关键词: 无震蠕滑, 独立分量分析, 相对强度指数, 峰度值, 累积分布函数, 慢滑移

Abstract:

Transient aseismic creep of the fault frequently induces earthquakes with high magnitude and destructiveness. In allusion to the challenges of slow sliding speed and difficult to detect, an automatic detection method of transient aseismic creep information for the fault was proposed, which based on the abnormal fluctuations of GNSS continuous coordinate time series. First, independent component analysis was used to improve the signal-to-noise ratio (SNR); then the relative strength index and kurtosis value of the fluctuations of the coordinate time series were calculated; finally a creep signal probability was converted through the cumulative distribution function, so as to the fault creep event was detected. In this paper, a 500-day GNSS surface displacement time series was simulated that included a 25-day transient creep signal. The experimental results show that the creep information of the fault can be effectively detected when the signal strength was at least equivalent to the noise level. After calculating the GNSS data for three consecutive years in the Akutan Zone, onecreep signalwas detected and it might be an aseismic creep signal related to the strong movement of the volcanic. In accordance with the processing results of seven-year GNSS data from 18 stations of the Crustal Movement Observation Network of China in Sichuan province, four abnormal signals was found. The analysis results indicated that these signals may be closely related to the abnormal displacement caused by stress accumulation and fault creep caused by the earthquake.

Key words: aseismic creep, independent component analysis, relative strength index, kurtosis value, cumulative distribution function, slow slip