用动态自回归模型预报大华北地区强震的发震时间

Prediction of Occurring Time of Strong Earthquakes in the Great North China by Dynamic AR Model

本文采用动态自回归模型研究了大华北地区(32°N—42°N,105°E—123°E)自1900年以来Ms≥6.0地震的发震时间。结果表明,动态自回归模型时变参数(时变系数)的变化是有规律的,其增量大体上是一些简单周期函数的叠加。把它们用富里叶级数展开,并选取一定数量的主周期,即可对未来大华北地区Ms≥6.0地震的发震时间做出某种程度的预报。该法内符检验较好,作者认为它可能是预报强震发震时间的一种有效方法。

In this paper,the occurring time of earthquakes ( Ms≥6. 0) from 1900 up to now in the great North China region (32°N-42°N; 105°E-123°E) was studied by dynamic AR model. The results showed that the change of time-varying parameters (coefficient) in dynamic AR model has a regularity. Its increments are piled up by some simple period functions. They are spread out with Fourier series and can be used to predict the occurring time of earthquakes ( Ms ≥6. 0) in the great North China region in some main periods. This method is better in inner accord with testing. The authors consider that this method is promising for predicting the occurring time of strong earthquakes.