Go to the list of seismic processes.
Go to SIOSEIS introduction.
Photographs from the cruise.
R/V Ewing photographs.
Photographs of seismic equipment.
The EW9903 ODP site survey cruise aboard the R/V Ewing, Doug Wilson
and Alistair Harding Chief Scientists, used SIOSEIS in a similar manner
to that described on the EW9807 web example. The main difference was use
of a new Syntron recording system with 480 channels and a 6km streamer.
There were three onboard processing objectives:
1) Quality control.
2) Near realtime brute stack.
3) Tape copy and reformat.
The Syntron recording system wrote the MCS data onto IBM 3490E tapes.
With a 12 second recording length, a 2 millisecond sample interval, 480
recording channels and a 37.5 meter shooting interval, tapes were filled in
less than 25 minutes.
Unfortunately, the "offline" 3490E tape drive could not be attached to
the Ewing Sun Ultra Enterprise servers because they don't accept the 3490E
drive's fast wide differential interface. Neither HP computers brought from
Scripps could accept this interface either. The only machine with fast
wide differential was a Sun SparcStation 10, which also had a DLT, HP Dat,
Exabyte, 4GB disk, and was on the network.
The SparcStation 10 could not process the large quantity of data
efficiently, so the quality control and brute stack were limited to
using 1/4 of the traces and decimating to 4 mils. Another problem with
the SparcStation 10 was that it only had 32MB of memory and could not
run two SIOSEIS jobs at the same time. The SparcStation 10 is called
heezen.
Consequently, we had the 4GB disk on the SparcStation exported, so
that other machines could NFS mount it. All tape reading and writing
had to be done on the SparcStation since it had the tape drives.
Several subdirectories were created on the 4GB disk under the
/ldata/realtime. Subdirectory stack was used to read and stack the 3490Es.
The stack job also wrote every 50th shot into a circular disk file in
/ldata/realtime/shots. The stack job wrote the brute stack in a file in a
subdirectory of /ldata/realtime/stacked. Subdirectory /ldata/realtime/plots
had scripts to plot the shots and the the stacks. Another subdirectory,
/ldata/realtime/copy contained a sioseis script to reformat the SEG-D
3490Es to a DLT tape in SEG-Y format.
I added the following lines to my .cshrc to simplify processing:
set path=($path /export/sioseis/bin )
alias xl "xloadimage -r 90 sunfil &"
alias sd "/export/sioseis/doc/siodoc"
alias s "suntops < sunfil > psfil"
The tape reformat step was done after all the MCS data were collected.
While the reformat step theoretically takes the same length of time as the
collection step, it took much longer because manual intervention was required
for every tape.
The brute stack script has several features:
- The script was executed with the line name on it so the output
filename was unique.
- Read SEG-D tapes, discarding data before 4 seconds and after 8 seconds.
Only alternate traces between traces 1 and 479 were used and the data
were decimated to 4 mils. (A SparcStation 10 is not very fast!)
- Write every 50th shot to a "circular" file for QC (quality control)
- Assign streamer geometry
- Trace gather according to common midpoint
- NMO and MUTE based on Hydrosweep water depth
- Stack
- Disk output for other use
A subtle processing trick was used in the NMO process in order to
laterally vary the NMO velocity function. Earlier work in the area
showed that a velocity function could be designed for the area.
The velocity varied with water depth. Graham Kent had modified NMO
such that the parameter vtrkwb (velocity track water bottom) modifies
the normal SIOSEIS spatial interpolation by changing the RP number
into the water depth determined by Hydrosweep. The velocity function
was described by water depth so that the spatial interpolation is
done by water depth. SEGDIN has always put the Hydrosweep depth in
the SEG-Y trace header.
A simple brute stack script without spatially varying
the velocities converts the water depth to the water bottom time and
then "hangs" the NMO velocity function and mute from the water bottom
time.
The brute stack takes 15 minutes on the SparcStation 10, the same amount
of time as to collect it.
The stack plot script filters, gains, and creates
a SIOSEIS raster plot that is converted to an HP RTL file by
program SIO2HP and sent to the Ewing's HP DesignJet 755. It was
determined that the HP lp command did not work correctly with the
Ewing's plotter spooler by adding a character. The lpr command:
lpr -Pplotter hpfil
from the Ewing's Sun server (grampus) worked correctly though.
The shot plot script creates a Sun rasterfile and displays it
on the screen with xloadimage. A PostScript plot was made with:
suntops -w 8 -h 10.5 < sunfil > psfil
however 480 traces is quite dense. Running the plot script on some
other computer other than the SparcStation 10 worked well.
The copy script reformats the SEG-D 3490E tapes onto
SEG-Y DLT tapes. Each 3490E tapes is written as a SEG-Y file (process
output parameter newfile), even though the SEG-Y standard does not allow
multiple files on a tape. The output tape further violates the SEG-Y
standard by writing in IEEE format.
Both the brute stack and copy scripts utilize segdin parameter
OFFLINE, which rewinds and ejects the input SEG-D tape after it has
been read. This enables the "operator" to know that a new tape should
be "mounted". After the new tape is entered, a file named "in" is
created with the tape unit number int it. SEGDIN waits for the
creation of this file and starts reading the new tape. SEGDIN
then deletes the file "in" so that it waits at the end of tape for
yet another tape.
RP sorted gathers were written to a DAT tape using the gather script.
This script uses several convenient parameters. NFSKIP 41 was used in
process INPUT because there are 70 SEG-Y files on each DLT tape.
INPUT and OUTPUT parameter DEVICE was used because neither tape
drive was /dev/nrst devives.
Go to the list of seismic processes.
Go to SIOSEIS introduction.