NAME
qest - perform Q estimation on seismic traces
SYNOPSIS
qest [ -Nntap ] [ -Ootap ] [ -O2ftap ] [ -Aatap ] [ -rsrs ]
[ -rere ] [ -wlwin ] [ -fccut-off frequency ] [ -ththreshold
] [ -nullnull ] [ -smoothsmoothing on ] [ -ordlmord ] [ -M ]
[ -top ] [ -cen ] [ -bot ] [ -R ] [ -PF ] [ -PA ] [ -L2 ] [
-h ] [ -V ] [ -? ]
DESCRIPTION
qest computes an estimate of effective Q from seismic data.
Output from the program consists of the estimated Q and the
peak, or dominant, frequency for each time sample. The
methodology of the procedure is based on a method of spec-
tral ratios in which the spectral estimates are made at each
time sample. This procedure employs either a maximum
entropy algorithm or a recursive DFT to compute the spectral
estimate, converts the power spectrum to dB, finds the dom-
inant frequency, computes the Q estimate from the slope
(using only those spectral values which exceed a specified
threshold) of the spectrum relative to that frequency, and
assigns the estimated Q value and dominant frequency to the
center of the window for output.
The data input to qest must have proper mutes applied, e.g.
the water column data should contain zeroes. This prevents
bogus Qs from influencing the deeper data.
Description of the basic theory behind the procedure can be
found in the references given at the end of this man page.
Normally effective Qs are expected to increase with depth
and qest defaults to output these values. But the option is
provided to turn off the time weighted portion of the Q cal-
culation. This will give a better representation of relative
changes in Q with time or depth. The option also exists in
program qint to take effective Qs and unwrap the time
weighted function. It is recommended that the latter course
be taken if the user wishes to view relative Qs since run-
ning qint is much faster than rerunning qest.
Data preparation is a key element to successful analysis
using this program. The data may be stacked and migrated or
may be input in prestack form. Mixing the data prior to
analysis is advisable, but other data processing should be
consistent with AVO processing, with appropriate scaling,
deconvolution, and multiple removal. If the data has been
processed with AGC or with a DAFD-like process, do NOT
expect significant results.
Output from program qest consists of three (3) data sets
(two are optional). The first data set contains the Q esti-
mate for each time sample. The second [OPTIONAL] data set
contains the peak frequency values. The third [OPTIONAL]
data set contains the peak amplitude values. Piping of the
first (Q estimate) data set only is permitted unless run
inside XIKP. Inside XIKP the peaf frequency data (and peak
amplitude data) are sent to units 3 (and 4) respectively on
the qest XIKP panel. A very close approximation of the
second data set, if desired, can be derived by computing and
heavily smoothing the instanteous frequency.
As the Q traces are output the time to the first nonzero
sample is placed in the restore mute trace header word
VPick1 for subsequent optional application in the inverse Q
filter program qifl. This is because subsequent Q smoothing
could introduce funny Q values at this transition zone (this
could cause the inverse Q filting problems). This disturbed
zone will exist above the mute line so it can be corrected
by flagging qifl with a -R which will zero out the Q traces
down to the VPick1 time.
Note: this code has been entirely recoded and is consider-
ably faster. The new default mode is the fourier spectral
estimator (not MEM). Also there are options for slope fit-
ting that were not in the old code.
qest gets both its data and its parameters from command line
arguments. These arguments specify the input and output
file names, the starting and ending records to process, the
analysis window length, and the bandwidth over which to pre-
form the regression for slope computation.
Command line arguments
-N ntap
Enter the input data set name or file immediately after
typing -N unless the input is from a pipe in which case
the -N entry must be omitted. This input file should
include the complete path name if the file resides in a
different directory. Example -N/b/vsp/dummy tells the
program to look for file 'dummy' in directory '/b/vsp'.
-O Q estimate Output
Enter the name of the Q estimate data set immediately
after typing -O. This output file name is not required
if piping to another process. The output data set,
when specified, also requires the full path name (see
above).
-O2 Peak Frequency estimate Output
Enter the name of the [OPTIONAL] peak frequency data
set immediately after typing -O2. This data set may
NOT be piped to the output except when run in XIKP in
which case this is output socket #3. The data set,
when specified, also requires the full path name (see
above). See -PF flag below.
-A Peak Frequency estimate Output
Enter the name of the [OPTIONAL] peak amplitude data
set immediately after typing -A. This data set may NOT
be piped to the output except when run in XIKP in which
case this is output socket #4. The data set, when
specified, also requires the full path name (see
above). See -PA flag below.
-rs rs
Enter the sequential index of the first record to be
processed. (Default = first).
-re re
Enter the sequential index of the last record to be
processed. (Default = last)
-w lwin
Enter the length, in ms, of the analysis (DFT) window.
Note: if you want to attempt to observe changes in pore
pressure, use a window length greater than 300 ms
(e.g., 1000 ms). (Default = 300 ms).
-fc cut-off frequency
Enter the maximum frequency, as a fraction of the
Nyquist frequency, at which to truncate the regression
for the slope of the power spectrum. The value entered
may not exceed 1.0.(Default = 1.0)
-th threshold
Enter the threshold (in unsigned dB) below which values
in the power spectrum are ignored in the linear reqres-
sion. For example, if you want to include spectral
values no more than 18 dB below the maximum, specify
the threshold ast 18. (Default = 12.)
-null null
Enter the value of Q to be placed in the trace mute
zone. The trace mute zone contains hard zeroes and so
will cause the spectral calculations to blow up and if
replaced with zeroes might cause downstream Q smoothing
operations to be compromised. Default = 500
-smooth smoothing on
If present, the output Q estimate and Peak Frequency
traces are vertically smoothed with a 15-point running
average operator. If this flag is not present, that
smoothing is turned off. (Default = on)
-ord mord
Enter the order of the maximum entropy calculation. The
default is 2 which means the spectra is dominated by a
single peak. Each additional peak in the spectrum will
add 2 to the order. Unless careful testing of orders is
done using program mesa then the default should be
used.
-M Enter the command line argument -M to use the maximum
entropy method of sprectral estimation, otherwise the
fourier method is used. Spectral computations within
the sliding window based on the maximum entropy method
tend to minimize the effects of the analysis window.
This fits an autoregressive (AR) recursive model to the
data. It is most effective when the AR order is low.
The order is always a multiple of two and if the order
is divided by two the result gives an idea of the max-
imum number of possible peaks in the spectrum.
-top Enter the command line argument -top to hang the output
Q values from the top of the analysis window (default).
-cen Enter the command line argument -cen to hang the output
Q values from the center of the analysis window.
-bot Enter the command line argument -bot to hang the output
Q values from the bottom of the analysis window.
-R Enter the command line argument -R to output the raw
(relative) Qs; the default is to out time weighted
(absolute) Qs. The data input to qest must have proper
mutes applied, e.g. the water column data should con-
tain zeroes. This prevents bogus Qs from influencing
the deeper data.
-PF Enter the command line argument -PF to output the
optional output peak frequency data set.
-PA Enter the command line argument -PA to output the
optional output peak amplitude data set.
-L2 Enter the command line argument -L2 to use the least
squares slope fitter; otherwise the L1 norm fitter is
used.
-h Enter the command line argument -h to get online help.
The program terminates after the help screen is
printed.
-V Enter the command line argument -V to output verbose
printout.
-? Enter the command line argument '-?' to get online
help. The program terminates after the help screen is
printed.
EXAMPLE
qest -Nindata -w300 -fc.45 -Oodata
where the sliding window is 300ms, the cutoff frequency is
.45 Nyquist, and we use the default maximum entropy spectral
method.
REFERENCES
Jacobson, R.S., Shor, George G. Jr., and Dorman, LeRoy M.,
1981, Linear Inversion of Body Wave Data - Part II: Attenua-
tion versus Depth Using Spectral Ratios, Geophysics, Vol
46, pp 152-162.
Stainsby, S.D., and Worthingto, M.H., 1985, Q Estimation
from Vertical Seismic Profile Data and Anomalous Varaiations
in the Central North Sea, Geophysics, Vol 50, pp 615-626.
Bickel, S.H., and Natarajan, R.R., 1985, Plane-wave Q Decon-
volution, Geophysics, Vol. 50, pp 1426-1439.
AUTHOR
Richard Crider E&PTG/IP Houston, Paul Gutowski EPTG
Tulsa
COPYRIGHT
copyright 2001, Amoco Production Company
All Rights Reserved
an affiliate of BP America Inc.
Man(1) output converted with
man2html