by M.Kikuchi & Y.Yamanaka
(EIC: Earthquake Information Center)

Prior to the summit eruption around 4 a.m. on July 14, at 2:12 long-period (duration of about 50 sec.) earthquake with a possible hypocenter around the Miyake-jima was observed at broadband seismograph station. Similar phenomena with various amplitudes had recurred after that envet. Those earthquakes had a different feature from the one accompanying the eruption on the evening of July 8, about which we mentioned in EIC seismological note No. 82. We believe those earthquakes manifest subsurface volcanic activity that failed to bring about eruption.

Broadband waveform data at NIED's FREESIA stations: Tateyama(TYM), Nakaizu(JIZ), and Tsuru(SGN) were collected. The data were band pass filtered (period 10 - 100 sec.) and transformed into displacement records. We analyzed the waveform data assuming two kinds of source models. One (Model 1) was moment tensor source with all 6 components including volumetric component. The other (Model 2) was vertical single force source.

[Model 1]Results are shown in Figure 1. Mechanism solution indicates "Push" in all ray directions (explosive source).
Source parameters are as follows.

Origin Time	02:12:13 JT
Moment tensor: Mij i,j = 1 (North) , 2 (East) , 3 (Down)	(2.5 -0.7 -0.4 -0.7 2.9 1.6 -0.4 1.6 3.7) [x 10**17Nm]
Isotropic component	3.0 x 10**17Nm (Mw = 5.6 equiv.)
Volume expansion DV	6 x 10**6 [m**3] (bulk modulus k = 50GPa)
Source duration	40 s
Focal depth	about 2 km

[Model 2] Results are shown in Figure 2. Vertical upward force was applied for 30 sec. The agreement between the observed and computed waveforms are rather poor. Source parameters are as follows.

Origin Time	02:12:19 JT
Max. Impulse	Ft = 2.9 x 10**12 Ns
Mass transport	M = Ft/v ( presumed average speed v =50m/s) = about 6 x 10**10 [kg]
Volume transport	V = M/r= about 3 x 10**7 [m**3]

Source mechanism obtained can be well interpreted in terms of primarily volumetric expansion with other minor moment components. What is the cause of expansion? In addition, abrupt elevation (steps in tilt) of the volcano is observed to be followed by a gradual relaxation process. What is the cause of the change indicated with tiltsteps? We believe the following idea can be the answer. Underground water running into a high temperature area is heated into steam, and suddenly pushes all around; Gas goes out through fissures, and steam changes back into water.(Figure 3) These ideas may explain all the above observation. Alternatively, magma's drain-back is the other candidate of the long-period source. It seems, however, difficult to suppose that the drain-back plays a major role because the Model 2 representing the drain-back process does not adequately explain the observed waveform.