Ballistic ejecta and eruption condition of the vulcanian explosion of Shinmoedake volcano, Kyushu, Japan on 1 February, 2011
The physical condition of the 1 February, 2011, vulcanian explosion at Shinmoedake volcano, Japan, is estimated based on the size of impact craters created by ballistic ejecta, using a ballistic trajectory model and a scaling law for impact crater formation. The initial velocity, impact velocity and mass of ejecta were estimated at 240–290 m/s, 140 ± 20 m/s and 1–3 ton, respectively. The gas mass fraction at the source was calculated to be 0.04–0.1, using the initial velocity and a theoretical model of vulcanian explosion. This gas mass fraction is higher than the petrologically estimated value for pre-eruptive magma. Low-angle jets from the explosion and the estimated depth and size of a pressurized gas region suggest a shallow source inside the lava dome. The observation and results imply that segregation and accumulation of gas in a shallow conduit played a role in an increase of excess pressure immediately below the dome surface, prior to the vulcanian explosion.
Key wordsBallistic ejecta impact crater lava dome vulcanian Shinmoedake
A transient volcanic explosion is one of the typical eruption styles of andesitic volcanoes. It is characterized by explosions separated by intervals of minutes to days, and a violent nature with an instantaneous release of pressurized gas and solid/molten materials (e.g., Morrissey and Mastin, 2000). This eruption style, called Vulcanian, occurs as a consequence of magma ascent in a conduit and a gas-rich accumulation region at the top of the magma column (Self et al., 1979; Wilson, 1980). If pathways of gas to the surface are efficiently sealed by a coherent rock, gas pressure may increase to a point at which the sealing rock fails catastrophically, so that the solid/molten materials and gas are violently expelled in a transient explosion (Self et al., 1979; Wilson, 1980). In certain cases, a lava dome may act as a cap, and its failure may trigger transient explosions (e.g., Galeras volcano, Stix et al., 1997; Soufriere Hills volcano, Robertson et al., 1998; Novarupta volcano, Adams et al., 2006). Velocities of ejected materials may become supersonic in the surrounding atmosphere, as evidenced by the observation of shock waves (e.g., Nairn, 1976). The intensity, duration, and course of vulcanian-type explosions are varied, and the transitional behaviors from/to subplinian and strombolian eruptions are occasionally observed. A discrete, short-lived explosion like a cannon and a relatively-long, sustained explosion like a fire hose are probably two end-member types of vulcanian explosions (McBirney, 1973).
The determination of the initial physical condition, such as the initial velocity and ejection angle of materials, is important to constrain the mechanism and process of explosion and to assess resultant hazards. Modeling studies of this type of eruption furnish a relationship between the velocity with which blocks are expelled from an explosion site, and pressure in the trapped gas that produces the explosion (Minakami, 1942; Wilson, 1980; Fagents and Wilson, 1993; Woods, 1995). However, it is not easy to determine the eruption condition from direct observations (Nairn and Self, 1978; Steinberg and Babenko, 1978; Druitt et al., 2002; Formenti et al., 2003). In certain cases, geological and petrological evidence for explosions offer constraints on this condition (e.g., Stix et al., 1997; Robertson et al., 1998; Adams et al., 2006).
2. Outline of the Vulcanian Explosion on 1 February, 2011
The Shinmoedake eruption in 2011 began with small precursory phreatomagmatic explosions on 19 January, followed by successive subplinian explosions during 26–27 January. During the subplinian stage, the column height of each explosion reached 6–7 km above the summit crater (Maeno et al., 2012; Shimbori and Fukui, 2012). After this stage, a lava dome began to grow upward at the center of the summit crater beginning 28 January. The dome was initially a small black-colored pile of diameter ~30 m, and then it grew to fill the summit crater (Nakada et al., 2013). In two days, the dome diameter reached about 500 m (Fig. 1). Steam rose vigorously from the top of the dome, and lava wrinkles strongly developed on its surface.
Numerous ballistic ejecta were expelled from the summit crater through the expanding explosion cloud. The explosion cloud reached the rim of the summit crater (~300 m) in a few seconds. Finer materials were entrained into the rising plume. The plume height reached 2 km above a vent based on visual observation by JMA, while Shimbori and Fukui (2012) reported that the plume height reached 7 km above sea level based on Doppler radar observation. The plume then moved toward the southeast. Pyroclastic flows were not observed. The maximum amplitude of seismicity (3767 µm/s) was recorded at the Shinmoedake Southwest station, 1.7 km from the summit crater, and maximum air pressure (458.4 Pa) was measured at the Yunono station, 3 km southwest of the crater. The amplitudes of seismicity were the largest in the historical record of Shinmoedake eruptions (Japan Meteorological Agency, 2011). In the city of Kobayashi, over 10 km from the crater, roofs and windows of buildings were partially destroyed by small fragments of lava transported by the plume. In areas at the southern foot of the volcano, windows were broken by the shock wave. In some places to the southwest where ballistic ejecta and their fragments were landed, fires were started. Tephra distribution was unknown; hence tephra volume is not constrained, but probably is on the order of 104 m3 based on a comparison to tephra data our group obtained for similar-scale vulcanian explosions during February to March 2011 (Unpublished data of Earthquake Research Institute).
In the same morning after the vulcanian explosion, the lava dome diameter was measured at about 500 m by a JMA airborne survey. This size was nearly the same as that on 31 January. The dome height was slightly reduced because the explosion destroyed its surface. Although the change of dome height was not precisely measured, it seemed very small compared to the thickness of the lava dome. The dome center was covered by blocky lava, but lava wrinkles at the outer part of the dome were still observed. The observation indicates that the explosion site was confined to the dome center, approximately within 250 m in diameter (half the size of the dome). The Geospatial Information Authority of Japan (2011) estimated the altitude of the top of the lava dome to be 1360 m, based on Daichi Phased Array type L-band Synthetic Aperture Radar (PALSAR) intensity images on the same day after the explosion. The dome growth then slowed and almost ceased the next day. Based on these observations, we assumed that ballistic ejecta were launched from the lava dome center at the altitude of 1360 m. The thickness of the lava dome at the explosion site was estimated to be about 130 m, based on topographic data of pre- and post-explosion.
3. Impact Craters Produced by Ballistic Ejecta
Characteristics of impact craters produced by ballistic ejecta.
Crater radius (m)
Crater depth (m)
Impact angle (°)
Ballistic blocks were dense but have a few bubble textures, and there was no evidence of delayed vesiculation such as bread crust surfaces (Fig. 5(a)). This suggests that ballistic ejecta found in this area originated from solid lava that would seal pressurized gas beneath the dome but were still at a high temperature, so they caused fires and burned grass when they emplaced.
4. Petrological Characteristics of Ballistic Ejecta
The chemical composition of lava is andesite with 57.9 wt.% of SiO2. The lava contains 30 vol.% of phenocrysts (plagioclase, orthopyroxene, clinopyroxene and Fe-Ti oxide). Detailed petrological features of erupted materials from this explosion are described by Suzuki et al. (2013a). They suggest that the andesite was produced by syneruptive magma mixing before the explosion, and that the magma temperature was 960–980°C based on phenocryst compositions and thermodynamic calculations. The temperature inside the lava dome is unknown. We therefore assumed that the gas temperature was almost the same as the magma temperature (970°C) or much lower (770°C) in the calculation of the gas/mass fraction at the time of the explosion (Subsection 5.4). The water content of the melt in the conduit magma is uncertain, but less than 3 wt.% is plausible. This value is estimated from a phase equilibrium relationship in a storage condition of pre-eruptive magma (at 125 MPa) and H2O analyses of melt inclusions, as discussed in Suzuki et al. (2013a). The density of lava blocks expelled from the dome surface is uncertain. Hence, we assumed this to be 2100–2400 kg/m3. The minimum value of this range is a one measured by Y. Suzuki (personal comm.).
5. Estimation of Eruption Conditions
5.1 Calculation of trajectories of ballistic ejecta
Parameters used in the calculations of ballistic trajectories.
Temperature at sea level
Thermal lapse rate
Speed of tailwind*
Diameter of cubic-shaped block
Density of block
Extent of zone of reduced drag
Elevation of takeoff point
Elevation of landing point
5.2 Scaling law for impact crater formation
The impact velocity vf of ballistic ejecta is correlated with impact crater size r, using a scaling law for crater formation (Housen et al., 1983; Holsapple, 1993; Housen and Holsapple, 2011). This scaling law has been applied to a wide range of cratering phenomena, such as water drops (on the order of 1 m/s) to meteorite impact (on the order of 1 km/s), when the impact angle θf is >45° (Housen and Holsapple, 2011). Impact processes are generally characterized by an inverse Froude number Open image in new window, where g is the gravitational acceleration and d is the impactor diameter, assuming a sphere (Holsapple, 1993). When the inverse Froude number is large and the effect of gravity is greater than that of the strength of basal materials, the impact process is categorized as the gravity regime (Holsapple, 1993). The impact of ballistic ejecta caused by vulcanian explosions is generally thought to fall within this regime, because the inverse Froude number (on the order of 10−4) is almost equivalent to previously estimated values for cratering in this regime (Holsapple, 1993).
Considering the above relationship (Fig. 9), for example, to produce an impact crater of 3.5 m radius in soil with μ = 0.41, an impact velocity of 100 m/s is required for 3.2 ton ejecta, 155 m/s for 1.8 ton ejecta, and >190 m/s for less than 1.2 ton ejecta. It is important to note that for producing the same sized crater, heavier impactors require a smaller impact velocity, but lighter ones require a higher impact velocity. We neglected the effect of density difference (2100 and 2400 kg/m3) because it was very small compared to errors from other factors (<2% error for impact velocity).
5.3 Initial condition of ballistic ejecta
Based on the results of trajectory calculations (Fig. 8), to simultaneously achieve impact angles 50–60° and distance 3.4 km (location A), the impact velocity must be 138–164 m/s for 3.2 ton ejecta, 126–150 m/s for 1.2 ton, and 108–126 m/s for 0.3 ton. However, to produce an impact crater with R = 3.5 m, the impact velocity must be <100 m/s for 3.2 ton and >200 m/s for 0.3 ton (Fig. 9). Ejecta of 0.8–2 ton can satisfy the scaling relation for a 3.5-m crater. For a 4-m crater at 3.1 km (location B), 1.2–3.2 ton ejecta can satisfy this relation.
If Eq. (5) is integrated twice, the maximum velocity of the expanding envelope, corresponding to the initial ejection velocity, vi, in the trajectory model, is obtained. The implication is that the initial drag forces acting on the blocks are essentially zero, only becoming significant as the gas velocity decays. Fagents and Wilson (1993) showed that with allowance for this effect, pre-explosion gas pressures completely compatible with rock tensile strengths were predicted for a number of well-documented examples. Our approach is to infer the mass ratio of solid material (ms) to volatile concentrations (mg) responsible for the ejection velocity of ballistic ejecta, calculated with Mastin (2008) as discussed above. The vulcanian explosion would begin once the gas pressure (P0) exceeds that from the overlying dome mass, or when the cool outer dome material fails in a brittle manner. Although Fagents and Wilson (1993) modeled an explosion from a large pressurized gas cavity enclosed by a solid caprock, this approach can also be adapted to a system where gas and magma are intermixed (Robertson et al., 1998). In the following discussion, we focus only on the case where a large gas cavity exists.
5.4 Gas/solid mass fraction and source condition
Ballistic blocks with a few bubble textures suggest that the lava dome surface was degassed and solidified prior to disruption and ejection. The pre-explosion gas pressure (P0) would therefore have had to overcome a bulk lava strength consisting of the yield strength of the molten interior and the tensile strength of the solid exterior of the lava dome. For previously studied vulcanian explosions, it was demonstrated that explosion pressure is on the order of 1–10 MPa, entirely compatible with the expected range of tensile strengths of caprock materials (e.g., Nairn and Self, 1978;Self et al., 1979; Fagents and Wilson, 1993). Field examinations of hot dacitic lava blocks have indicated that their strengths may be low. These blocks were easily shattered by blows from a hammer (Mt. St. Helens, Mellors et al., 1988; Mt. Unzen, Sato et al., 1992). In fact, fragmented and shattered ballistic blocks that we surveyed also indicate that the strength of dome lava was very low (Fig. 5(b)). These lower strengths may be attributed to thermal stresses, trapped residual gas, or high strain caused by the presence of a large proportion of crystals (30 vol.% in the Shinmoedake case, Suzuki et al., 2013a) within the viscous melt. It follows that caprock consisting of hot juvenile material should display similar failure properties. The strength of silicic magma is not well constrained, but is likely to be within 1–10 MPa (Romano et al., 1996) and may be reduced by the presence of crystals and bubbles caused by local stress amplification (Spieler et al., 2004). Thus, it may be argued that maximum tensile strength is on the order of 10 MPa. Given uncertainties in the bulk strength of the dome material, we therefore adopted a range of initial pressure P0 = 5–50 MPa. Although this choice of pressure is arbitrary, it allows us to bracket the range of plausible explosion pressures.
A mass ratio of gas to solid material (mg/ms) is estimated by Eqs. (5) and (6), if we assume a large pressurized gas cavity with height ri enclosed by a solid caprock with height rc (= rs − ri). For the same variation of gas temperature and ejecta density as in Fig. 11(a), the ratio of caprock height to gas region height is calculated to be approximately 0.02–0.2 (Fig. 11(b)). In the calculations, we adopt results that the initial acceleration of ejecta occurred over a short distance (less than a few hundred meters), which lie on a range of error for estimation of the flight distance of ballistic ejecta.
The initial velocity of ballistic ejecta was estimated to be 240–290 m/s. This value does not contradict previous studies of vulcanian-type explosions, in which initial velocities of ballistic ejecta were estimated to be 100–400 m/s (e.g., Morrissey and Mastin, 2000; Parfitt and Wilson, 2008). For the Sakurajima volcano in Japan, those for representative vulcanian explosions in 1982–1987 were estimated to be 110–240 m/s, based on analyses of ballistic trajectories, and the explosion pressure was constrained to be 14–63 MPa (Iguchi et al., 1983; Ishihara, 1985,1990). For the Soufriere Hills volcano, Montserrat, ballistic ejecta of the explosion on 17 September, 1996, was expelled with a velocity of 180 m/s and an explosion pressure of 27.5 MPa (Robertson et al., 1998). Druitt et al. (2002) and Formenti et al. (2003) estimated the exit velocities of jet accompanying ballistics at < 150 m/s and the initial pressure at a few to 10 MPa for other vulcanian explosions at the Soufriere Hills volcano. Our estimation of the initial velocity of ejecta for the Shinmoedake explosion is almost equivalent to the maximum case for the Sakurajima volcano and higher than examples from the Soufriere Hills volcano.
It is noteworthy that the initial velocity of ballistic ejecta is not correlated with eruption scale, such as erupted volume and plume height. For typical vulcanian explosions accompanying ballistic ejecta at the Soufriere Hills volcano, the plume height reached > 10 km above sea level, and the erupted volume was on the order of 105 ~ 106 m3 (Robertson et al., 1998; Druitt et al., 2002). On the other hand, for the vulcanian explosion at Shinmoedake on 1 February, 2011, the plume height was ~7 km above sea level (based on Doppler radar observation, Shimbori and Fukui, 2012) and the erupted volume was probably on the order of 104 m3. Explosion events in Soufriere Hills created large plumes sustained by a gas stream, and a large amount of vesicular magma was erupted from a deeper part of the conduit (a depth of 0.5–2 km or more; Druitt et al., 2002; Formenti et al., 2003). The explosion at Shinmoedake volcano was more violent, but with a smaller volume and lower plume height. Ejected material was probably derived from the lava dome itself. The difference between these explosion styles indicates two end-member types of vulcanian explosion (McBirney, 1973): ‘cannon-like’ (for Shinmoedake volcano) and ‘gas-stream’ (for Soufriere Hills volcano).
A noticeable observation for the explosion at Shinmoedake volcano was that the jet angle of the explosion was very low (20–40°) and directed toward the east and west (Fig. 2). This type of low-angle jet is likely to reflect the source depth. Ohba et al. (2002) studied the relationship between shapes of explosion clouds, explosion energies (J) and scaled depths that are defined as the explosion depth divided by the cube-root of the explosion energy (m/J1/3). In their experiments, funnel-shaped, low-angle jets were observed in shallow-depth explosions. Since cloud shapes were similar at the same scaled depths, they asserted that the scaling law is a good approximation for the shapes of explosion clouds. The scaling arguments for explosion phenomena are useful to constrain the source parameters, so that they were applied to some examples of volcanic explosions (e.g., Yokoo et al., 2005). Based on a comparison between the shape of the observed eruption cloud (Fig. 2) and the explosion experiments of Ohba et al. (2002), it is plausible that the vulcanian explosion source was at a very shallow depth, where the scaled depth is at least <0.002 m/J1/3.
Although the explosion crater was not clearly observed at the dome surface, the dome height was slightly decreased and the dome center was covered by blocky lava due to the explosion. Our interpretation is that the area within a half of dome diameter (∼250 m in maximum) was destroyed and fragmented by the explosion and the depression was buried by fallback, because lava wrinkles were still well developed, except for the dome center. To further constrain the source condition, we assume that the (hidden) explosion crater was in the area of blocky lava, and also that the explosion follows the scaling law describing the relationship between explosion energy E and explosion crater size Dexp (Goto et al., 2001; Ohba et al., 2002).
The scaling law is written as log Dexp = 0.32 log E − 2.06. In order to produce the diameter of the explosion site (<250 m), E should be < 8 × 1013 J. Moreover, according to the relationship between the scaled depth (m/J1/3) and the scaled distance of the ballistic ejecta (m/J1/3) found in field experiments (figure 4 in Yokoo et al., 2005), the maximum scaled distance should be 0.2–0.3 m/J1/3 when the scaled depth is <0.002 m/J1/3, which corresponds to near-surface explosions. Based on this relationship, the explosion energy produced ballistic ejecta flying at least 3400 m is estimated to be > 1 × 1012 J. The estimated range of the explosion energy (1 × 1012 < E < 8 × 1013) constrains the dimension of the conical-shaped gas region characterized by ri, rc and rs through Eq. (7). The depth of the gas region top rc could be 4~20 m for P0 = 50 MPa (Fig. 12(a)), or 1~4 m for P0 = 5 MPa (Fig. 12(b)). These depths indicate that the pressurized gas region was in a very shallow level inside the lava dome with a thickness of approximately 130 m.
In fact, the mechanism of the violent explosion at Shinmoedake volcano on 1 February, 2011 is still ambiguous, because tephra deposit (grain size variation, etc.) and petrological characteristics are not constrained enough. Nevertheless, the estimated physical condition provides information of the explosion source. The mass ratio of gas to solid material, 0.04–0.1, estimated in this study exceeds the volatile content of andesitic magma beneath the Shinmoedake volcano (approximately 0.03–0.04 at 115 MPa) estimated from petrological analyses (Suzuki et al., 2013a). In fact, the volatile content could be less than 0.03 in a shallower conduit or lava dome. The discrepancy between the estimated mass ratio and the volatile content may be explained by vaporization of external water in porous country rocks in contact with the magma (e.g., Self et al., 1979), and/or by the concept that excess gas for driving the explosion is contributed by the degassing and crystallizing magma body (e.g., Stix et al., 1997; Druitt et al., 2002).
The idea of vaporization of external water is supported by the geological background that a shallow hydrothermal system is developed beneath the Kirishima Volcano Group (e.g., Kagiyama et al., 1996). This is also supported by the facts that most of the historical explosive activities in the Shinmoedake volcano is characterized by phreatic and phreatomagmatic eruptions (Imura and Kobayashi, 2001). Even for vulcanian explosions at the later stage of the 2011 event, magma-water interaction is suggested based on a microscopic analysis of ash particles (Suzuki et al., 2013b). However, this idea is questionable for the 1 February vulcanian explosion because it is likely that the shallow hydrothermal system had been mostly dried up during the subplinian and dome-building stages in January.
The vulcanian activity on 1 February, 2011, was the first explosive event of a series of transient explosions that continued over the next several months. The lava dome rapidly grew in the summit crater within a few days after subplinian explosions, so the overlying dome might efficiently seal a vent and prevent gas emission. Although there were some fumarolic activities on the dome, no open conduit or crater was observed on the dome surface before 1 February. Accumulation of gas inside the dome, or near the dome surface, might cause excess pressure. A system where gas and magma are intermixed is also possible for gas accumulation. With time, pressure eventually increased by further accumulation of gas and finally overcame the bulk strength of the lava dome, then catastrophically broke the dome surface. One of the plausible mechanisms of a pressurizing lava dome is the exsolution of magmatic volatiles from rapid microlite growth (e.g., Stix et al., 1997; Druitt et al., 2002). Even small amounts of crystallization (<20 vol.%) can produce overpressures of 5–20 MPa, which are sufficient to fracture rocks (Tait et al., 1989). Detailed petrological analyses of dome lava will be needed for further investigations of this scenario.
A vulcanian explosion from an unconfined, horizontal dome surface, as observed in Shinmoedake volcano, is probably common for andesitic volcanoes but has not been studied well. Further studies on the mechanism and dynamics of this type of vulcanian explosion will be interesting. They should be constrained based on geophysical observations such as seismic records and acoustic signals (shock waves), that we did not focus on in this study, as well as petrological studies on the vesicularity, crystallinity and volatile content of juvenile material which will be useful to constrain the physical condition of explosion. Determination of the initial physical condition, such as the initial velocity and ejection angle of materials, is important not only to constrain the mechanism and process of the explosion, but also to assess resultant hazards. Impacts of ballistic ejecta is one of the most hazardous phenomena in vulcanian explosions, as witnessed in recent eruptions at Colima volcano, Mexico (1999), Popocatepetl volcano, Mexico (2003), Asama volcano, Japan (2004), etc. The physical condition will contribute to the determination of a distribution of ballistic ejecta and their impacts around the volcano quantitatively. The combined approach using geological data of ballistic ejecta and modeling results will also enable making hazard zonation and establishing more sophisticated assessments of hazards.
Impact craters were created by ballistic ejecta during the vulcanian explosion of Shinmoedake volcano, on 1 February, 2011. Based on the size of impact craters and a ballistic trajectory model and scaling law for impact crater formation, the initial velocity and impact velocity of ballistic ejecta were estimated to be 240–290 m/s and 140 ± 20 m/s, respectively. The mass of ballistic ejecta was estimated to be 0.8–2 ton for a location 3.4 km distant, and 1.2–3.2 ton at a location 3.1 km distant. The gas mass fraction at the source was calculated to be 4–10 wt.%, using a theoretical model of vulcanian explosion. This fraction is larger than the petrologically-estimated value for pre-eruptive magma (0.03–0.04 at 115 MPa, Suzuki et al., 2013a). Assuming a large pressurized gas cavity enclosed by a solid caprock, the ratio of caprock height to gas region height was computed to be 0.02–0.2. Low-angle jets of the vulcanian explosion and the estimated depth and size of the gas region indicate that the explosion occurred at a very shallow depth inside the lava dome. The observation and results imply that segregation and accumulation of gas in a shallow conduit played a role in increasing excess pressure prior to the vulcanian explosion.
We thank the JMA (Japan Meteorological Agency) and the city of Kirishima for a field survey in the Shinyu area. We are also grateful to the JMA and Kagoshima Prefecture for providing video images of the explosion of Shinmoedake volcano on 1 February, 2011. We are also grateful to Y. Suzuki for discussion about the petrological features of erupted magma, and to M. Ichihara for providing information on sequential photographs of the explosion. The manuscript was improved by insightful comments from two anonymous reviewers. This study was supported by a Grant-in-Aid for Scientific Research (No. 22900001) from MEXT, Japan.
- Baker, W. E., P. A. Cox, P. S. Westine, J. J. Kulesz, and R. A. Strehlow, Explosion Hazards and Evaluation, 807 p., Elsevier, Amsterdam, 1983.Google Scholar
- Druitt, T. H., S. R. Young, B. Baptie, C. Bonadonna, E. S. Calder, A. B. Clarke, P. D. Cole, C. L. Harford, R. A. Herd, R. Luckett, G. Ryan, and B. Voight, Episodes of cyclic Vulcanian explosive activity with fountain collapse at Soufriere Hills Volcano, Montserrat, in The Eruption of Soufriere Hills Volcano, Montserrat, from 1995 to 1999, edited by T. H. Druitt and B. P. Kokelaar, Geol. Soc. London Mem., 21, 483–516, 2002.CrossRefGoogle Scholar
- Geospatial Information Authority of Japan, Crustal deformations around Kirishima Volcano, Rep. Coord. Comm. Predict. Volcan. Erup., 108, 197–220, 2011.Google Scholar
- Iguchi, M., K. Ishihara, and K. Kamo, On ejection of velocity and pressure of volcanic explosion reduced from photo-trajectory of volcanic bombs, Ann. Disas. Prev. Res. Inst., Kyoto Univ., 26(B-1), 9–21, 1983 (in Japanese with English abstract).Google Scholar
- Imura, R. and T. Kobayashi, Geological map of Kirishima volcano, Geological Survey of Japan, pp. 8, 2001 (in Japanese with English abstract).Google Scholar
- Ishihara, K., Pressure sources and induced ground deformation associated with explosive eruptions at an andesitic volcano: Sakurajima volcano, Japan, in Magma Transport and Storage, edited by M. P. Ryan, 335–356, Wiley and Sons, New York, 1990.Google Scholar
- Japan Meteorological Agency, Volcanic activity of Kirishima Volcano— February, 2011–May, 2011—, Rep. Coord. Comm. Predict. Volcan. Erup., 109, 139–167, 2011.Google Scholar
- Kagiyama, T., H. Utada, M. Uyeshima, F. Masutani, W. Kanda, Y. Tanaka, H. Masuda, H. Murakami, I. Shiozaki, M. Ichiki, T. Yukutake, T. Mogi, K. Amita, N. Oshiman, and M. Mishina, Resistivity structure of the central and the southeastern part of Kirishima volcanoes, Bull. Volcanol. Soc. Jpn., 41, 215–225, 1996 (in Japanese with English abstract).Google Scholar
- Maeno, F., M. Nagai, S. Nakada, R. E. Burden, S. Engwell, Y. Suzuki, and T. Kaneko, Constraining tephra dispersion and deposition from cyclic subplinian explosions at Shinmoedake volcano, Kyushu, Japan, 2011, Japan Geoscience Union Meeting, SVC50-07, 2012.Google Scholar
- Mastin, L. G., A simple calculator of ballistic trajectories for blocks ejected during volcanic eruptions, version 1.4, U.S. Geol. Sur. Open File Rep., 01–45, 2008.Google Scholar
- Minakami, T., On the distribution of volcanic ejecta (Part I). The distributions of volcanic bombs ejected by the recent explosions of Asama, Bull. Earthq. Res. Inst. Univ. Tokyo, 20, 65–91, 1942.Google Scholar
- Morrissey, M. M. and L. G. Mastin, Vulcanian eruptions, in Encyclopedia of Volcanoes, edited by H. Sigurdsson, 463–476, Academic Press, San Diego, 2000.Google Scholar
- Parfitt, E. A. and L. Wilson, Fundamentals of Physical Volcanology, pp. 230, Blackwell Science Ltd., 2008.Google Scholar
- Robertson, R., P. Cole, R. S. J. Sparks, C. Harford, A. M. Lejeune, W. J. McGuire, A. D. Miller, M. D. Murphy, G. Norton, N. F. Stevens, and S. R. Young, The explosive eruption of Soufriere Hills Volcano, Montserrat, West Indies, 17 September, 1996, Geophys. Res. Lett., 25, 3429–3432, 1998.CrossRefGoogle Scholar
- Shimbori, T. and K. Fukui, Time variation of the eruption cloud echo height from Shinmoe-dake volcano in 2011 observed by Tanegashima and Fukuoka weather radars, Rep. Coord. Comm. Predict. Volcan. Erup., 108, 2012 (in Japanese).Google Scholar
- Suzuki, Y., A. Yasuda, N. Hokanishi, T. Kaneko, S. Nakada, and T. Fujii, Syneruptive deep magma transfer and shallow magma remobilization during the 2011 eruption of Shinmoe-dake, Japan: Constraints from melt inclusions and phase equilibria experiments, J. Volcanol. Geotherm. Res., 257, 184–204, 2013a.CrossRefGoogle Scholar
- Suzuki, Y., M. Nagai, F. Maeno, A. Yasuda, N. Hokanishi, T. Shimano, M. Ichihara, T. Kaneko, and S. Nakada, Precursory activity and evolution of the 2011 eruption of Shinmoe-dake in Kirishima volcano— insights from ash samples, Earth Planets Space, 65, this issue, 591–607, doi:10.5047/eps.2013.02.004, 2013b.CrossRefGoogle Scholar
- Yokoo, A., F. Maeno, and H. Taniguchi, Asama explosion of September 1st, 2004—On the damage to glass windows and estimation of explosion energy, Bull. Volcanol. Soc. Jpn., 50, 195–201, 2005 (in Japanese with English abstract).Google Scholar