Page 377 - Acquisition and Processing of Marine Seismic Data
P. 377

368                          7. SUPPRESSION OF MULTIPLE REFLECTIONS

           deconvolution is based on the predictability of  shows this process on a schematic CDP gather
           the multiples; stacking, f-k filtering, or Radon  containing three primary reflections (P)and a
           velocity filtering in the τ-p domain is based on  seabed multiple reflection (M). Because the
           the moveout time, dip, or velocity differences.  surface-related multiples propagate in the water
           In addition, the wave equation multiple rejec-  column and in the uppermost sediments, their
           tion (WEMR) method uses wave equation        propagation velocity generally equals the aver-
           modeling and subtraction, and surface-related  age water velocity (approximately 1500 m/s).
           multiple elimination (SRME) employs the auto-  This velocity is considerably low when compared
           convolution approach. Most of these methods,  to the velocities obtained from the primary reflec-
           except common depth point (CDP) stacking,    tions with the same zero offset time on CDP
           are applied to prestack data, commonly to the  gathers, which makes the curvatures of the mul-
           preconditioned shot or CDP gathers.          tiple reflections, or their moveout times, higher
                                                        than those of the genuine reflections (Fig. 7.3A).
                                                        This difference is particularly pronounced for
                       7.1 CDP STACK                    far offset traces. Since the velocities of primary
                                                        reflections are used during the NMO correction,
              Because of the moveout time differences   which is higher than the velocity of multiples
           between primary and multiple reflections after  (Fig. 7.3B), the reflection hyperbolas of multiples
           NMO correction, most of the multiple ampli-  are not perfectly flattened after NMO correction
           tudes are suppressed during stacking. CDP    (Fig. 7.3C).
           stacking is the earliest technique which success-  While in-phase amplitudes of primaries are
           fully suppresses the multiples based on the resid-  strengthened in the stacked trace, out-of-phase
           ual moveout differences. Fig. 7.3 schematically  amplitudes of the multiples tend to diminish




























           FIG. 7.3  (A) A schematic CDP gather involving three primary (P) and one multiple (M) reflections, (B) schematic NMO
           velocity functions for primary (black) and multiple (red) reflections, (C) same schematic CDP gather after NMO correction
           using the velocity function of primary reflections, and (D) suppressed multiple energy on the stacked trace.
   372   373   374   375   376   377   378   379   380   381   382