Page 464 - Acquisition and Processing of Marine Seismic Data
P. 464

9.4 QC IN VELOCITY ANALYSIS                        455




























           FIG. 9.29  Utilizing NMO correction during velocity picking to discriminate the primary and multiple velocities, especially
           in shallow water areas. (A) If the primary reflection velocities are picked (e.g., velocity function of dashed blue line), primaries
           (P) are completely flattened after NMO correction while the multiples (M) are still hyperbolic and undercorrected. (B) If veloc-
           ities of multiples are picked (e.g., velocity function of dashed red line), then the multiples are flattened whereas the primaries are
           overcorrected.


           not easy to discriminate, since they may strongly  be applied after the dip moveout (DMO) pro-
           interfere. In shallow parts of the supergathers,  cess,  which  removes  multiple  enclosures
           even small variations in the picked velocities  appearing at the same zero-offset times in the
           result in erratic changes on the flattening of  semblance plots by solving the conflicting dip
           the events, and therefore the processors must  issues on the seismic data (Section 11.10).
           be extra careful when picking the velocities in  The same RMS velocity field, or interval
           shallow areas where primaries and multiples  velocity field derived from RMS velocities using
           interfere.                                   the Dix equation, is also used to migrate the seis-
              In a complete seismic processing, generally  mic data. Some of the migration algorithms are
           more than one velocity analysis is necessary.  quite sensitive to lateral velocity variations
           After sorting the data into CDP gathers, a first-  and may produce artifacts in different shapes
           pass velocity analysis is applied. Then, this  and sizes if there are issues or erratic lateral var-
           velocity field is used in multiple suppression  iations in the derived velocity field. Fig. 9.30
           since some of the multiple removal methods   shows an example velocity issue on a migrated
           require 2D or 3D velocity field to be known, such  section. There are two apparent normal faults
           as a Radon velocity filter. After removing the  on the migration section in Fig. 9.30A forming
           multiples, it is expected that the semblance plots  a small-scale graben, which occurs due to a
           will be greatly improved, and therefore a    gap in the RMS velocity field (Fig. 9.30B) because
           second-pass velocity analysis is applied to  of an incorrectly picked velocity function at the
           obtain a more appropriate velocity field. Fur-  location of the graben. Fig. 9.30C shows a
           thermore, a third-pass velocity analysis may  migrated section obtained by the correct 2D
   459   460   461   462   463   464   465   466   467   468   469