科研之友
微信
新浪微博
Facebook
分享链接
摘要
How to identify seismic wave velocity anomalies and to study their characteristics has practical significances for earthquake study and prediction. Almost all disastrous earthquakes in China's mainland have occurred in the Earth' s crust within depth of 5 similar to 25 km. Therefore, dynamic monitoring physical parameters of the Earth's crust is an effective approach to predicting earthquakes. Seismic Wave Radar (SWR) is a type of mechanical and electrical devices. It continuously excites Linear Frequency Modulation (LFM) signals into the Earth' s crust, and these signals are recorded by high sensitivity seismographs at the deployment sites. How to retrieve impulsive seismic waveforms from the recorded data which include signals and different kinds of noises is very important for calculating accurate travel times and temporal velocity variation. In this work, we propose a new method for processing the SWR data, which is based on accumulating energy in time-frequency domain. LFM is a non-stationary signal and its frequencies vary linearly as a function of time and have a good energy concentration, so it is very suitable for time-frequency analysis. We use Short-Time Fourier Transform (STFT) to analyze the data recorded by the monitoring stations, and acquire the time-frequency distribution, in order to validate whether the excitation time and frequencies are consistent with those set by the control system of SWR. Then, two methods are introduced for the waveform retrieval from the SWR data. One is the Wigner-Ville Distribution (WVD) algorithm with the best time-frequency concentration capability for the LFM signals, and the other one is the Wigner-Hough Transform (WHT) which is helpful to suppress the cross-time interference in the signal detection and parameter estimation for the multi-component LFM signals. By aggregating energy of the excited signals of the SWR on the LFM line with WVD method and accumulating the concentration energy to a crest by WHT, we can extract the peak row according to the inclination angle of the LFM line and compose the seismic wave travel-time curve after calculating the travel time. Firstly, in order to quantify the repeatability of the SWR, we calculated cross-correlation coefficients of all waveform pairs retrieved from different time windows. Within the total 126 hours' SWR data, cross-correlations coefficients of 123 hours' are bigger than 0.999. The high correlations between the waveforms indicate the excellent repeatability of the SWR. In addition, the results obtained from STFT methods show that the excitation time and frequencies of the LFM signals met the control requirements and the repeatability is more than 99.9%. Finally, the crustal velocity structure obtained from the H-21 section near the location of this experiment was used to calculate the reduced travel-time curves of Pg, Sg, PmP and SmS of the detecting line, which were compared with those retrieved by our method. The results demonstrate that the proposed method is an effective way for retrieving waveform, identifying seismic phase and measuring travel time. By using this method, we can clearly identify Pg and Sg, and high frequency seismic phases such as PmP and SmS with strong amplitude in some epicenter distances can be discernable. In all, the SWR with high repeatability can be easily applied to monitoring temporal changes and imaging the spatial variations of the subsurface structure. The proposed method provides a feasible way to process the SWR data and advance its application in monitoring the crustal processes.
- 出版日期2015-4
- 单位中国地震局地震预测研究所
