Seismic instruments: article

Аnnotation: Multichannel signal processing algorithms have been developed, which allow detecting signals of weak seismic events observed against the background of natural seismic noise. The algorithms of weak seismic signal detection are based on the estimation of coherence measure of P-waves, recorded by a small aperture array. The advantage of these algorithms over the known single-channel STA/LTA detector is the ability to detect time intervals of P-waves with a signal-to-noise ratio (SNR) less than one. Using the robust to noise phase bimforming algorithm [Zhang et al., 2008; Kushnir, Varypaev, 2017], an apparent slow-motion vector can be then estimated by the revealed time intervals of the seismic records, where the values of the coherence measure exceed a certain threshold. The developed multichannel signal processing algorithms were tested with the real seismic records by a small aperture array, deployed in the area of the Kursk Magnetic Anomaly, directly near the mining facilities for extraction of iron ore. Within 20 hours of seismic records, signals were revealed of both strong (underground mine explosions) and weak seismic events. In some cases it was possible to show that seismic events have a complex focal mechanism. On the basis of the optimal Capon filter [Capon, 1969] we have discovered the similarity (up to sign) of P-waves generated by the different weak events, the duration of which does not exceed 0.2 s and the shape is similar to the known function of the Ricker pulse [Bording, 1996].