VI. National Project for Prediction of Volcanic Eruptions

2. Joint Volcanological Experiment on Volcanic Structure and Magma Supply System

Since 1994, joint experiments have been conducted in Japan to reveal the structure and the magma supply system by the scientist group of national universities under the National Research Project for Prediction of Volcanic Eruptions. The experiments were conducted in Kirishima Volcanoes in 1994, Unzen Volcano in 1995, Kirishima Volcanoes in 1996, Bandai Volcano in 1997 and Aso Volcano in 1998 (Fig. 25). The experiments were carried out by seismological, electromagnetic and other geophysical methods. These experiments succeeded in detecting some anomalous regions related to magma activity. A new task in the following 5 years is to develop a new method to detect a temporal change of the structure. From this aspect, the next experiment is now planned in Izu-Oshima Volcano, which has eruptions frequently. The results of the previous experiments are briefly presented as follows.

Experiments in Kirishima Volcanoes in 1994 and 1996

Experiments were conducted in Kirishima Volcanoes in 1994 and in 1996. The first seismic explosion experiments were carried out on December 1, 1994. Observations were made along a 30-km major line lying in the NNW-SSW direction and other sub-lines which across the major line in and around the Kirishima Volcanoes. Along these lines, 6 shots with a charge size of 200-250 kg, and 163 temporary observations were arranged. The newly developed data logger was used for those temporal observations, and the position of each site was determined by GPS. All 6 shots were successfully fired and significant data were obtained by most of the loggers (Kagiyama et al., 1995). Velocity structure of Kirishima was revealed by refractive analysis (Tsutsui et al., 1996) and by 3-D inversion (Nishi, 1997). Mikada (1996) applied reflective analysis to this data. Magnetotelluric survey was conducted in September 1994 by using more than 10 sets of portable MT instrument. Simultaneous magnetic and electric field variations were measured in the ULF, ELF and VLF band in and around Kirishima (Kagiyama et al., 1996). The second experiment was carried out in 1996 with some new additional view. Observation points for seismic explosion experiment were arranged just linearly, but distributed at intervals of 50 m crossing Iwo-Yama, one of the anomalous areas in Kirishima detected by the previous experiment. This experiment clarified that high-density observation arrangement is useful for relection analysis. MT survey was carried out along the E-W lines, which across the Kyushu Island, to clarify the regional resistivity structure around Kirishima Volcanoes (Munekane et al., 1997, 1998).

The experiment found some anomalous regions suggesting magmatic activity, and revealed that the Kirishima volcanic area is divided into two sub-areas; the northwestern volcanic area and the southeastern volcanic area. According to the MT survey, volcanoes in the northwestern volcanic area such as Iwo-Yama and Shinmoe-Dake are characterized by a simple four-layer structure; relatively resistive overburden, very conducting second layer at the depth of a few hundred meters, resistive basement, and deep conductor (Fig. 26). The low-resistivity second layer (Shallow Low Resistivity Region) is interpreted as a water-saturated porous layer, which is widely distributed throughout Kirishima. This layer plays an important role in controlling types of eruption and in generating precursory phenomena of volcanic eruptions through interaction of the water with ascending magma. The deep conductor (Deep Low Resistivity Region) appears about 10 km below the surface in average, but shallower beneath the presently active volcanoes, such as Iwo-Yama or Shinmoe-Dake, up to a few km in depth. On the other hand, volcanoes in the southeastern volcanic area such as Ohachi and Mi-Ike are characterized by a rather resistive second layer and lack of deep conductor. Seismic explosion experiment found seismic reflectors, which is probably related to the Deep Low Resistivity Region, at 10 km beneath the northwestern area and at several km beneath Iwo-Yama, respectively (Fig. 27). Precise seismological and other geophysical observations clarified a remarkable migration of anomalous phenomena described bellow. An outbreak of earthquake swarm occurred at the top of the DLRR beneath Shinmoe-Dake. This activity expanded upward, and followed by a dilatational crustal deformation in the resistive third layer, volcanic tremor and thermal demagnetization within the SLRR, and finally steam and ash emission from the crater. These results suggest that the DLRR or the seismic reflective zone be related with magma.

Seismic refraction experiment and gravimetric survey also suggest that a structural discontinuity exist between northwestern volcanoes and southeastern volcanoes. High-velocity (high-density) basement (Shimanto Group) is depressed and covered with the low-velocity (low-density) surface layer in the northwestern area, while the basement is mostly bared in the southeastern area (Tsutsui et al., 1996; Nishi, 1997; Kobayashi et al., 1995). The difference of resistivity structure may be related to the difference in the magma-supplying system. The results of the MT survey and of other seismological and petrological investigations indicate that beneath the northwestern Kirishima volcanoes, magma is stored about 10 km in depth, while on the other hand beneath southeastern volcanoes, magma is supplied from a deeper part.

Experiment in Unzen Volcano in 1995

Unzen Volcano started eruption in November 1990, and it continued until early 1995. A lava dome was formed in this eruption, and the total volume of erupted lava reached approximately 200 million m**3.

In order to reveal the subsurface structure and magma feeding system of Unzen Volcano, a seismic experiment was conducted using artificial explosion sources on November 30, 1995. Fig. 28 shows the location map of the experiment. About 290 geophones were deployed along two survey lines (N-S and E-W) and at 4 small-aperture arrays. The E-W and N-S lines are about 26.5 and 12.5 km long in the WNW-ESE and NNE-SSW directions, respectively. These lines cross on the western flank of the volcano. Six shots of 200-250 kg dynamite were fired on and at the end of the lines, and the seismic signals were recorded on a compact data-logger with precise GPS clock at each geophone site. Outlines of the experiment and the travel-time data were reported by Matsushima et al. (1997).

A velocity structure model was obtained from the travel-time analysis using a ray tracing method (Shimizu et al., 1997). The subsurface structure of Unzen Volcano consists of materials with compressional wavespeeds 1-1.9, 2.1-3.5, 4-4.5 and 6-6.1 km/s, layered from the top (Fig. 29). The layers of 4-4.5 km/s and 6-6.1 km/s rise toward the western flank of the volcano. Beneath this swelling, pressure sources (magma chambers) derived from geodetic measurements are located with depth of 4-5 and 7-10 km. Although these magma chambers could not be detected by the refraction analysis, Tsutsui et al. (1997) successfully detected some reflectors beneath the western flank of the volcano using reflected waves recorded at a seismic array in this experiment. The survey lines are crossing the active faults that form the Unzen graben. The obtained velocity model, however, does not show the apparent subsidence of layers corresponding to the graben structure. The seismic velocity of the volcanic rocks filled in the graben may be comparable or rather higher than that of the Tertiary sediment, which forms the basement of Unzen Volcano.

Since the artificial electromagnetic noises are very high around Unzen compared with Kirishima Volcanoes, TDEM (Time Domain Electro Magnetic) method was newly developed for the experiment in Unzen Volcano in 1995. High quality transient magnetic field data were obtained from measurements made at 26 sites. This method successfully found a low resistive region elongated in the east-west direction several km in depth beneath Unzen Volcano (Kanda et al., 1997).

Experiment in Bandai Volcano in 1997

Bandai volcano had experienced the great eruption in July 15, 1888 accompanied with a large phreatic explosion, which created a horseshoe-shaped crater. It is great important to investigate volcanic interior in order to clarify mechanism of such an eruption.

At Bandai volcano, a seismic exploration using active sources was carried out in October 1997. Ten national universities and JMA participated in this experiment. Geographical distribution of seismic stations and shot points was arranged to reveal the 3-D structure of the volcano. Such a 3-D exploration is the first attempt in Japan. We deployed 292 seismic stations on and around the volcano within 15 km from the summit. The stations were distributed with 0.5 km-intervals along mountain paths within 5 km from the summit. In the outside the above area, the interval was 1 km. Active sources with 100-250 kg dynamite were fired at the 6 points in the margin of the target region and at the 2 points adjacent to the summit. We successfully obtained 2157 first motion arrival times.

We inverted the first arrival times into 3-D velocity structure using pseudo-bending method and the damped least square. Reliability of the obtained model was examined with checkerboard test. The test confirmed that the structure of the region within 5 km-distance from the summit and depth shallower than 2 km from the ground surface are resolved with much confidence. The 3-D image revealed that a prominent high velocity body (with Vp=4.8-5.2 km/s) protruding vertically with horizontal extent less than 2 km, is located just beneath the crater of the volcano.

(Tsuneomi Kagiyama, Hiroshi Shimizu and Satoru Tanaka)

References

*Kagiyama, T. et al. (1995) 1994 explosion experiment in Kirishima Volcanoes. Bull. Earthq. Res. Inst., 70, 33-60.

*Kagiyama, T. et al. (1996) Resistivity structure of the central and southeastern part of Kirishima Volcanic region. Bull. Volcanol. Soc. Japan, 41, 215-225.

*Kagiyama, T., Utada, H., Mikada, H., Tsutsui, T. and Masutani, F. (1997) Structure of the Kirishima Volcanic Region and its magma supply system. Bull. Volcanol Soc. Japan, 42, S157-S165.

Kanda, W., Utada, H., Kagiyama, T. and Tanaka, Y. (1997) Resistivity model around Unzen Volcano inferred from TDEM experiments. Proc. Unzen International Workshop, pp. 50-55.

Kobayashi, S., Shichi, R., Nishinaka, H., Watanabe, H. and Onizawa, S. (1995) Dense gravity survey in the Kirishima Volcanoes and its surrounding calderas, Southern Kyushu, Japan. Bull. Earthq. Res. Inst., 70, 103-136.

*Matsushima, T. et al. (1997) 1995 Explosion experiment in Unzen Volcano. Bull. Earthq. Res. Inst., 72, 167-183.

*Mikada, H. (1996) A seismic reflection analysis on refraction data from the 1994 Kirishima Explosion Experiment. Bull. Volcanol. Soc. Japan, 41, 159-170.

*Munekane, H. et al. (1997) Resistivity survey across the southern part of the Kyushu Island. Bull. Earthq. Res. Inst., 72, 19-65.

Munekane, H., Kagiyama, T. and Utada, H. (1998) The resistivity structure in the southern Kyushu area (abstract). EOS, 79, W106.

**Nishi, K. (1997) Three dimensional subsurface P-wave velocity structure of Kirishima Volcano derived from shot data inversion. Bull. Volcanol. Soc. Japan, 42, 165-170.

**Shimizu, H. and Explosion Seismic Research Group of Unzen Volcano (1997) Subsurface structure of Unzen Volcano derived from the 1995 explosion experiment. Proc. Magma Exploration, pp. 9-15.

*Tsutsui, T. et al. (1996) Seismic velocity structure beneath Kirishima Volcanoes with differential analysis of explosion experiment. Bull. Volcanol. Soc. Japan, 41, 227-241.

**Tsutsui, T., Nakaboh, M., Mori, T., Matsumoto, Y. and Yoshikawa, S. (1997) A detection of reflectors or scatterers beneath Unzen Volcano, western Kyushu. Proc. Magma Exploration, pp. 17-26.

(* In Japanese with English abstract. ** In Japanese)