[1] We conducted 1-year continuous measurements of in situ compressional wave velocity and attenuation across a distance of 12 m in a vault at the coast of Miura Bay, Japan, using a piezoelectric transducer as the ultrasonic wave source. We detected, for the first time, periodic variations in both velocity and attenuation, with amplitudes of 0.3%, and of 4%, respectively, corresponding to the diurnal and semidiurnal lunar and solar tides. These variations correlate well with the in situ areal strain change due to tidal ocean loading estimated from the strain records about 100 m away from the observational site. Relative minimum velocity and relative maximum attenuation occur at every time of relative maximum areal strain (expansion), suggesting that opening/closure of pores or cracks of in situ rocks is responsible for these periodic variations. The velocity variation shows a remarkable 14-day periodicity corresponding to the spring and neap tides, arising from a nonlinear response of the velocity change to tidal strain change such that the velocity change depends on how slowly the tidal dilatation departs from its peak value.

Uniaxial compression tests have been conducted on Oshima granite under various constant axial strain rates ranging from 10−8 to 10−4. The results showed that the strength and the acoustic emission rate increased exponentially with increasing strain rate. The inelastic volumetric strain rate defined by the differentiation with respect to the stress increased with decreasing strain rate. The redistribution of microcracks due to subcritical crack growth was considered theoretically, and the equations derived from the theory were compared with the experimental results. The agreement between the theoretical and experimental results shows that stress corrosion plays not only a major role in the brittle creep under constant load but also dominates the strain rate effects on strength and dilatancy observed in the constant strain rate loadings.

Employing polyhedral specimens, longitudinal and shear wave velocities were measured in various directions of propagation and polarization. Sound velocities showed orthorhombic elasticity in all of the rocks. With the assumption of orthorhombic elasticity the nine stiffness constants of all samples were determined by the sound velocities under atmospheric pressure and Kelvin‐Christoffel's equation. Twenty‐one stiffness constants of Oshima granite, determined without assuming any symmetry, also showed orthorhombic features. Directions of the symmetry axes agreed well with the orientation of the preexisting cracks. Akaike's Information Criterion showed that the orthorhombic model with nine nonzero elastic stiffnesses was better than the model having 21 nonzero elastic stiffnesses for Oshima granite. The polyhedrons of two granitic rocks were loaded under hydrostatic pressure. All components of the stiffness constants increased with pressure. Under pressure of more than 120 MPa, two granitic rocks were approximately isotropic. The results show that oriented microcracks are mainly responsible for the orthorhombic elasticity of the granitic rocks and also indicate that dry oriented cracks can not be a cause for the anisotropic elasticity of granites at depths of more than 6–8 km.

No abstract