The inverse scattering series developed by Weglein et al is utilized for the prediction of internal multiples. The algorithm is implemented in 1.5 dimensions in the tau-p domain. While the method has shown promise in its ability to predict multiples there are still challenges to overcome to become a standard in processing. The issue addressed here is the computational expense of the method, which is shown to be decreased using Seismic Unix and parallel processing. The computational time for the chosen model is reduced by a factor of approximately 120 in comparison to the MATLAB implementation. It is also shown how artifacts from the prediction in 1.5D tau-p can be minimized through a time domain tau-p transform and as shown by Sun and Innanen (2015) a spatial cosine taper.
View full article as PDF (0.68 Mb)