Preparatory and precursory processes leading up to the 2014 phreatic eruption of Mount Ontake, Japan
© Kato et al. 2016
Received: 15 April 2015
Accepted: 2 July 2015
Published: 16 July 2015
We analyzed seismicity linked to the 2014 phreatic eruption of Mount Ontake, Japan, on 27 September 2014. We first relocated shallow volcano tectonic (VT) earthquakes and long-period (LP) events from August to September 2014. By applying a matched-filter technique to continuous waveforms using these relocated earthquakes, we detected numerous additional micro-earthquakes beneath the craters. The relocated VT earthquakes aligned on a near-vertical plane oriented NNW–SSE, suggesting they occurred around a conduit related to the intrusion of magmatic–hydrothermal fluids into the craters. The frequency of VT earthquakes gradually increased from 6 September 2014 and reached a peak on 11 September 2014. After the peak, seismicity levels remained elevated until the eruption. b-values gradually increased from 1.2 to 1.7 from 11 to 16 September 2014 then declined gradually and dropped to 0.8 just before the eruption. During the 10-min period immediately preceding the phreatic eruption, VT earthquakes migrated in the up-dip direction as well as laterally along the NNW–SSE feature. The migrating seismicity coincided with an accelerated increase of pre-eruptive tremor amplitude and with an anomalous tiltmeter signal that indicated summit upheaval. Therefore, the migrating seismicity suggests that the vertical conduit was filled with pressurized fluids, which rapidly propagated to the surface during the final 10 min before the eruption.
KeywordsMount Ontake 2014 phreatic eruption Precursor Earthquake Relocation Matched-filter technique b-value
It is generally difficult to predict the timing and likely size (VEI) of phreatic eruptions because their measurable precursors are weak and highly localized (e.g., Barberi et al. 1992) or absent (e.g., Maeda et al. 2015). Phreatic eruptions often occur during periods of elevated seismicity and high heat flow through the volcanic system, which can be potential precursors (e.g., Barberi et al. 1992; Aoyama and Oshima 2008; Mordret et al. 2010). However, the fundamental processes related to preparatory and precursory stages of phreatic eruptions are poorly known. Because seismicity is one of the most powerful tools for identifying volcanic processes that could precede phreatic eruptions (e.g., Chouet 1996; Roman and Cashman 2006), it is essential to investigate the spatial–temporal evolution of earthquakes leading up to the 2014 phreatic eruption of Mount Ontake.
A manually constructed earthquake catalog for the region under the summit of Mount Ontake shows that an average monthly rate of volcano tectonic (VT) and long-period (LP) earthquakes (magnitude < 1.0) during 1 year before the eruption was about 3.3/month, and 0.17/month, respectively (Japan Meteorological 2014). Then, the micro-seismicity sharply increased beginning around 2 weeks before the eruption. To constrain the fundamental processes associated with the phreatic eruption, we relocated VT and LP earthquakes beneath the summit, using double differential travel times extracted from waveform cross-correlation method. We then searched for new earthquakes by applying a matched-filter detection technique (e.g., Shelly et al. 2007; Kato et al. 2013) to continuous seismograms recorded near the Mount Ontake summit (Fig. 1 b) from 23 August to 30 September 2014, using the waveforms of well-located earthquakes as template events. We discuss the spatial–temporal evolution of the elevated seismicity, including temporal changes in b-values from the beginning of September to the time of eruption. Furthermore, we focus on the immediate precursors that began less than 10 min before the phreatic eruption, including the up-dip and lateral migrations of earthquake hypocenters, an accelerated increase in the amplitudes of pre-eruptive tremors, and tiltmeter changes observed near the summit.
Data and methods
Relocation of template earthquakes
The numbers of absolute arrival times used in the relocation are 979 for P-waves and 722 for S-waves. The numbers of differential arrival times for the manually picked P- and S-waves are 6697 and 4003, respectively. We also used differential arrival times obtained by waveform cross-correlation method. The correlation measurements were conducted using a 0.8-s window length beginning 0.4 s before each visually picked arrival time for waveform data bandpass filtered between 8 and 16 Hz. This produced a data set of differential arrival times that contained 2477 P-wave and 367 S-wave observation pairs, all of which had a normalized cross-correlation coefficient of ≥0.90. The root mean square of the travel time residual was reduced from 0.118 to 0.06 s after 21 iterations.
To assess the reliability of the relocated hypocenters, we applied a statistical resampling approach termed as bootstrap method to the events. We calculated synthetic arrival time data by adding random Gaussian noises with a standard deviation of 0.05 s (average of 0 s) to the observed arrival time data. We then relocated events to determine the shift in location based on the corrupted arrival times. The process was repeated 500 times with new Gaussian noise generated for each iteration. Based on the cumulative results, the average horizontal and depth errors are 160 and 340 m, respectively. The error ellipsoids of each hypocenter on the horizontal surface are slightly elongated along the NE–SW directions. The hypocenter accuracy of the LP events are not so well compared with that of the VT events, due to a shortage of high-frequency components at onsets of P- and S-wave arrivals. Furthermore, absolute locations were sensitive to the assumed velocity structure, but the relative distribution of earthquake was less sensitive. For an example, relocated hypocenters tend to be deeper and more clustered for an assumed velocity model with a small velocity gradient near the surface.
Detection of earthquakes by a matched-filter technique
To more precisely characterize pre- and post-eruption temporal changes in seismicity, we searched for additional events using a matched-filter technique (e.g., Shelly et al. 2007; Kato et al. 2012, 2013). Waveform data from the relocatable earthquakes were used to construct a dictionary of template events. We then matched these waveforms against continuous data recorded between 23 August 2014 and 30 September 2014.
Calculation of b-values
Using the newly constructed earthquake catalog, we calculated b-values using the maximum likelihood estimate (Utsu 1965) in overlapping 400-event windows shifted by 20 events per calculation. We simultaneously determined the temporal change in the catalog completeness magnitude using the goodness-of-fit method (Wiemer and Wyss 2000). We used the 90 % goodness-of-fit level as the completeness magnitude, which implies that 90 % of the variability in the observed frequency–magnitude distribution can be modeled with a synthetic power law.
Figure 4a shows that the relocated hypocenters are tightly clustered to the southwest of the Mount Ontake summit, with ~500-m absolute offset. The VT and LP events occurred in the same area at shallow depths close to sea level. The relocated hypocenters (98 events) roughly coincide with alignments of eruptive vents identified by remote sensing images obtained by Geospatial Information Authority of Japan (Geospatial Information Authority of Japan 2014) (white masked area in Fig. 4 a). The epicentral distribution was oriented roughly NNW–SSE. Depth section for relocated hypocenters shows that the earthquakes are aligned along a vertically dipping plane (Fig. 4b) approximately 1.5 km high (dip direction) and 0.7 km wide. Additionally, while VT earthquakes before the eruption were located at greater depths, post-eruption VT earthquakes occurred at depths shallower than the preceding seismicity by approximately 1.0 km.
Figure 6a, b shows space–time diagrams of the newly detected seismicity. From 6 September 2014, the source region of VT earthquakes slightly expanded vertically as well as laterally along the NNW–SSE feature, accompanied by a gradual increase in VT and LP earthquakes. The VT earthquake rate peaked on 11 September and remained relatively high until the eruption, with a slight decay (Fig. 6c). The number of LP events showed a sharp increase between 15 and 16 September, a delay of about 5 days from the peak of VT seismicity.
After the eruption, the depths of VT earthquakes became systematically shallower by ~1.0 km relative to pre-eruption hypocenters (Fig. 6b), even after adding numerous events with the matched-filter technique. Furthermore, we recognized a slight lateral expansion of the earthquake alignment in the NNW–SSE direction just after the eruption (Fig. 6a).
Discussion and conclusions
By relocating shallow VT and LP earthquakes and applying a matched-filter detection technique to seismic data recorded before the 2014 phreatic eruption of Mount Ontake, we reconstructed the earthquake catalog beneath the craters, which provides new insights into the preparatory and precursory stages of the eruption. The relocated VT earthquakes are aligned on a near-vertical plane oriented NNW–SSE, defining a near-vertical conduit related to the intrusion of hydrothermal fluids (steam, gases, and liquid water) into the craters (Fig. 4).
Before the minor phreatic eruption in 2007, a very-long-period (VLP) earthquake with equivalent moment magnitude 3.4, as well as a burst of LP events, was recorded by broadband seismic stations (Nakamichi et al. 2009). Waveform inversion yielded a location 0.6 km above sea level beneath the craters of Mount Ontake and suggests that the source mechanism was inflation followed by deflation of the crack associated with vaporization of water and the subsequent discharge of steam into the hydrothermal system. Interestingly, the azimuth of the crack exciting the VLP (N20W) was sub-parallel to that of the vertical conduit indicated by the relocated hypocenters in the present study (N15W). Whether reactivation of an old conduit or coincidental similarity, the similar geometries suggest the preferred orientation of pre-existing cracks along NNW–SSE beneath the summit of Mount Ontake. However, a VLP event was not recorded during the elevated VT seismicity before the 2014 eruption.
During the period of increased LP seismicity, b-values increased to nearly 1.7 (Fig. 8). Based on spatial mapping of b-values at many volcanoes (e.g., Wyss et al. 1997; McNutt 2005; Bridges and Gao 2006), high b-values have been widely observed in areas with high thermal gradients due to the interaction of hot fluids (vapors, gases, and liquid water) with adjacent magma bodies. Thus, we interpret that the high b-values reflect an increase in the density of smaller faults/cracks due to the infiltration of hot fluids (Fig. 9a).
Between 11 and 16 September 2014, the combination of an increase in LP events and an increase in b-values suggests that hot fluids infiltrated into pre-existing faults/cracks above a magma chamber. Fluid pressures within the faults/cracks further increased, resulting in a reduction of the effective normal stress on faults/cracks (e.g., Hubbert and Rubey 1959; Terakawa et al. 2013; Terakawa 2014). This reduction in fault strength facilitated the growth of small faults/cracks into larger ones, producing higher-magnitude earthquakes (Fig. 6d), which can also explain the gradual decrease in b-values after 16 September 2014 (Fig. 8).
Based on these observations, we postulate that hot fluids were pervasive to pre-existing faults/cracks and pressurized by heat supplied from the underlying magma chamber from the middle of September 2014 (Fig. 9a). The shallow portion beneath the summit perhaps withstood stress concentration derived from the increase in internal pressure of the underlying hydrothermal system.
During the 10 min prior to the phreatic eruption, the up-dip and lateral migrations of VT earthquakes coincided with an increase in pre-eruptive tremor amplitude and an anomalous tiltmeter signal near the craters. Because the tiltmeter signal was consistent with summit upheaval (Figs. 1 and 9b), the migrating seismicity indicates the rapid propagation of pressurized fluids (vapor, gases, and liquid water) through a vertical conduit. The lateral propagating speed of the conduit tip ranged from 0.35 to 0.7 m/s (Fig. 7a). Interestingly, these speeds are roughly comparable to those of migrating VT earthquakes triggered by dike intrusion (0.06–0.3 m/s) (e.g., Hayashi and Morita 2003; Shelly and Hill 2011). Finally, failure of the sealed surface of the vertical conduit led to the 2014 phreatic eruption. The sealed surface may be produced by the precipitation of altered clay minerals or hydrothermal secondary deposits near the surface (e.g., Bertrand et al. 2012; Maeda et al. 2015).
To our knowledge, the present study was the first to describe the rapid propagation of a vertical conduit during the accelerating stage prior to a phreatic eruption. It is important to capture short-term precursors as early as possible and to raise the alarm for climbers to evacuate from potential eruption sites. Further understanding of the excitation mechanism of phreatic eruptions will require continuous monitoring of seismicity and crustal movement in the vicinity of the summit because the geophysical signals related to the pre-eruption propagation of a vertical crack are too localized to be observed in the far field.
We thank JMA, NIED, Nagano Prefecture, and Gifu Prefecture for allowing us to use waveform data collected by their permanent stations. JMA provided the tiltmeter data recorded near the summit (V.ONTN). We are grateful to Y. Shibayama for the dedicated help with the data processing. We thank D. Roman and one anonymous reviewer for useful comments and suggestions. Figures were created using GMT (Wessel and Smith 1995) and spectrogram analysis tool (Miyakawa and Sakai 2008). This study was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, under its Earthquake and Volcano Hazards Observation and Research Program and under KAKEN (26900002). This study was also supported by the Earthquake Research Institute cooperative research program.
- Aoki Y, Takeo M, Aoyama H, Fujimatsu J, Matsumoto S, Miyamachi H, Nakamichi H, Ohkura T, Ohminato T, Oikawa J, Tanada R, Tsutsui T, Yamamoto K, Yamamoto M, Yamasato H, Yamawaki T (2009) P-wave velocity structure beneath Asama Volcano, Japan, inferred from active source seismic experiment. J Volcanol Geotherm Res 187:272–7Google Scholar
- Aoyama H, Oshima H (2008) Tilt change recorded by broadband seismometer prior to small phreatic explosion of Meakan-dake volcano, Hokkaido, Japan. Geophys Res Lett 35:L06307. doi:10.1029/2007GL032988 Google Scholar
- Barberi F, Bertagnini A, Landi P, Principe C (1992) A review on phreatic eruptions and their precursors. J Volcanol Geotherm Res 52(4):231–46. doi:10.1016/0377-0273(92)90046-G View ArticleGoogle Scholar
- Bertrand EA, Caldwell TG, Hill GJ, Wallin EL, Bennie SL, Cozens N, Onacha SA, Ryan GA, Walter C, Zaino A, Wameyo P (2012) Magnetotelluric imaging of upper-crustal convection plumes beneath the Taupo Volcanic Zone, New Zealand. Geophys Res Lett 39:L02304. doi:10.1029/2011GL050177
- Bridges DL, Gao S (2006) Spatial variation of seismic b-values beneath Makushin Volcano, Unalaska Island, Alaska. Earth Planet Sci Lett, 245, 408–415 doi:10.1016/j.epsl.2006.03.010.
- Chouet B (1996) Long-period volcano seismicity: its source and use in eruption forecasting. Nature 380:309–16View ArticleGoogle Scholar
- Cyranoski D (2014) Why Japan missed volcano’s warning signs. Nature. doi:10.1038/nature.2014.16022 Google Scholar
- Earthquake Research Institute in the University of Tokyo (2014), http://www.data.jma.go.jp/svd/vois/data/tokyo/STOCK/kaisetsu/CCPVE/shiryo/130/130_no01.pdf. Accessed 1 April 2015.
- Geospatial Information Authority of Japan (2014), http://www.gsi.go.jp/BOUSAI/h26-ontake-index.html. Accessed 1 April 2015.
- Hayashi Y, Morita Y (2003) An image of a magma intrusion process inferred from precise hypocentral migrations of the earthquake swarm east of the Izu Peninsula. Geophys J Int 153:159–74View ArticleGoogle Scholar
- Hirose F, Nakajima J, Hasegawa A (2008) Three-dimensional seismic velocity structure and configuration of the Philippine Sea slab in southwestern Japan estimated by double-difference tomography. J Geophys Res 113:B09315. doi:10.1029/2007JB005274 Google Scholar
- Hubbert MK, Rubey WW (1959) Role of fluid pressure in mechanics of overthrust faulting 1. Mechanics of fluid-filled porous solids and its application to overthrust faulting. Geol Soc Am Bull 70(2):115–66View ArticleGoogle Scholar
- Ikami A, Yoshii T, Kubota S, Sasaki Y, Hasemi A, Moriya T, Miyamachi H, Matsu’ura R, Wada K (1986) A seismic refraction profile in and around Nagano Prefecture, Central Japan. J Phys Earth 34:457–74View ArticleGoogle Scholar
- Japan Meteorological Agency (2014), Rep. Coordin. Committee on Prediction of Volcanic Eruption, http://www.data.jma.go.jp/svd/vois/data/tokyo/STOCK/kaisetsu/CCPVE/shiryo/130/130_no01.pdf. Accessed 1 April 2015.
- Kato A, Iidaka T, Kurashimo E, Nakagawa S, Hirata N, Iwasaki T (2007) Delineation of probable asperities on the Atotsugawa fault, central Japan, using a dense temporary seismic network. Geophys Res Lett 34:L09318. doi:10.1029/2007GL029604 View ArticleGoogle Scholar
- Kato A, Obara K, Igarashi T, Tsuruoka H, Nakagawa S, Hirata N (2012) Propagation of slow slip leading up to the 2011 Mw 9.0 Tohoku-Oki earthquake. Science 335:705–8. doi:10.1126/science.1215141 View ArticleGoogle Scholar
- Kato A, Fukuda J, Obara K (2013) Response of seismicity to static and dynamic stress changes induced by the 2011 M9.0 Tohoku-Oki earthquake. Geophys Res Lett 40, doi:10.1002/grl.50699.
- Kumagai H, Chouet BA (2000) Acoustic properties of a crack containing magmatic or hydrothermal fluids. J Geophys Res 105(B11):25,493–25,512View ArticleGoogle Scholar
- Lahr JC, Chouet BA, Stephens CD, Power JA, Page RA (1994) Earthquake classification, location and error analysis in a volcanic environment: implications for the magmatic system of the 1989–1990 eruptions at Redoubt Volcano, Alaska. J Volcanol Geotherm Res 62:137–51View ArticleGoogle Scholar
- Maeda Y, Kumagai H, Lacson RJ, Figueroa MS, Yamashina T, Ohkura T, Baloloy AV (2015) A phreatic explosion model inferred from a very long period seismic event at Mayon Volcano, Philippines. J Geophys Res Solid Earth 120:226–42. doi:10.1002/2014JB011440 View ArticleGoogle Scholar
- McNutt SR (2005) Volcanic seismology. Annu Rev Earth Planet Sci 32:461–91View ArticleGoogle Scholar
- Miyakawa K, Sakai S (2008) Development of a spectrogram analysis tool for seismic waveform data and its application to MeSO-net for noise survey, Technical Research Report, Earthquake Research Institute, the University of Tokyo, 14, 13–22.Google Scholar
- Mordret A, Jolly AD, Duputel Z, Fournier N (2010) Monitoring of phreatic eruptions using Interferometry on Retrieved Cross-Correlation Function from Ambient Seismic Noise: results from Mt. Ruapehu, New Zealand. J Volcanol Geotherm Res 191:46–59View ArticleGoogle Scholar
- Nakajima J, Hasegawa A (2006) Anomalous low-velocity zone and linear alignment of seismicity along it in the subducted Pacific slab beneath Kanto, Japan: reactivation of subducted fracture zone? Geophys Res Lett 33:L16309. doi:10.1029/2006GL026773 View ArticleGoogle Scholar
- Nakamichi H, Kumagai H, Nakano M, Okubo M, Kimata F, Ito Y, Obara K (2009) Source mechanism of very-long-period event at Mt. Ontake, central Japan: response of a hydrothermal system to magma intrusion beneath the summit. J Volcanol Geotherm Res 187:167–77View ArticleGoogle Scholar
- Roman DC, Cashman KV (2006) The origin of volcano-tectonic earthquake swarms. Geology 34:457–60View ArticleGoogle Scholar
- Shelly DR, Hill DP (2011) Migrating swarms of brittle‐failure earthquakes in the lower crust beneath Mammoth Mountain, California. Geophys Res Lett 38:L20307. doi:10.1029/2011GL049336 Google Scholar
- Shelly DR, Beroza GC, Ide S (2007) Non-volcanic tremor and low-frequency earthquake swarms. Nature 446:305–7. doi:10.1038/nature05666 View ArticleGoogle Scholar
- Shi Y, Bolt BA (1982) The standard error of the magnitude-frequency b-value. Bull Seism Soc Am 72:1677–87Google Scholar
- Smithsonian Institution (2015) Bulletin of the Global Volcanism Network, http://www.volcano.si.edu/search_eruption.cfm. Accessed 1 April 2015.
- Tanaka S, Hamaguchi H, Nishimura T, Yamawaki T, Ueki S, Nakamichi H, Tsutsui T, Miyamachi H, Matsuwo N, Oikawa J, Ohminato T, Miyaoka K, Onizawa S, Mori T, Aizawa K (2002) Three-dimensional P-wave velocity structure of Iwate volcano, Japan from active seismic survey. Geophys Res Lett 29, 10.1029/2002GL014983.Google Scholar
- Terakawa T (2014) Evolution of pore fluid pressures in a stimulated geothermal reservoir inferred from earthquake focal mechanisms. Geophys Res Lett 41:7468–76. doi:10.1002/2014GL061908 View ArticleGoogle Scholar
- Terakawa T, Yamanaka Y, Nakamichi H, Watanabe T, Yamazaki F, Horikawa S, Okuda T (2013) Effects of pore fluid pressure and tectonic stress on diverse seismic activities around the Mt. Ontake volcano, central Japan. Tectonophysics 608:138–48View ArticleGoogle Scholar
- Utsu T (1965) A method for determining the value of b in a formula log n = a - bM showing the magnitude frequency for earthquakes. Geophys Bull Hokkaido Univ 13:99–103Google Scholar
- Waite GP, Chouet BA, Dawson PB (2008) Eruption dynamics at Mount St. Helens imaged from broadband seismic waveforms: interaction of the shallow magmatic and hydrothermal systems. J Geophys Res 113, B02305, doi:10.1029/2007JB005259.
- Wessel P, Smith WHF (1995) New version of the generic mapping tools released. Eos Trans AGU 76:329View ArticleGoogle Scholar
- Wiemer S, Wyss M (2000) Minimum magnitude of completeness in earthquake catalogs: examples from Alaska, the western United States, and Japan. Bull Seismol Soc Am 90:859–69View ArticleGoogle Scholar
- Wyss M, Shimazaki K, Wiemer S (1997) Mapping active magma chambers by b-value beneath the off-Ito volcano, Japan. J Geophys Res 102:20413–22View ArticleGoogle Scholar
- Zhang H, Thurber CH (2003) Double-difference tomography: the method and its application to the Hayward fault, California. Bull Seismol Soc Am 93:1875–89. doi:10.1785/0120020190 View ArticleGoogle Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.