introduction

The last decade has seen an exponential growth in the use of 3-D seismic imaging. Contemporaneous with this development, imaging techniques have become more complex in the effort to account for multi-pathing in complex media and to produce ``true amplitude'' migrated pictures of the subsurface. Since multiples are not accounted for in the physical model that leads to these migration methods, they can severely affect the final migration result producing erroneous interfaces or amplitude artifacts; consequently, the multiples have to be removed from the data. As pointed out by Weglein (1999), the multiple attenuation techniques may be divided into two families: (1) filtering methods which exploit the periodicity and the separability (move-out discrepancies) of the multiples and (2) the wavefield prediction/subtraction methods, where the multiples are first predicted Verschhur et al. (1992); Weglein et al. (1997) and then subtracted Brown et al. (1999); Clapp and Brown (1999); Doicin and Spitz (1991); Dragoset and MacKay (1993); Spitz (1999).

As oil companies lead exploration towards more complex geological structures (e.g., salt plays) and use 3-D surveys intensively, the attenuation of the multiples becomes more challenging. Spitz (2000, Personal communication) recently asserted that multiples are the number one problem in seismic processing. Traditionally, filtering techniques are the method of choice for multiple processing because of their robustness and cost. However, they have some limitations when tackling multiples in complex media (predictive deconvolution) and in the preservation of primaries'amplitude (f-k filters). Wavefield methods overcome these limitations, therefore they are becoming more popular in the seismic industry. Nonetheless, they are often arduous to tune, generally slow, and very difficult to extend in 3-D for coverage reasons.