Further Development of the Matched Filter Method for Solving Seismological Problems

The progress of seismology and geophysics as a whole depends on the detection thresholds and the accuracy of measuring physical signals. The repeatability of geophysical phenomena, such as earthquakes, allows for the efficient use of the matched filter as a detector of repeating signals. The matched filter detection threshold depends on the noise characteristics. For example, microseismic noise is created by many physical sources and includes signals that may be partly coherent with the templates. Stochastic noise, which is not present in a mathematically defined form in nature, is optimal for the matched filter. In this paper, stochastization of microseismic noise is achieved by adding stochastic noise generated by a random number generator to waveforms on seismic arrays. The signal-to-noise ratio on seismic arrays increases by a factor of 5 in the presence of noise coherent to the signal in the microseismic noise. For the microseismic noise without dominant components, adding stochastic noise provides a gain of 2–3 times. A decrease in the rate of false alarms is observed, which also lowers the detection threshold.