Automatic Microseism Location by Full-waveform Triangulation

We present a method for locating microseism hypocenters in both space and time, by correlating multiple reconstructed wavefields corresponding to different groups of receivers. The method is mathematically and conceptually similar to triangulation, but does not require arrival picking. The fundamental idea is that for correct velocity for the wavefield from all receivers arrives at the spatio-temporal location for the microseism simultaneously but on distinctly different paths. Thus, the wavefields for subsets of receivers do not coincide in space and time until the wavefields are at the hypocenter location. To capture the coincidence of the wavefields, we propose a new imaging condition for passive seismic datasets that is similar to the zero-lag cross correlation imaging condition but uses the cross-correlation of multiple receiver wavefields instead of the source and receiver wavefields. We demonstrate this idea on a synthetic microseismic dataset by reconstructing the wavefield using time-reversed propagation for a few groups of receivers that have distinctly different angular aperture and then applying the proposed imaging condition. This method is able to correctly image the microseisms in space and time, even though the signal-to-noise ratio is low.