| Investigation of noise sources and assessment of noise levels can provide reference for the use of seismic waveform data from seismic arrays.Using continuous seismic waveform data from nine substations of the Hailar seismic array from 2017 to 2021,the frequency domain characteristics of noise and the periodic variations were analyzed using power spectral density and power spectral probability density function methods,and a noise model of the array was obtained.The overall noise level of the array was relatively low,with little variation from 2017 to 2021.A prominent noise in the frequency band of 1-2 Hz has been present in the array for a long time,which appeared after 2002.Periodic changes with annual cycles were found within two frequency bands.Low-frequency noise around 0.10.5 Hz had lower noise levels from April to November,and it was inferred that the low-frequency noise was related to oceanic activity based on its frequency and periodic characteristics.Highfrequency noise around 2-5 Hz had higher noise levels from April to November,and it was inferred that this high-frequency noise was related to human activity.From April to November and from 2 am to 9 am each day,B3 sub-station exhibited prominent high-frequency noise of 8-16 Hz,which was inferred to be related to vehicle traffic on nearby roads.Power spectral density and power spectral probability density function images also reflected the time of data missing and instrument malfunctions and the corresponding power spectral distribution characteristics of waveform data.Analysis of signal correlation and noise correlation is the foundation of earthquake array data processing.The cross-correlation coefficients between noise and seismic signals at different filtering frequency bands were calculated using the Hailar seismic array.In the 1-2 Hz frequency band with an inter-station spacing of 1 km,the maximum negative correlation coefficient was around-0.4.With increasing frequency,the maximum negative correlation coefficient gradually decreased to around0.05 in the 2-4 Hz and 4-8 Hz frequency bands.However,there were no negative correlation coefficients in the 0.5-1 Hz frequency band.The correlation coefficients of the P-phase of near,far,and deep earthquakes filtered accordingly were around 0.8,and overall,they decreased with increasing station spacing.The study showed that there was no systematic pattern in the variation of the correlation coefficient of the P-phase with the back azimuth after the earthquake event.Three explosion events occurred in the North Korean region on January 6,2016,September 9,2016,and September 3,2017.Using data from the Hailar seismic array,the amplitudes of P-wave and Lg-wave for these three explosion events were picked,and the amplitude spectrum,power spectrum,earthquake magnitude,and explosive yield were calculated.The short-period body wave magnitudes mb(P),which were calculated using the national standard for magnitude determination,were 5.3±0.1,5.6±0.1,and 6.1±0.1 for the three explosion events respectively.The short-period body wave magnitudes mb(Lg)were also calculated using the third peak value amplitude method of the Lg wave,and were 4.33±0.05,4.56±0.4,and 5.60±0.03 for the three events respectively.The low-frequency component of the Lg wave of the 2017 event was more developed with higher amplitudes compared to the previous two events,likely due to differences in the explosive destruction mechanism or technology process for the 2017 event.The average yield of the three events was obtained from the yield-magnitude empirical formula using mb(P)and mb(Lg).The explosion yields were 14.4±4.3 kt,34±11.6 kt,and 190.4±45.4 kt for the three events respectively.Therefore,the energy of the 2017 explosion event was about 5.6 times that of the September 2016 event and about 13.2 times that of the January 2016 event. |
PDF Full Text Request