Connecting optical holography and seismic imaging

An optical hologram is a recording of the interference between two different wavefields. One wavefield interacts with the objects to be imaged (the “object wave”) and the other does not (the “reference wave”). If only the reference wave is then shone through the recorded hologram, the objects are reconstructed visually in 3D. This occurs because the interference pattern encodes the phase information of the object wave.

In seismic imaging, the phase information of the seismic wavefield is directly recorded by seismometers. Therefore, a reference wave is not needed and the object can be reconstructed without. Because there is no real-world equivalent to shining light through the hologram in seismology, computational methods are needed. The closest seismic equivalent to optical holography is full-waveform inversion.

We demonstrate that a similar connection between optical holography and seismic imaging arises naturally in surface-wave propagation. To image any structure with surface waves, phase or group velocities are measured at different frequencies. The resulting dispersion curves are commonly biased by wavefield interference. Closely connected bias patterns can also be observed in the apparent arrival angles of wavefronts and the peak amplitude of surface wavegroups. Similar to the optical holography, the surface-wave patterns can also be explained as interference of the “object” and “reference” wave. In seismology, however, these interference patterns can be produced by two distinct mechanisms: the “object” wave could be emitted either by heterogeneous structure in case of deterministic surface waves propagating from earthquakes, or, in seismic interferometry, by additional isolated noise sources. Our results suggest that this effect a) should be carefully considered as a potential source for bias in imaging applications, and b) may reveal new opportunities for seismic imaging.