Three basic methods for suppressing multiples exist in published literature. Deconvolution methods use the periodicity of multiples for suppression and are effective in suppressing short-period free-surface multiples generated at shallow reflectors. Filtering methods use differential moveout between primaries and multiples that are separate in the f-k, tau-p, or Radon domains. These filtering methods can successfully suppress multiples generated at moderate to deep reflectors where multiples are well- separated from their primaries. The third group of methods, wavefield prediction and subtraction, based on the wave equation, use recorded data to predict multiples by wave extrapolation and inversion procedures. These wavefield methods obtain multiple-free data by subtracting the predicted multiples and can suppress all multiples generated by any complex system of reflectors. This can be accomplished as long as the recorded wavefield has complete internal physical consistency between primaries and multiples. The most striking advantage of wavefield prediction and subtractions over other methods is its ability to suppress multiples that interfere with primaries without coincidentally attenuating the primaries.
Wavefield prediction and subtraction methods are the most promising methods for multiple suppression, but they have considerable cost and are limited by data acquisition and processing more than other methods. Therefore, the choice of multiple suppression methods should be based on the effectiveness, cost, and processing objectives, and depends on how well a particular data set fits the assumptions of each multiple attenuation method.
View full article as PDF (2.42 Mb)