High-resolution time reverse imaging for microseismic event location

Abstract

Summary Time reverse imaging has become a standard technique for locating and characterising seismic events. No identification of events or their onset times is required for locating events with time reverse imaging. Nevertheless, because of the resolution limits of the source signals, it can not reliably locate the sources that are close to each other, i.e., a small concentrating source distribution. We propose a new time reverse imaging method to address this issue. First, we divide the wavefields into several small parts according to the bounds of the maximum absolute amplitude at each time step. The neighboring wavefields of each small part are extracted, and they are centred at the picked points that correspond to the maximum absolute amplitude of each small part and given by a circle with a radius of half the dominant wavelength of the source signal. Then we introduce the Gaussian-type weights to weight these neighboring wavefields. Finally, these extracted wavefields are cross correlated. The crosscorrelation creates a new imaging condition. It yields good location results, deviating from the actual source locations by far less than half the prevailing wavelength of the signal, even in the case of sparse acquisition and poor S/N ratio.