Page 152 - Acquisition and Processing of Marine Seismic Data

2.6 QC IN DATA ACQUISITION 143

and an optional gain application. RMS ampli- • Maximum acceptable gun timing error is
tude displays of all available channels for each 1 ms.
shot can be used to locate suspicious shots, • Maximum acceptable air pressure variation
which can be then individually investigated. for each string is 2000 psi 100 psi.
Various QC applications, including raw/filtered • Maximum allowable gun array drop-out
data displays, RMS noise displays, power and f- tolerance is 1.5 dB.
k spectra of the data and producing brute stacks, • A minimum of three depth sensors per string
are performed to identify noise problems and to must be operational during the recording.
assess the compliance of the data based on var- • Maximum allowable source depth variation
ious acceptance criteria. averaged over the sequences is 0.5 m.
Analyzing of seismic data by an onboard pro- • Maximum allowable variation in source
cessing for QC implementations reveals if the separation is 4m.
data contains unacceptable problems. This is • Maximum allowable variation in string
done by a simple comparison of the observed (subarray) separation is 2m.
issues with the tolerance values determined • Maximum allowable variation in streamer
before the survey. Tolerance standards may separation is 10%.
change from one service provider to another. • Maximum acceptable feather angle value for
The following are example acceptance toler- any of the streamers is 10 degrees.
ances for seismic data:
In some specific cases, reshooting of an ini-
• Maximum acceptable bad records are 5% of tially collected line is necessary depending on
total shots for a line or for a segment. the offline QC results. Reshooting is done by col-
• A maximum of 10 bad records in 50 lecting the data in the same direction as the ini-
consecutive shots is acceptable. tial survey, generally with a 10 shot points
• Maximum three consecutive bad records are overlap, and is performed anytime during the
allowable. survey. In addition, if a seismic line is aborted
• Maximum allowable bad records are 5% of within 10 km of the line start, it is totally reshot.
total shots per line. Reshooting of a line takes place whenever the
• Maximum acceptable dead traces are 4% of previous example situations are observed dur-
the total channels per streamer. ing QC analyses.
• Maximum 5 consecutive dead traces are Analyzing the amount and specific types of
acceptable for a shot record. noise embedded in the seismic data is a crucial
• Maximum acceptable dead traces are 2 for 20 part of the online and offline QC implementa-
front channels of each streamer. tions. At the start and end of each line, at least
• 1 m streamer depth variation is acceptable one noise record is acquired and analyzed in
for maximum 30 consecutive shots. terms of ambient background and operational
• 1.5 m streamer depth variation is noise levels both on shot records and RMS
acceptable for maximum 20 amplitude plots, which allows us to identify
consecutive shots. noisy traces for further evaluations. An example
• Maximum allowed positioning error for front noise record analysis is given in Fig. 3.2. During
of the streamers is 2 m. the offline QC analyses, the following noise
• Maximum allowed positioning error for gun types are specifically reported in the per-line
strings is 2 m. QC reports:
• Maximum allowed positioning error for tail
• Streamer bend noise
buoys is 5 m.
• Operational noise