Elastic Born modeling in an ocean-bottom node acquisition scenario

Discussion and conclusions

The primary assumption of PZ summation for OBN data is that the vertical geophone contains only pressure wave energy. I have shown that doing a PZ summation on OBN-acquired hydrophone and vertical geophone data, where the vertical geophone component contains significant amounts of non-pressure wave data, will result in an introduction of the non-pressure wave energy into the summation result and subsequently degrade the image. In the simple case I used, the only non-P energy was shear waves. Running acoustic RTM with shear wave data as input will invariably image the shear wave energy away from the true reflector position, since shear waves do not propagate with an acoustic velocity. However, since the shear waves have a different moveout from the pressure waves, the stacking procedure may be sufficient to remove any artifacts they may create in the image. With elastic RTM, the shear and the pressure waves may both be imaged at correct reflector positions, but additional artifacts appear as a result of mode-conversion at the data injection point.

The realism of the scenario I've modeled here can certainly be called into question. The shear velocity in the sea bottom tends to be a rather smooth gradient, which can range from a few tens of meters per second in the topmost unconsolidated sediment, to a few hundred meters per second as depth increases and the material becomes more consolidated. Therefore the very obvious P-to-S conversion in the incident wavefield (Figure 4(c)) is probably very weak. This means that the SS reflection in the Born-modeled data I've shown will be likewise very weak , and not contribute to image artifacts. The reason for choosing a minimum shear velocity of $500 m/s$ was purely practical - so that the modeling grid would not have to be too fine. What could conceivably contribute to artifacts are the P-to-S conversions within the solid layers, as these are much stronger shear waves and are unaccounted for by the PZ summation's ``P-only'' assumption.

PZ summation is a more robust method than the methods mentioned in the introduction (Schalkwijk et al. (2003); Wapenaar et al. (1990); Amundsen (1993); Dankbaar (1985)), in that it requires less subsurface parameters to operate. It cannot however tell pressure wave energy from other energy, and assumes all data recorded on the hydrophone and vertical geophone must be P-wave energy.

I am currently working on a method which will be medium independent, and which will extract the pure pressure energy from the geophones. This hypothetical P-wave-only geophone data can be fed into the standard PZ summation in order to separate upgoing from downgoing wavefields in the data.

Elastic Born modeling in an ocean-bottom node acquisition scenario