Multigrid surface consistent deconvolution

John Millar, John C. BancroftABSTRACT


solutions, especially in the intermediate wavelengths in the frequency domain (Millar and Bancroft, 2004), (Millar and Bancroft, 2006).

The importance of the errors in these intermediate wavelengths in the deconvolution of seismic data is demonstrated on synthetic data. Multigrid and Gauss-Seidel surface consistent operators improve on trace by trace deconvolution when ground roll is included in the model.

Adding offset and midpoint consistent terms in the equations helps further remove the effect of ground roll from the surface consistent deconvolution operators. The multigrid deconvolution operators are more stable, and produce a more white spectrum than Gauss- Seidel operators. This leads to improved resolution throughout the seismic section.

The intent of this paper is to encourage the use of four term surface consistent deconvolution of seismic data aquired on land. Additionally, we suggest that the calculation of the operators used for surface consistent deconvolution can be improved using a multigrid method, as opposed to a Gauss-Seidel or conjugate gradient method.

View full article as PDF (1.55 Mb)