Strategic research of effective prediction to increase the probability gain statistical space-time model for diagnosing abnormal seismic activity

(1)完成了适应地震活动区域特征的分层时空ETAS模型等贝叶斯模型软件的测试版。已为加州预报试验(UCERF3)提供了部分程序。(2)主震成功后一天内发生余震概率的实时预报研究。(3)从经验上论证了基于前震统计判别预报大地震的概率。(4)我们建立了一个分析重复地震的非平稳BPT模型，以可视化海底下板块边界应力变化的时空变化。(5)建立了一个非平稳的ETAS模型，用于从震群资料中捕捉断层上孔隙流体压力的变化。

(1) We have completed beta version of the software of Bayesian models, such as hierarchical space-time ETAS models, which can be adapted to regional characteristics of seismic activity. Some programs have been provided for prediction experiment in California (UCERF3). (2) Research for real-time prediction of the probability of aftershocks within one day after the mainshock has been successful. (3) We have empirically demonstrated that predicting the probability of a major earthquake based on statistical discrimination of foreshocks. (4) We have developed a non-stationary BPT model for analyzing repeating earthquakes to visualize the space-time variation of stress changes on the plate boundary beneath the seafloor. (5) We have developed a non-stationary ETAS model to capture the pore fluid pressure changes on the faults from the earthquake swarm data.