Seismic Multiples can be Responsible for Completely Masking Underlying Data: A Case Study Shows a Remarkable Example of Hidden Information Revealed

John DeWildt, Derek Skoyles, Halle Rhea, and Nigel Payne
Spectrum EIT Inc, Houston, TX

In situations where the multiple moveout velocity is very close to the primary moveout velocity algorithms such as F/K and Radon are often ineffective. This has resulted in a number of essentially no data areas where primary reflections are completely masked by multiples resulting in potentially overlooked opportunities.

This poster describes a situation whereby the interpreter knew from other sources of information to expect dipping beds in one direction whilst the conventionally processed seismic data suggested ‘normal' behaving data in the opposite direction. Other than the interpreters expectations there was nothing abnormal to indicate a multiple trend, stacking velocities were normal and the processors had no idea that the primary reflections were being obscured.

In this case the discriminating factor to differentiate between primary and multiple was dip. The processing analysts were able to model the multiple generator in 3D space and remove the multiples through application of an Eigenvalue filter. Once the multiples were removed a second pass of stacking velocity analyses revealed a subtly different velocity range and the resulting stack was startlingly different.

This technique illustrates one of the eye opening situations in data processing where once the multiples are removed the interpretation of the underlying geology is completely different. The teamwork between interpreter and analyst can unveil an opportunity that might otherwise be completely missed.