用小波方法系统检验强震“前驱波”

Systemically checking-up strong earthquake precursory waves with wavelet analysis

采用山东省泰安台观测精度达10-9量级的钻孔体应变仪观测资料,深入分析了2001—2005年MS≥7.5强远震前应变变化,对"前驱波"进行了客观、系统的检验.用高通滤波消除了周期大于128min的固体潮汐影响,然后用回归分析方法消除了气压的干扰.用小波变换方法将2—128min信号分解到6个不同频段进行分析.小波分析结果表明,该方法能提取出信号中微弱变化信息.根据小波变换提取的类似"前驱波"基本特征,提出了"超限率"方法,并对这种波形是否普遍存在进行了检验.该方法将单位时间内超过标准差的数据个数定义为单个超限率,类似"前驱波"波形存在时段内的超限率将高于正常时段.据此得出震前30天的超限率时间序列.用直线拟合该序列,最后对直线斜率进行统计分析.用超限率方法分析了所有小波分解的细节部分,得出了超限率时间序列的回归直线斜率的平均值和标准差,并分析其正负比.这3项统计数据显示,强远震前15天之内不能普遍检测到"前驱波".

In-depth analysis on strain variations before strong earthquakes （Ms ≥7. 5） during 2001 to 2005 was carried out based on the observations from borehole volume strainmeter with a resolution of 10^-9 at Tai＇an station in Shandong Province. Objective and systematic tests on precursory waves were carried out. High-pass filter was used to eliminate the effect of solid tides with periods greater than 128 minutes, and the influences by atmospheric pressures was removed through linear regression method. Wavelet transform method was applied to decompose the signals from 2 to 128 minutes into 6 different frequency bands for analysis. The result of wavelet analysis shows that weak anomalous variation information in signals could be extracted with this method. According to the basic characteristics of the waves similar to precursory waves, which were extracted with wavelet transformation, the analysis method using the overrun rate analysis （ORA） was developed for testing the waves. An ORA, which is the amount of data exceeding standard deviations within a certain time, was defined. The ORA within the time window with anolamous waves would be greater than the ORA of normal time. An ORA time series were obtained by computing ORA during 30 days before strong earthquakes. The series was fitted by a straight line with regression method and a statistical analysis was finally made on the slope of the line. All detailed parts decomposed by wavelets were analyzed with this method. The average value and the standard deviation of the slopes were obtained, and the positiveegative ratioes of the slopes were analyzed. These three statistical quantities show that precursory waves can not he generally detected during 15 clays before strong earthquakes afar.