Page 494 - Acquisition and Processing of Marine Seismic Data
P. 494

10.4 SPECIFIC STACKING METHODS                      485




































           FIG. 10.27  Schematic illustration of CRS stack surface. (A) A 2D earth model of a curved reflector (bottom), and hypothet-
           ical travel time curves (blue curves) from the reflector on common offset gathers (top). NMO stacking is done along the red curve
           to obtain a zero-offset stacked amplitude represented by a zero-offset ray (red line) at point P 0 where h ¼ 0 and x m ¼ x 0 ,
           whereas all the amplitudes on the CRS stack operator (green surface) consisting of the reflections originated from a curved
           segment on the reflector (blue segment on the reflector) are summed up to obtain a CRS stack amplitude. (B) Schematic display
           of seismic data at half offset (h)-midpoint (x m ) coordinates. The CRS stacking surface, or CRS operator, is indicated by the green
           surface. The frontal plane indicated by A is a zero-offset section, like a conventional stack, and the side plane indicated by B is
           the gathers sorted according to their half offsets.

           seismic line with a low S/N ratio. In a CRS stack,  NospecificprocessisnecessaryforaCRSstack-
           stack quality is widely improved, especially for  ing procedure and a conventional processing
           the areas of the thrust fault where the data qual-  sequence is applied to the data prior to the CRS
           ity is poor and high-amplitude random noise  stack algorithm. CRS processing produces four
           dominates. In the 3D case, the CRS stack can  optimized sections: simulated zero-offset, α, R N ,
           be expressed by eight parameters: emergence  and R NIP sections, and a coherency section. Using
           or dip angle α, emergence azimuth β, and three  the CRS parameters, a poststack time migration
           independent radius values for R N and R NIP .  can also be applied to the simulated zero-offset
           However, only α, β, and R NIP are considered  section without a preliminary velocity model,
           since an eight-parameter search is extremely  since the CRS stack parameters provide the back-
           expensive. It not only improves the S/N ratio  ground velocity. In addition, the preliminary
           of the stack data, but also widely enhances the  velocityfieldfromaconventionalvelocityanalysis
           quality of time slices from 3D cubes (Trappe  can be used as a confinement during a CRS search
           et al., 2015).                               to avoid stacking of the multiples (Kumar, 2008).
   489   490   491   492   493   494   495   496   497   498   499