supply and magma partitioning: the role of tec- mechanism of volcanic unrest and disaster mitigation measures during the crisis. tonic settings. J. Volcanol. Geotherm. Res. 93, 93–110. https://doi.org/10.1016/ Earth Planet. Space 70, 68. https://doi.org/10.1186/s40623-018-0844-2. S0377-0273(99)00082-7. Marti, J., Felpeto, A., 2010. Methodology for the computation of volcanic susceptibility: Takada, A., Ishizuka, Y., Nakano, S., Yamamoto, T., Kobayashi, M., Suzuki, Y., 2007. An example for mafic and felsic eruption on Tenerife (Canary Islands). J. Volcanol. Characteristic and evolution inferred from eruptive fissures of Fuji Volcano, Japan. Geotherm. Res. 195, 69–77. https://doi.org/10.1016/j.jvolgeores.2010.06.008. In: Aramaki, S. (Ed.), Fuji Volcano. Yamanashi Institute of Environmental Sciences, Matsubara, M., Obara, K., Kasahara, K., 2008. Three-dimensional P- and S-wave velocity pp. 183–202 (in Japanese with English abstract). structures beneath the Japan Islands obtained by high-density seismic stations by Takada, A., Yamamoto, T., Ishizuka, Y., Nakano, S., 2016. Geological map of Fuji volcano. seismic tomography. Tectonophysics 454, 86–103. https://doi.org/10.1016/j.tecto. In: Miscellaneous Map Series 12, Geological Survey of Japan. National Institute of 2008.04.016. Advanced Industrial Science and Technology. Mikada, H., Watanabe, H., Sakashita, S., 1997. Evidence for subsurface magma bodies Takahashi, N., Kodaira, S., Tatsumi, Y., Yamashita, M., Sato, T., Kaiho, Y., Miura, S., No,

281 Y. Aoki, et al. Earth-Science Reviews 194 (2019) 264–282

T., Takizawa, K., Kaneda, Y., 2009. Structural variations of arc crusts and rifted Ukawa, M., 2005. Deep low-frequency earthquake swarm in the mid crust beneath Mount margins in the southern Izu-Ogasawara arc-back arc system. Geochem. Geophys. Fuji (Japan) in 2000 and 2001. Bull. Volcanol. 68, 47–56. https://doi.org/10.1007/ Geosyst. 10, Q09X08. https://doi.org/10.1029/2008GC002146. s00445-005-0419-5. Thelen, W.A., Malone, S.D., West, M.E., 2010. Repose time and cumulative moment Vicari, A., Alexis, H., Del Negro, C., Coltelli, M., Marsella, M., Proietti, C., 2007. Modeling magnitude: a new tool for forecasting eruptions? Geophys. Res. Lett. 37, L18301. of the 2001 lava flow at Etna volcano by a Cellular Automata approach. Environ. https://doi.org/10.1029/2010GL044194. Model. Softw. 22, 1465–1471. https://doi.org/10.1016/j.envsoft.2016.10.005. Togashi, S., Takahashi, M., 2007. Geochemistry of rocks from the Fuji volcano, Japan: Wessel, P., Smith, W.H.F., 1998. New, improved version of Generic Mapping Tools re- constraints for evolution of magmas. In: Aramaki, S. (Ed.), Fuji Volcano. Yamanashi leased. EOS Trans. Am. Geophys. Union 79, 579. https://doi.org/10.1029/ Institute of Environmental Sciences, pp. 219–232 (in Japanese with English abstract). 98EO00426. Tsukui, M., Sakuyama, M., Koyaguchi, T., Ozawa, K., 1986. Long-term eruption rates and Wessel, P., Smith, W.H.F., Scharroo, R., Luis, J., Wobbe, F., 2013. Generic Mapping Tools: dimensions of magma reservoirs beneath Quaternary polygenetic volcanoes in Japan. improved version released. EOS Trans. Am. Geophys. Union 94, 409–420. https:// J. Volcanol. Geotherm. Res. 29, 189–202. https://doi.org/10.1016/0377-0273(86) doi.org/10.1002/2013EO450001. 90044-2. Yamamoto, T., Takada, A., Ishizuka, Y., Miyaji, N., Tajima, Y., 2005a. Basaltic pyroclastic Tsumura, N., Horiuchi, S., Hasegawa, A., Kasahara, K., 1993. Location of the upper flows of Fuji volcano, Japan: characteristics of the deposits and their origin. Bull. boundary of the Philippine Sea plate beneath the eastern part of Yamanashi pre- Volcanol. 67, 622–633. https://doi.org/10.1007/s00445-004-0398-y. fecture estimated from S to P converted waves. J. Seismol. Soc. Jpn. 46, 109–118 (in Yamamoto, T., Takada, A., Ishizuka, Y., Nakano, S., 2005b. Chronology of the products of Japanese with English abstract). Fuji Volcano based on new radiometric carbon ages. Bull. Volcanol. Soc. Jpn. 50, Tsutsui, T., Oikawa, J., Kagiyama, T., Reserch group for seismic exploration of Fuji vol- 53–70 (in Japanese with English abstract). cano, 2007. Seismic reflection profilings of Fuji volcano, Central Japan. Butsuri- Yamamoto, T., Takada, A., Ishizuka, Y., Miyaji, N., Tajima, Y., 2007. Basaltic pyroclastic Tansa 131–144 (in Japanese with English abstract), 60. flows of Fuji volcano, Japan: characteristics of the deposits and their origin. In: Tsuya, H., 1955. Geological and petrological studies of Volcano, Fuji, V. 5. On the 1707 Aramaki, S. (Ed.), Fuji Volcano. Yamanashi Institute of Environmental Sciences, pp. eruption of Volcano Fuji. Bull. Earthq. Res. Inst., Univ. Tokyo 33, 341–383. 245–254 (in Japanese with English abstract). Tsuya, H., 1962. Geological and petrological studies of Volcano Fuji (VI). 6. Geology of Yoshimoto, M., Fujii, T., Kaneko, T., Yasuda, A., Nakada, S., 2004. Multiple magma re- the Volcano as observed in some borings on its flanks. Bull. Earthq. Res. Inst., Univ. servoirs for the 1707 eruption of Fuji volcano. Japan. Proc. Japan Acad. Ser. B. 80, Tokyo 40, 767–804. 103–106. https://doi.org/10.2183/pjab.80.103. Uesugi, Y., 1990. Standard tephra columns for the eastward area of Volcano Fuji, part 1: Yoshimoto, M., Fujii, T., Kaneko, T., Yasuda, A., Nakada, S., Matsumoto, A., 2010. S-25 – Y-141. Bull. Assoc. Kanto Quat. Res. 16, 3–28 (in Japanese). Evolution of Mount Fuji, Japan: Inference from drilling into the subaerial oldest Ukawa, M., 1991. Collision and fan-shaped compressional stress pattern in the Izu block volcano, pre-Komitake. Island Arc 19, 470–488. https://doi.org/10.1111/j.1440- at the northern edge of the Philippine Sea plate. J. Geophys. Res. 96, 713–728. 1738.2010.00722. https://doi.org/10.1029/90JB02142.

282