The temporal variation of seismic velocity near the Nojima Fault, which ruptured during the 1995 Kobe earthquake (Hyogo-ken Nanbu earthquake), was detected using an accurately controlled routine-operated seismic source (ACROSS). The source generates elastic waves by a centrifugal force of an eccentric mass rotating around an axis. The mass is driven with an AC servomotor whose angular position is accurately controlled with reference to a very accurate global positioning system (GPS) clock. The error of the mass’ position is less than 0.002 radian and does not accumulate. As a result, the source generates sinusoidal waves of very narrow spectral peaks enabling their detection with an excellent signal-to-noise ratio. Although the stability of the rotation is quite excellent, a large daily variation was found, which seems to be caused by changes in atmospheric temperature. The daily variation was 10% in amplitude and 0.1 radian in phase of the signal observed at the 800 m borehole seismometer. A significant variation was found to be due to that of coupling between the rotational source and the foundation made of reinforced concrete in which the source was situated. In order to make a correction on the signal of the 800 m borehole seismometer, the vibration of the foundation was measured and modeled assuming a rigid body movement. The correction successfully reduced the daily variation by approximately 90%, resulting in a variation of 1% in amplitude and 0.01 radian in phase. The phase variation of 0.01 radian corresponds to 100 μs and less than 0.1% in velocity over 1000 m between the source and the receiver.
[1]
C H Scholz,et al.
Earthquake prediction: a physical basis.
,
1973,
Science.
[2]
Keiiti Aki,et al.
A precise, continuous measurement of seismic velocity for monitoring in situ stress
,
1974
.
[3]
K. Doi,et al.
In situ measurements of elastic wave velocity in a mine, and the effects of water and stress on their variation
,
1988
.
[4]
M. Wyss,et al.
Mapping the frequency-magnitude distribution in asperities: An improved technique to calculate recurrence times?
,
1997
.
[5]
David J. Varnes,et al.
Predictive modeling of the seismic cycle of the Greater San Francisco Bay Region
,
1993
.
[6]
H. Kanamori,et al.
State of stress before and after the 1994 Northridge Earthquake
,
1997
.
[7]
D. E. Gawith,et al.
Reservoir monitoring of the Magnus Field through 4D time-lapse seismic analysis
,
1996,
Petroleum Geoscience.