The present fault model available in the field of seismology is the highcut source model. Since detailed behaviors of short-period components at the source are still unknown, it is quite difficult to reproduce observed accelerations (or Intensities) by modelling the earthquake source.
In this paper, applying a low-pass filter to observed strong-motion accelerograms in the frequency range from O Hz to ν Hz, relations between maximum amplitudes and cutoff period T (=1/ν) are obtained. The empirical equations for maximum acceleration, um (gal), maximum velocity, u_m (kine), and maximum displacement, u_m (cm), as a function of T (sec) are
found to be
and
If values at T=T0 are known, maximum values for short periods, for in-stance at T=0.1sec, can be estimated.
For dislocation source models corresponding to M=5 to 8, theoretical seismograms are computed. The above cutoff period is defined as the period at which the spectral amplitude is equal to or less than 1/1, 000 of the maximum value and maximum short-period accelerations at T=0.1sec are estimated by using the above empirical relations.
Although, in this study, only an infinite medium is considered and expected accelerations at prescribed stations computed by using I-R relations of SHIMA (1977) have large variations, it can be said that maximum accelerations obtained for the present fault models give plausible values.
In the present paper, possibility is only investigated whether the fault model, assuming the simple infinite medium, can realize or not the shortperiod maximum accelerations we experience during the earthquake. It is quite important to study spatial distributions of maximum accelerations as well as maximum velocities in the period range, which bring about great damage in the epicentral area, by adopting a pertinent hypothetical earthquake model and taking more realistic multi-layered surface structures into account.

View full abstract

A calculation scheme for Rayleigh wave solutions at higher frequency range and/or higher modes was investigated. Process of long significant figures was made clear. By combining the computational algorithms developed by HASKELL (1953) and DUNKIN (1965), the loss of significant figures in evaluating the internal displacement was found to be improved. A suggestion was proposed for controlling the minimum layer number required to be put into calculation to maintain sufficient accuracy.

View full abstract

Analytical expressions are derived for the strain fields at an arbitrary depth due to an inclined fault in a semi-infinite medium. These expressions, being composed of elementary functions, are suitable for numerical computations. Basing upon these expressions, we develop a computer program, by which contour maps of deformation fields can be drawn in an arbitrary area designated by input parameters.
Using this program, several examples of the strain fields in a semi-infinite medium are shown. Effect of the free surface is clarified by comparing patterns for a semi-infinite medium with those for an infinite medium. For example, in the vertical cross sections, this effect yields saddle points and concentration of contour lines and also the "reverse area", where sense of the deformation is opposite to that for the case of an infinite medium. The reverse area appears clearly in the shear strain components given by differentiating horizontal displacements with respect to depth.

View full abstract

The time-dependent equations of thermal convection in a fluid with variable viscosity are solved numerically. The method allows computation for the medium in which the vertical viscosity profile contains a minimum. Some computer techniques are also developed to observe the time development of convective patterns on a computer graphic display unit.
Two types of thermal convection are studied. One is the free convection model, in which the upper boundary condition of a convection cell is free from stress. The other is the forced convection model, in which the upper boundary condition is kept at a constant horizontal velocity. The convective current in the underlying layer is induced by this motion. Three different viscosity distributions with depth for the free convection model and one for the forced convection model are investigated. In all cases, the initial convective cell having an aspect ratio of 3, breaks up, as the convection develops, into several cells of approximately same size with an aspect ratio of about 1.

View full abstract

Adiabatic elastic moduli for a NaCl single-crystal under high pressures up to 80kbar at room temperature were determined through ultrasonic sound velocity measurements in a cubic anvil-type high pressure apparatus with a solid pressure transmitter. Ultrasonic measurements far both the longitudinal and transverse waves were carried out simultaneously by a pulse-echo-overlap technique. Pressures were determined by X-ray measureinents of NaCl based upon Decker's equation of state. The adiabatic bulk modulus K_s calculated from the Decker's formulae as well as various pressures was combined with the ultrasonic sound velocity to provide pressure dependence of elastic moduli C_ij for NaCl.
Results of measurements yield zero-pressure values of C₁₁=5.16×10¹¹, C₁₂=1.22×10¹¹ and C₄₄=1.36×10¹¹dynes/cm². The plots of C_ij versus pressure are fit by a least square method with straight lines for pressure range zero to 80kbar, to give the following slopes: ∂C₁₁/∂p=8.72, ∂C₁₂/∂p= 1.22 and ∂C₄₄/δp=0.31. In the overlapping pressure range, ∂C_ij/∂p values obtained by this work ae slightly smaller than those obtained by others.

View full abstract