A compact gravity observation system using a force-balance accelerometer, which enables efficient measurement of gravity on moving carriers, is under development (Matsuo et al., 2012). The observation records obtained by the device are contaminated by various noises, including those whose sources are not identified. A low-pass filter was not able to remove all of such noises and efficient noise removal methods are needed.
 This paper proposes to use methods which remove noises exploiting statistical independence of noise signals and gravity observation data. We discuss applicability of Second Order Blind Identification (SOBI) and Thin-Independent Component Analysis (Thin-ICA). They are originally developed for Blind Signal Separation (BSS), which identifies original signals from the time series data that is a mixture of several original signals.
 The gravity measurements using a ship were carried out in the Toyama Bay and the observation records were processed by those two methods. The gravity signal obtained under appropriate conditions showed a good agreement with the high quality data obtained by AIST (National Institute of Advanced Industrial Science and Technology), indicating that methods using statistical independence can be promising noise removal methods. They did not work efficiently under poor conditions, and it is necessary to clarify the required conditions for the observation.

View full abstract

  We proposed a new waveform inversion method to estimate a 2-D velocity structure of deep sedimentary layers from Love wave in earthquake records. In this method, first, we conduct particle motion analysis to investigate the characteristics of Love wave propagation in earthquake records observed at strong motion stations located along a target line for validating a 2-D approximation in its propagation. Next, we conduct waveform inversion of Love wave at the stations in the target line for a two-dimensional S-wave profile. We applied the method to synthetic Love wave generated from ground motion simulation of the 2011 Mt.Fuji Region Earthquake (M_J6.4) considering existing 3-D model of Kanto basin. A 2-D velocity structure was well reconstructed in the waveform inversion. We, furthermore, applied the method to Love wave observed at stations in Sagamihara to Setagaya in the western part of Kanto basin during the same earthquake. The observed Love wave was well reproduced, and estimated S-wave velocity structure was thinner than the model by Headquarter of Earthquake Research Promotion (2009). Our inverted model can explain observed horizontal-to-vertical ratio of coda wave better with theoretical Rayleigh wave ellipticity than that of the previous model.

View full abstract

 We applied a method, used for the estimation of the dip angle of two-dimensional structures such as dikes, to the estimation of the dip angle of a fault or structure boundary. Prior to the application, we undertook numerical tests to evaluate whether this method could be applicable to the estimation of such dip angles. It was established that the method could estimate the dip angle of a fault or structure boundary and that it could recognize the difference between a normal-fault type boundary and a reverse-fault type boundary. As an application with field data, we applied the method to the Bouguer gravity anomaly in and around the area from Iyo-Nada, across Beppu Bay to the Hohi volcanic zone. We obtained fault and/or structural boundary dip angle distributions that were in agreement with previous studies.

View full abstract

 In cold areas such as Hokkaido in Japan, underground water freezes in winter to produce frozen soil, while it thaws to become a liquid in other seasons. The ground in this situation is called seasonally frozen ground.
 Frozen ground causes damages to the ground, roads and ground water pipes. On agricultural land, the freezing of the ground results in poor drainage and slows the rise in soil temperature during the spring. Furthermore, freezing promotes slope failure on bare land. Meanwhile, the thawing process generates cracks and depressions in a road surface. It is thus important to understand temporal variations in the location and expanse of frozen regions for the maintenance of constructions and the prevention of disasters.
 We made measurements of the air temperature, soil temperature, frozen depth, and resistivity for 5 years at Obihiro in Hokkaido, which is a Japanese prefecture known to suffer deep freezing of the soil. To clarify spatial and temporal variations in the frozen region, a series of two-dimensional resistivity analyses were conducted approximately every 2 weeks during the period from November 5, 1995 to May 1, 1996, which covers the whole period from the beginning of freezing to the end of thawing.
 In the freezing process, highly resistive regions formed discretely in a shallow layer and extended deeply as the soil temperature decreased. In the thawing process, the resistivity structure around the boundary between frozen and unfrozen regions changed drastically between the days before and after the date on which the maximum frozen depth was recorded by the frozen depth meter. As thawing proceeded, resistivity in the highly resistive regions decreased by factors ranging from 10 to 100, and these ‘moderately’ resistive regions moved both laterally and in the deep direction, before finally vanishing.

View full abstract

[in Japanese]

View full abstract