Go to the list of seismic processes.      Go to SIOSEIS introduction.
SIOSEIS was used aboard the R/V Ewing during the EW0008 cruise to
Blake Ridge (September 2, 2000 to October 17, 2000; Newark, N.J)

Step 1:
Reformat SEG-D 3490 to SEG-Y DLT script includes assigning the
streamer geometry (trace ranges and CMP numbers) as well as extracting
the GPS clock, LDEO shot number, and streamer depth.  The script also
writes every 20th shot to a circular disk file of QC plotting.
   This script was run immediately after each 3490 tape was ejected
from the Syntron acquisition system.  Each tape was read another time
after the copy/reformat was completed (see step 2 below).
   This script had to be changed each time the streamer was deployed
since the tow leader and streamer bird locations changed each time
the streamer was deployed.
   During EW0008 a new feature was added to SIOSEIS process GEOM so
that the shot to shot distance can be determined on each shot using
the realtime GPS locations that are in the SEG-D and SEG-Y headers.
EW0008 leased a realtime differential GPS system  (diagram)  from
John E. Chance (written description).   While in Port Newark, this
system performed within our 1 meter specification.


Step 1a:
QC "plot.shot" script was run occassionally to check on the data
quality.  Generally, this plot was used to check on noisy channels.
In one case this plot showed that alternate shots were dead before the
watch recognized the problem.

Step 1b:
The frequency analysis script was run in the early part of the cruise
to show the effect of different tow depths of the streamer and guns
as well as to reassure ourselves that a ,002 seconds (2 mil) sample
interval was sufficient. e.g.
freq_tr.2mil /export/home/EW0008/Realtimestack/latest.shot.segy 301
    This frequency plot shows that the data is dominated by steamer
noise, which can be filtered out using a (sioseis script).
    The frequency plot of the filtered trace shows the predominant
frequency to be around 40Hz.

Step 2:
The brute stack script (version 1) and (version 2) has many unique
features.  The most important objective was to provide a hardcopy
plot for geologic and geophysical purposes as well as another
quality control step.  This script generates a plot file which is
constantly being read by program atlantek that writes to the
Atlantek thermal plotter.  The brute stack plot thus contain data that
were collected around one hour previously.
    All traces are read by SEGDIN, assigned streamer geometry and
gathered according to cmp so that every 50th gather could be written
to a disk file for velocity analysis.  New parameter LIMITS was used
in process TREDIT to exclude all ranges except for the 1000 meters
closest to the ship because the velocities were not known well enough.

     Before starting the velocity analysis, a quick plot was made
of a few cmp gathers.   The script that generated the plot has a
couple of interesting twists.  New plot parameter RECSP YES was used
so we didn't have to worry about the number of traces in each gather.  
Process header was used to convert the Hydrosweep water bottom
time from seconds to milliseconds because process plot can only
annotate integers.
     This vpick script runs on Matlab version 5.3.0 which uses
function ginput rather than getline.  These velocities were checked
by plotting each moved-out cmp gather individually using a C-Shell
script.  These picks weren't satisfactory when used in the stack.
There was only one machine with Matlab on the ship and the license
was for a single user and other users were using Matlab.  So, we used
the constant velocity stack method provided by SIOSEIS (scripts).
     The stack script was short and simple, with only 5 velocities
necessary.  NMO velocity interpolation parameter vintpl 3 should be
used for smooth velocity interpolation.  The mute parameters were
selected to mute out the far half of the streamer down through the
BSR because of NMO stretch.  The velocities generally increase, except
for the BSR.  e.g.
fno 309701 vtp 1510 3.65 1527 3.9 1535 4.15 1515 4.2 1555 4.25 1570 4.5
               1590 4.85 1650 6 end
where the BSR is .55 seconds after the water bottom.
      The stacked data were then FK migrated with a velocity of 1520m/s,
reversed in order (since the line runs from west to east), the water
bottom picked, the muted to water bottom, filtered, and an exponential
gain applied from the water bottom.  The script uses a brand new
parameter in process gains; TADD.  e.g.
gains
  subwb yes tadd 3.  type 3  alpha 4 end end
subtracts the water bottom time of each trace and then adds 3 seconds.
Thus, each trace gets exactly the same gain function, as if each trace
started at 3 seconds.

     A comparision of realtime and post-processed navigation
was performed.  The LDEO post-processed ts.n file was converted to 
SIOSEIS navigation file using PERL script ts2sio.  This comparison
was done on a relatively calm day.  Many factors should be considered
when deciding which method to use for cmp binning:
The realtime nav block is a predicted location.
The realtime location has more ship pitch, roll, yaw in it.
The guns and streamer are somewhat decoupled from ship pitch, roll, yaw.


*******************************************************************
**************        RECOMMENDATION        ***********************
*******************************************************************
TS.N FILES SHOULD BE USED FOR POST-PROCESSING NAVIGATION AND BINNING.



Return to SIOSEIS examples.

Go to the list of seismic processes.      Go to SIOSEIS introduction.