Abstract
Decomposing an earthquake motion phase into the linear delay and fluctuation parts we investigated stochastic characteristics of phase difference of the fluctuation part. We took up the probability density characteristic of the mean phase gradient which is the quotient of a phase difference with respect to the concerned circular frequency interval. The probability density function of this mean phase gradient was uniquely defined for a rather wide range of circular frequency interval. It, moreover, had very unique distribution characteristics with thick tail and multiple crests. We proposed a method to express this distribution characteristics as a weighted sum of Levy-Flight probability density functions. We also developed an innovative algorithm to simulate the stochastic process obeying this peculiar probability density function by modifying the already proposed method for simulating the stochastic process which violates the central limit theorem.
