Taiwan: Sort, Velocity Analysis, NMO, Stack
Returning to the Taiwan data, we now have 5 seconds of data, we zeroed the bad hydrophones, applied spherical divergence correction, and band-pass filtered the shot gathers. The result is Tshot5kgf.su. We will now
sort these shot gathers to CMP order,
analyze the CMPs for stacking velocities,
use the stacking velocities to apply NMO, and
stack the CMPs.
Sorting to CMP order is a one-line command:
We send the sorted data to “Tdata” since this file is just as large as the file of windowed (0-5 seconds) shot gathers: 133 Mbytes.
Below is the surange output of Tcmp5kgf.su.
We have 790 CMP gathers (4816-4027+1). Figure 15.1 shows that the full fold gathers (at least 34 traces per gather) are approximately 4175-4700. (We describe the program, sukeycount, which generated the data for Figure 15.1 in Appendix C.) Despite the constant number of traces per shot, CMP fold varies (Figure 15.1) because ship speed and streamer position varied.
Below is the user area of iva.sh changed for the Taiwan data. To make reasonably detailed velocity analysis at a regular interval along the line, we choose to analyze every 25th CMP starting with CMP 4175 (lines 13-18). We set the perc value to 90 (line 31), an accommodation for real data. Semblance values go from 1000 m/s (line 38) to 5000 m/s (nvs*dvs + fvs). CVS panels range from 1000 m/s to 5000 m/s (lines 43-46).
We make one more accommodation for real data. When we use suximage following suvelan, we set perc=97