地震活动时序谱的涨落统计特征与地震活跃期

为寻求地震活动演化的统计判据，本文采用多项式拟合法展示地震时序谱之涨落谱，以分维函数法判断最小邻间距（NNS）分布的Poisson性，然后采用MKS无参量法判别涨落谱之统计稳定性，确定NNS分布的转变区，具体研讨了四川、云南若干震区地震活动时序谱的涨落统计分布，发现地震活跃期到来时NNS分布表现出非Poisson化的特征。进一步通过改变和调节时序数据样本集之震级下限，并逐段前推进行统计计算的方式进一步验证结果的可靠性，客观地确定了这些震区地震活动时序涨落谱统计特征的转变期，使根据该统计法判断地震活跃期的方法论渐趋完善。

THE STATISTICAL CHARACTERISTICS OF FLUCTUATION SPECTRA OF EARTHQUAKE TIME SERIES DATA AND THE EARTQUAKE ACTIVE PERIOD

It is an attempt to search for a statistical criterion on the evolution of earthquake actitities. The method of polynomial expansion fitting is adopted to unfold fluctuation spectrums of earthquake time series data.The Poissonization of the Nearest Neighbor Spacing (NNS) distri bution is determined by means of the fractal dimension function method. Furthermore, the nonparameteric MKS method is introduced to analyze the statistical stationarity of fluctuation spectra and to identify the transition periods in the NNS distributions. Some examples including Sichuan and Yunnan Province are then discussed. It turns out that the NNS distributions are characterized by nonpoissonization during the earthquake active periods. In addition, the reliability of these results is tested through both changing the lower limit of magmitude of the selected earthquake time series data and taking a progressive statistical calculation step by step respcctivily.