ABSTRACT
ABSTRACT Because the earthquake motion phase (EPM) has the fractal nature it has been discussed what kind of stochastic process can simulate EPM. But all efforts done so far have restricted within Gaussian regime. First we study the main cause of this reason.
We deconvolute EPM into the linear delay and fluctuation parts then investigate a non-Gaussian nature of the fluctuation part of phase (FPP). We find that the probability distribution function of the mean gradient of FPP is expressed by the Levy-flight distribution function. Because the variance of Levy-flight distribution does not exist we develop a new stochastic process not satisfying the central limit theorem to simulate FFP.
