After 12 years' quiescence, Oshima Volcano, Izu, Japan (Izu-Oshima Volcano) entered an active period by eruptions from the central cone on November 15, 1986. There were three stages in the volcanic activity. The first stage was a period when the volcano repeated Strombolian type eruptions from its central cone. The second stage was characterized by fissure eruptions extending outside the caldera rim, and the third stage is the recovery phase. It is here attempted to review the changes in volcanic activity from the geomagnetic and geoelectric points of view.

View full abstract

We have repeatedly measured the electrical resistivity of Oshima Volcano using bipole-bipole electrode arrangements across the central cone, Mihara-yama. Apparent resistivities obtained by three different pairs of electrodes varied according to the vicissitudes of the volcanic activity. Changes associated with the eruption on November 15, 1986, were particularly remarkable. The increase of the apparent resistivity obtained by the electrode pairs the nearest to the central crater vent became noticeable about one and half years before the eruption, and one day after the breakout of the eruption the apparent resistivity reached a value 17% higher than its original resistivity value. The increase of the apparent resistivity observed with these electrode pairs has been interpreted as a topographical effect caused by the ascent of magma through the volcanic vent. The apparent resistivity obtained by the farthest pairs of electrodes began to change about 10 months before the eruption. The decrease for the last 3 months before the eruption was extremely rapid and large, and the apparent resistivity became as small as about 50% of its original value. This is considered to reflect the change in the true resistivity at a few hundred meters' depth beneath the central crater.

View full abstract

The VLF and ELF magnetotelluric (MT) methods have been applied to study the electrical resistivity distribution in the Izu-Oshima volcano. We utilized a 17.4kHz artificial electromagnetic signal for the VLF method, and three fundamental Schumann resonance frequencies for the ELF method. Measurements were made in 1984 and 1985 at 57 sites including 30 on the central caldera floor, and one-dimensional resistivity models were determined from the results. The modeling results clearly indicate the dominance of a water reservoir beneath the caldera at the depth of a few hundred meters above the sea level.
On the crater floor of the central cone Mihara-yama, measurements were carried out at four sites. The result shows a very shallow conductor with a resistivity of less than 10Ω·m. We also found that the edifice of Mihara-yama tends to be more conducting in the southern part than in the northern part, extending in the east-west direction. The important thing is that, within the one-dimensional approximation, this anomalously conducting layer exists only 200-300m below the crater floor. We concluded this conducting layer directly reflects the present thermal activity in the central cone of Mihara-yama. Furthermore, the depth of the conducting layer tends to be shallow at the northern and southern parts of the caldera rim. This shows a peculiar feature of the groundwater distribution beneath the caldera rim.

View full abstract

Airborne VLF-EM and ground VLF-MT surveys were carried out to study the resistivity structure of Izu-Oshima volcano after its fissure eruptions in November, 1986. In both the airborne and ground surveys, we employed electromagnetic induction techniques with VLF radio waves (Very Low Frequency, 17.4kHz in our case) as the primary source field.
Airborne VLF-EM surveys measure the variations in the vertical component of the magnetic field which are coherent to those of the horizontal component, and map a quantity that corresponds to the magnetic transfer function derived from the ratio of the vertical to the horizontal component. Ground VLF-MT surveys, on the other hand, measure the horizontal electric field simultaneously with the horizontal magnetic field and compute the apparent resistivity.
Airborne VLF-EM surveys of Izu-Oshima volcano revealed that there are clear trends of alignment in the resistivity anomalies in many cases. A remarkable anomaly belt has been found to run along the fissures through which fountain eruptions occurred in November. Many of the anomaly belts are related to geological structures due to activity in historic times. However, there are some anomalies which do not clearly correlate with historical volcanism.
The results of the ground VLF-MT measurements along three survey lines are consistent with the airborne survey, confirming the existence of low resistivity anomalies.

View full abstract

We carried out CSAMT (controlled-source audiofrequency magnetotelluric) measurements across the 1986 C craters of Izu-Oshima Island one year after the eruptions. After 2D analyses of the two profiles, we obtained basically two-layered structures, except for shallow resistivity variations. We found a deep conductive layer below the resistive lava which corresponds to thermal water, compared with nearby wells. We also found shallow (<200m) conductive bodies whose locations correspond to an old vent (8th century) and one of the 1986 craters. We think these are fracture zones containing meteoric water. We did not get affirmative evidence for dyke-like structures, partly because we used TM mode excitation, and partly because the dyke hadn't transferred enough heat to the surrounding rocks to decrease resistivity.
Away from the profiles, we found a three layered structure. We think the three layers correspond to (top) resistive lava, (middle) a fresh water lens, and (bottom) sea water. We think the two profiles lacked the fresh water lens due to mixing of fresh and saline water by thermal convection.

View full abstract

At the time of the major volcanic eruptions on November 15 and 21, 1986 at Mt. Mihara, Izu-Ohshima Island, Tokyo territory, Japan, a detector from a multipoint direction-finding network to detect electromagnetic emissions was operating. Since October 20, 1986 impulsive noises at 82kHz were recorded, and these anomalous burst-like noise observations increased after November 3. At 10-16 JST (Japanese Standard Time, UT+9 hours) on November 14, several strong noise bursts were observed, and the first major eruption occurred at the main summit crater of this volcano at 16 JST on November 15. At 10-12 JST on November 21, noise bursts were observed and four hours later, 19 new craters erupted suddenly on the side slope of the mountain, both in and outside the caldera, simultaneously with the intrusion of dykes in the mountain body. These records are the first observations of electromagnetic emissions as precursors to volcanic eruptions ever recorded. The authors present this paper in the form of a short note for data presentation, because analysis of these data and comparison with other observations is still continuing.

View full abstract

Great eruptions occurred suddenly at the Izu-Oshima volcano Oshima Island, Japan from the evening of November 15 to November 22, 1986. In this period, atmospherics of 160kHz were observed at Koganei, Tokyo which is located 101km north of the volcano, and atmospherics of 80kHz were observed at Inubo, Chiba prefecture, located 194km northeast of the volcano.
For propagation distances above 100km, hop-mode waves, via ionospheric reflection, dominate over ground-mode waves. The atmospherics from the Izu-Oshima volcano were detected only during the night at Koganei and Inubo, since LF radio waves are strongly absorbed in the ionospheric D region produced by solar X-rays and solar EUV radiation during the daytime.
Nighttime intensity enhancements of 160kHz atmospherics with 10dB or more above the normal night levels of November 14 and 15 were observed at Koganei from November 16 to 20, 1986. The nighttime enhancements of 160kHz atmospherics from November 16 to 19 show characteristic time variations with periods from 10 minutes to 15 minutes, corresponding to the time intervals of repeated eruptions at the summit of the Izu-Oshima volcano. Nighttime intensity enhancements of 80kHz atmospherics of 5dB or more above the normal night level were also observed at Inubo from November 16 to 20, 1986.
According to weather maps issued by the Japanese Meteorological Agency, no meteorological phenomena occurred to cause the observed nighttime intensity enhancements of LF atmospherics, such as thunderstorms or rain clouds, in those districts during that time. Also, no local artificial noises occurred corresponding to the observed enhancements of LF atmospherics in the vicinity of the receiving sites in that period, and local noise was about 10dB lower than the enhanced level of LF atmospherics during the nighttime.
Attenuation of electromagnetic waves in magma, wet soil and dry crust are numerically examined. It seems improbable that LF electromagnetic waves propagate through magma and moist soil from a volcano's interior to the ground because of strong attenuation, even if the LF electromagnetic waves were generated inside the volcano.
Lightning discharges generated by eruption smoke have been reported for decades and in fact, some pictures of lightning discharges over the volcano were broadcast by television during the Izu-Oshima volcano eruptions. Therefore, the nighttime intensity enhancements of LF atmospherics observed at Koganei and Inubo from November 16 to 20, 1986 seem to have been generated by lightning discharges in eruption smoke and ascending warm air currents heated locally by hot lava over the crater lake.

View full abstract

We have examined secular changes in the geomagnetic total intensity at four observatories distributed almost along the meridian 140°E from 1968 to 1986, and found that the total intensity at a site (NOM) in Oshima Island was subjected to an anomalous decrease until 1978 and recovered to the normal trend around 1980. The rate of the anomalous decrease has been estimated to be -1.3nT/year. Additional observations started in 1976 and 1978 at the south and the north of the volcano show similar secular changes. This seems to suggest that the anomalous behavior observed at NOM is not a local phenomenon in Oshima Island but a regional one occurring simultaneously over the entire island, except for a site very close to the central cone of the volcano.
The time when the anomalous secular change returned to the normal variation coincides with the period when seismicity and crustal deformation such as shortening of the distance between Oshima Island and the Izu Peninsula changed its behavior. This implies that there was a change in the stress distribution over a large area between 1978 and 1980, which is supposed to have caused the anomalous secular change in Oshima Island. This regional stress change seems to have eventually led the volcano to erupt in 1986.

View full abstract

Before the eruption of Oshima Volcano on November 15, 1986, 5 proton magnetometers were operating on the volcano. One of them, installed about 500m south from the center of the central crater, recorded anomalous changes in the total intensity preceding the eruption. The anomalous variation began around 1981. The total intensity decreased gradually at a rate of -5.3nT/year (-0.44nT/month), amounting to a 27nT decrease in 5 years. The decrease of the total intensity accelerated about six months before the eruption. The rate of decrease was estimated to be about -2.2nT/month. At the time of the eruption, a very fast and significant variation was observed. The total intensity decreased 4nT in one minute. Following the eruption from the central cone, fissure eruptions took place on November 21. A decrease in the total intensity of about 7nT was also observed at this time at a site 2km southeast of the central cone.
The observed magnetic changes preceding the eruption are supposed to have been caused by the volcanic body being heated during the intrusion of magma, at the initial stage through the main vent, and at the later stage through a new path toward the newly formed crater, called A crater. The variation at the time of eruption is considered due to the disorientation of magnetic materials by the catastrophic outflow of gases and liquids through pores and cracks inside the volcano.

View full abstract

Remarkable changes in the total geomagnetic intensity were observed associated with the 1986 eruption of the Izu-Oshima Volcano. They are summarized as follows: (1) Long-term precursor: decrease in the total intensity at one station on the southern slope of the central cone, Miharayama, continued since 1981, with a constant rate of -5nT/yr. (2) Middle-term precursor: the decrease there accelerated beginning in April, 1986. (3) Short-term precursor: in October, 1986, an increase in the total intensity was observed at two stations on the north side of Miharayama, while readings at two southern stations decreased. (4) At the first stage of the Phase I eruption, step-like changes were observed at the two southern stations. (5) A decrease in the total intensity was observed at two other stations outside the caldera at the time of the fissure eruption on November 21, and a few survey points and one station showed significant changes. (6) For a few months after the eruption, no remarkable magnetic variations were found outside the caldera.
(1), (2) and (3) are interpreted as thermal demagnetization of vent materials due to high-temperature volatiles dispatched from ascending magma. (4) is most likely due to rotation of vent materials. (5) is attributed to the piezomagnetic effect, together with the thermal demagnetization, due to intrusive dykes. In the 1950 eruption a demagnetized area with a 2km radius was formed at a depth of 5km (RIKITAKE, 1951a). In contrast, during the Phase II activity on November 21 and 22, 1986, an intrusive event is suggested to have occurred beneath the southeastern region of the volcano, as inferred from the observed total field changes.

View full abstract

After the 1986 eruption of Izu-Oshima Volcano, proton magnetometers were installed at various sites in the island. Most of the sites started measurements by March, 1987. During the period from March to November 15, the day before the 1987 gas explosion of the volcano, large magnetic variations in total intensity were observed at the sites located on the caldera floor. The variations can be divided into three stages, March-June, July-August, and September-November. The first stage of the variations can be attributed to after-effects of the 1986 eruption, and the last stage could be a precursory change to the 1987 gas explosion. The location of the source region for each stage of the variation was estimated based on the spatial distribution of the rate of the variation. The center of the source region for the first stage was estimated to be located about 400m below the caldera floor under the pit crater. The source for the last stage is under A crater at a depth of 200m below the caldera floor. The direction of the magnetic moments responsible for the variations is anti-parallel to the geomagnetic field. Hence, the thermal demagnetization of the remanence of the rocks is a possible cause of the variation.

View full abstract

An aeromagnetic survey over Izu-Oshima Volcano had been conducted in 1978 by the Geological Survey of Japan. A similar survey was carried out by Nakanihon Air Service Co. Ltd. on November 29-30, 1986 soon after the eruption. Comparison of those results revealed a very good repeatability of the magnetic anomaly pattern. It was suggested that the difference between the 1978 and 1986 anomaly maps was caused by the insufficient line spacing of each survey, and that there was no largescale variation in the magnetic anomaly. Then, both the survey data were processed together to produce a more detailed anomaly map, and a magnetic structural model was proposed from the analysis of the anomalies.

View full abstract