Extraction of Rayleigh waves from earthquake signals has become an important topic in civil engineering and seismology. In this study, we propose an automatic filtering procedure to extract retrograde and prograde Rayleigh waves and their propagation azimuths from the earthquake signals, separately. In this method, the users do not need to specify an initial filtering limits. The proposed procedure has been first verified with synthetic signals. We show that the points where new retrograde and prograde elliptical motions appear in the signals can be easily distinguished by using the proposed method. Since the Rayleigh waves contents in real signals may change in different earthquakes, this method in comparison with previous methods will be more realistic in extracting Rayleigh from real signals. The proposed method is used to extract the retrograde and prograde Rayleigh waves from recorded earthquakes in Garner Valley Downhole Array. Recorded signals in 6 levels of earth depth have been studied in this work, at 0 m, 6 m, 15 m, 22 m, 50 m and 150 m. We find that the share of prograde Rayleigh waves become more than the share of retrograde waves in deeper levels.

从地震信号中提取瑞利波已成为土木工程和地震学的重要课题。在这项研究中，我们提出了一种自动滤波程序，分别从地震信号中提取逆行瑞普波和前进瑞利波及其传播方位角。在这种方法中，用户无需指定初始过滤限制。拟议的程序已首先通过合成信号进行验证。我们表明，使用所提出的方法可以轻松地区分信号中出现新的逆行和前进椭圆运动的点。由于真实信号中的瑞利波含量可能在不同的地震中发生变化，因此与以前的方法相比，该方法在从真实信号中提取瑞利时将更加现实。该方法被用于从加纳河谷井下阵列记录的地震中提取瑞利波的逆行和前进。在这项工作中，已经研究了6个地球深度级别的记录信号，分别是0 m，6 m，15 m，22 m，50 m和150 m。我们发现，在更深的层次上，前进瑞利波的份额变得大于逆行波的份额。