Enhancements Bug
and Fixes 330
Introduction
The purpose of this page is to notify users of the Computer
Program in Seismology (CPS) package of repairs or extensions made
to the code. Whenever this page is updated, the down loadable
version of CPS is updated at the download FTP site.,
December 15, 2012
- Corrected compile warnings in VOLIII/src/sdisp96.f,
VOLIII/src/tdisp96.f, VOLIX/src/cprep96.f, VOLIX/src/cpulse96.f
and VOLVIII/src/sacfilt.f
December 10, 2012
- Modified surf96, rftn96 and joint96.
Justin Ball of the University of Colorado was using the
'negative' layer thickness flag but this choice was not
maintained for the temporary output files. The 'negative'
layer thickness is a way to set the reference level and to
control the assignment of depths. It introduced for the
atmosphere - solid earth problem. Rather than define source and
receiver positions with respect ot the top of the model, one can
define the reference level as the atmosphere - solid interface,
and, as a consequence, positive values of source or receiver
depths indicate positions in the solid while negative ones
indicate elevations above the surfece.
While fixing the code, I use the opportunity to modify the
dispersion display program srfphv96 so that plots of
Love and Rayleigh dispersion use the same scales:
 |
 |
- Modified genplt to permit placing a legend within the
plot. This is implemented using two new command line flags: -L
lcmdfil -LPOS position. The command line syntax now
has
-C cmdfil
-A acmdfil
-L lcmdfil
where the following are examples of the different files:
cmdfil:
'HAM.RDD.val' 1 0.01 'CI'
'CUS.RDD.val' 1000 0.01 'CI'
'R41.RDD.val' 1025 0.01 'CI'
'S42.RDD.val' 1050 0.01 'CI'
'S43.RDD.val' 1075 0.01 'CI'
acmdfil:
'HAM.RDD.val' 1 0.01 'CI' 0.04
'CUS.RDD.val' 1000 0.01 'CI' 0.04
'R41.RDD.val' 1025 0.01 'CI' 0.04
'S42.RDD.val' 1050 0.01 'CI' 0.04
'S43.RDD.val' 1075 0.01 'CI' 0.04
cmdfil:
'HAM.RDD.val' 1 0.01 'CI' 0.04 'HAM'
'CUS.RDD.val' 1000 0.01 'CI' 0.04 'CUS'
'R41.RDD.val' 1025 0.01 'CI' 0.04 'R41'
'S42.RDD.val' 1050 0.01 'CI' 0.04 'S42'
'S43.RDD.val' 1075 0.01 'CI' 0.04 'S43'
|
 |
The postion parameter is a string of the form 'TL', 'ML', 'BL',
'TR', 'MR' or 'BR' for top-left, middle-left, bottom-left,
top-right, middle-right and bottom-right, respectively. The
legends plotted to the right are indented so that the entire
string fits within the image frame.
Note that one can use all three commands. To make the plot
shown, the following command line was used:
genplt -XMIN 100 -XMAX 1000 -YMAX 0.0001 -YMIN 0.000001 -XLOG -YLOG -L cmdfil -TX 'distance (km)' -TY 'Filtered Velocity' -LPOS 'BL'
- Modified do_mft to be more efficient and added
additional command line flags.
The new set of command line flags are as as follow:
-G The default name of the final dispersion file is StaCmp.dsp.
With this flag, the output dispersion file is the original filename with
.dsp appended, e.g., NMSLMBHZ01.dsp
-T Show the 'Tomo' button
-DMIN dmin only use traces with distances ≥ dmin km
-DMAX dmax only use traces with distances ≤ dmax km
-11MIN 11min only use traces with IHDR11 ≥ 11min
-11MAX llmax only use traces with IHDR11 ≤ 11max
When processing there are several steps that are performed on
different pages: select the trace on page 1, set the units on
page 2, set the multiple filter analysis and display
parameters on page 3 and then select the dispersion on page 4.
Then repeat the entire process for the next waveform. Normally
the parameters on pages 2 and 3 are not changed. The new Autofeed
button permits one to skip pages 2 and 3. This small change
saves a lot of time and mouse movements when working with
large data sets.
The second change was specifically written for use with
inter-station Green's functions computed by cross-correlating
ground noise. Since gsac is used, the stack
function saves the number of traces in the ihdr11
variable. The command line of do_mft can be used to
use only traces with a range of ihdr11 values. The
command line can also select distance ranges now, a feature of
use for earthquake data. Finally a new -T flag
introduces a Tomo button.
The Tomo button executes the script MFTDOOVERLAY
that is in the users PATH. It is currently used to determine
the tomographically predicted dispersion between the station
and epicenter for earthquakes or two stations for
cross-correlation, and then to overlay this onto the MFT
image. The two figures show the buttons and the predicted
dispersion.
The purpose of making this overlay is to assist in
determining the proper period range for dispersion. For
earthquake data, the spectral shape is often used as a guide,
but the comparison may be useful. Another obtain is to overlay
model predicted dispersion curves created from the -S flag of
sdpegn96. The script MFTDOOVERLAY can be
programmed to do a lot of things. For example a combination of
gsacGMT can be used to create the file map.eps
which can be viewed using gv or other PostScript view:
Tomography results and display scrips will be provided
shortly.
- The current release is NP330.Dec-10-2012.tgz.
October 22, 2012
- Corrected an error in the gsac
interpolation routine in gsac_in.c. The problem was a spurious value
for the last point. Shaoqian Hu fixed the error. The following
image compares the original trace in red to the interpolated
trace in blue. In order to make the presentation, the observed
trace was read using a cut, and the cut version saved. The cut
trace was then interpolated and saved. Finally both traces were
red using the command cuterr
fillz and cut
with wider limits to focus on the behavior at the end. Thus is
why both traces are zero at the end. The problem has long been
noticed in waveform plots from the moment tensor inversion -
however since only one point was involved, the moment tensor
inversion results do not change.
Original
code
|
Corrected
code
|
|
|
September 20, 2012
- Upgraded the distributed IRIS program rdseed
from 5.2 to 5.3. In order to support the CPS
architectures, e.g., CYGWIN (Windows), MacOS-X, LINUX and
Solaris, changes were made to these codes. The changes are given
in the links in this section.Only minor changes are made. The
comments in the Sac pole-zero files are a superset of rdseedv5.3
and rdseedv5.0.slu.
- Main/output_resp.c
- modify to permit extra output for use in seismic
network QC
- Main/output_sac.c -
modified comments in Sac pole-zero file
- Main/summary.c -
changed the call to fabsl
to fabs since there
was no fabsl in
CYGWIN and since the use of the code only required double and not long double.
- Parsers/parse_50.c - eliminate
non-printing characters in the site name
- Utilities/log_errors.c -
add #include
<string.h>
- Detailed changes
- Modified the gsac
ppk to recognize the
use of the 'm' key as the same as the '*' key to make the trace
amplitudes larger.
- The current release is NP330.Sep-20-2012.tgz.
September 9, 2012
- The include file calplot.h
now has prototypes for the routines void gcont(float x, float y);
void gmove(float x, float y); (Thanks to Shaoqian Hu)
- Fixed the bsort and
dbsort subroutines in sprep96.f, sdisp96.f, tprep96.f, tdisp96.f, scomb96.f and
tcomb96.f. The problem
was that the bsort routine sorts and array, and then keeps only
the unique values. Thus the length of the array returned can be
smaller than the initial array. The last line of the
subroutinhe was written as n=n-m instead of nn=n-m. (Thanks to ruoshan at USTC)
- The current release is NP330.Sep-09-2012.tgz.
August 6, 2012
- Fixed an error in the Setup script so that the rdseed now compiles under
Cygwin. The current release is NP330.Aug-06-2012.tgz
July 26, 2012
- hpulse96 now uses list
direction IO instead of formatted IO for user defined source
time function
- The program
srfker96 is now part of the package.
This program reads the temporary files from surf96 to list all partial
derivatives as a function of depth. A tutorial on the use
of this program is available at the tutorial web page http://www.eas.slu.edu/eqc/eqc_cps/TUTORIAL/
.
- After creating the above tutorial, I realized that I cannot
compute the partials of the anelastic attenuation coefficient γ
with respect to changes in P- and S-wave velocity or
layer thickness. These partials created in subroutine
PROGRAMS.330/VOLIV/src/jsamat.f now set dgdv = 0 for
each layer.
- The furrent distribution is NP330.Jul-26-2012.tgz
April 9, 2012
- Upgraded the distributed IRIS program rdseed
from 5.0 to 5.2. In order to support the CPS
architectures, e.g., CYGWIN (Windows), MacOS-X, LINUX and
Solaris, changes were made to these codes. The changes are given
in the links in this section.Only minor changes are made. The
comments in the Sac pole-zero files are a superset of rdseedv5.2
and rdseedv5.0.slu.
- Main/output_resp.c
- modify to permit extra output for use in seismic
network QC
- Main/output_sac.c -
modified comments in Sac pole-zero file
- Main/summary.c -
changed the call to fabsl
to fabs since there
was no fabsl in
CYGWIN and since the use of the code only required double and not long double.
- Parsers/parse_50.c - eliminate
non-printing characters in the site name
- Utilities/log_errors.c -
add #include
<string.h>
- Detailed changes
March 26, 2012
-
Changes to
gsac:
-
outcsv
command adds a
column header which now indicates Freq(Hz) for spectral
files in addition to the previous Time(sec) for time
series.
- prs (plotrecordsection)
permits annotation using the title command. However this is not perfect
since the location of a title at the top of the page conflicts
with the annotate option
of prs and the
location on the left conflicts with the y-axis label
- prs command now
behaves better with the use of the prs title
title_text command to supersede the normal
axis label. The annotation is required since the Sac header
only permits a few of the variables to have associated text
strings, and if the USER1 happens to be set for a user defined
Angle, one can arrange the record section accordingly with the
command prs user1 title
"Angle (deg)"
- tregn96,
tdisp96, sregn96, scomb96, tcomb96: cleaned up compiler warnings
- hspec96: Fixed a
problem with flagging a receiver position is a fluid to compute
the pressure fields in the fluid for the different sources. This
only appeared with an atmosphere model was used which required refdep to be other than 0.0
- tspec96: Fixed a
problem with flagging a receiver position is a fluid to compute
the pressure fields in the fluid for the different sources. This
only appeared with an atmosphere model was used which required refdep to be other than 0.0. Hoever I have
to check the travel times for a fluid.
February 8, 2012
- Modified genplt to
have additional plotting controls on the size for the symbol
shape. This means that the other program tgenplt is deprecated. The
syntax of the current program and an example of its use is given
in the line GENPLT/nindex.html.
genplt will run as
before. The new capabilities are invoked with the -A acmdfil command-line
option.
- The current distribution of Computer Programs in Seismology is
now NP330.Feb-09-2012.tgz
January 14, 2012
Increased the size of the outstr
string in do_mft
to prevent overflow with long file names
November 23, 2011
- Fixed an error in the synchronize
command of gsac. The problem arose
when reading just the headers and then synchronizing. The
code has been fixed. The current version of gsac is now V1.1.44 21 NOV 2011.
- The command line
parsing of the program fmplot
was rewritten,
- The current distribution of Computer Programs in Seismology is
now NP330.Dec-01-2011.tgz
November 1, 2011
- In response to user requests the following changes were made
to gsac. The current
version of gsac is now
V1.1.43 31 OCT 2011.
- At the request of H. Benz, gsac
now sets a header value corresponding to the time of the
maximum (depmax) and
minimum (depmin)
amplitudes. The previously unused float/real header
values in the 70 position real header, FHDR65 and FHDR66, are
now called TIMMAX, TIMMIN, respectively. This request was made
to use the times of the maximum and minimum points of a trace
in external programs.
- These other codes were changed to reflect these new
parameters: scmxmn
routine in SUBC/sacsubc.c,
SUBS/sacsubc.h, SUBS/sacsubf.f . The programs that
use these general Sac I/O routines were re-compiled.
- A new command, ricker,
was added to gsac.
This permits one to convolve a time series with a Ricker
wavelet of a given frequency. This was implemented because of
difficulties with the spulse96
-f wavelet_file and hpulse96 -f wavelet_file implementation.
- The current use of Sac header variables by Computer
Programs in Seismology is as follows:
- USER0 - Gaussian filter parameter, alpha, used in saciterd, saciterdc, rftn96 and joint96
- USER1 - minimum period of filter. This is introduced so that
users do not falsely accept dispersion values outside the
bandpass of a filter
- USER2 - maximum period of filter.
- USER4 - ray parameter for joint96
and rftn96. Set
from command line by saciterd
and saciterdc.
- USER5 - percent fit value. Appears in pltsac, wvfgrd96, wvfmt96,
wvfmtd96, hrftn96, trftn96,
joint96 and rftn96.
- USER9 - time shift for waveform fit. Used by pltsac and set
by wvfgrd96, wvfmt96 and
wvfmtd96.
- TIMMAX - offset with respect to reference time indicating
time of trace maximum
- TIMMIN - offset with respect to reference time indicating
time of trace minimum
- IHDR20 - set by gsac
as part of the ppk qc
command to indicate whether a trace passed QC. I obtain
this value from a sac file in a shell script with a call to saclhdr, which is then
use in an awk script.
- The current version of Computer Programs in Seismology is NP330.Nov-01-2011.tgz.
October 20, 2011
- C. Buffoni noted a problem with the presentation shaded area
seismic trace plots in PostScript using the command plotnps. It took a while to
track this down, but the problem was repaired. In addition the
shaded area plot correctly for the command plotxvig -R, although this
would rarely be used. In addition plotgif -R works correctly.
OLD
(problem)
|
NEW
(Repaired)
|
|
|
- Modified the CALPLOT functions pend() and gend()
to reset the pen width to the default 0, and the pen to 1
(black). This assists merging files.
- NOTE:
The S-wave receiver function, e.g., Z/R, computed by hrftn96 is not
stable. This may be due to the problem of deconvolving
with a wavelet that is not minimum phase. This is a major issue
that is deferred to a future update.
- The current version of Computer Programs in Seismology is NP330.Oct-27-2011.tgz.
October 13, 2011
- The mt (moment
tensor) command of gsac
now supports the specification of the moment tensor source on
the command line, e.g.,
- mt to ZRT MXX 0.0000000 MXY -1.00000002E+20 MXZ 0.0000000
MYY 0.0000000 MYZ 0.0000000 MZZ 0.0000000 AZ 25. FILE
/backup/rbh/PROGRAMS.310t/GREEN/CUS.REG/0100/030000100
- The current version of gsac
is V1.1.42 13 OCT 2011
September 23, 2011
- Major change in gsac:
- refr command - filters are 2-pole 1 pass and not 2-pole 2
pass - thus they are causal!!
- plot command - for many traces displayed sequentially using
the perplot command the y/n/b query is now y/n/b/q so that one
can quit
- mt command - fixed code so that the t0 and t1 arrival times
of SV and SH are correctly read and then saved in the headers.
This arose during conversion of horizontals to N E
- prs command - do not annotate trace if KSTNM is -12345 and
add LAST kolor FIRST color to permit different colors. I
wanted to compared one observed trace to many predicted traces
using different colors for observed and predicted.
- csstime.c - MAJOR change. I used an incorrect rounding when
going from epoch time to integers NZYEAR NZJDAY NZHOUR NZMIN
NZSEC and NZMSEC. This inadvertently messed up the seconds by
one unit.
- Minor changes to sacmat96,
sepegn96.
- Changes to the VOLIV surface-wave inversion programs to permit
computation of du/da and dc/da, partials with respect to P-wave
velocity. these are not used in the surface-wave of joint
inversion; the changes were made for the design of a new
inversion of P, S first-arrivals and dispersion for shallow
exploration
- hrftn96 in VOLVI
changed the sign of the Z and R synthetics for SV incident to
agree with expected particle motion.
- In VOLVII, wvfmt96,
wvfmtd96 sets USER5 to the percentage of fit,
e.g., 100 * correlation coefficient squared
- In VOLVIII/src changed output format for the file
elocate.sum in elocate
- In VOLVII/src introducted new program saciterdc which is C
version of the FORTRAN saciterd for speed, and easy
transliteration of routines to java. The -Q flag suppresses
output at each iteration
- in SUBS put in an error check for the routine bwsac in
the file sacsubc.c
October 24, 2010
- Fixed bugs in the interpolate function of gsac which affected the
convolve comand. The current version of gsac is GSAC - Computer Programs in
Seismology [V1.1.39 24 OCT 2010].
- The current release of Computer Programs in Seismology is
NP330.Oct-24-2010.tgz.
October 12, 2010
- The current release of Computer Programs in Seismology is
NP330.Oct-12-2010.tgz
- Corrected Makefiles for OSX in VOLVIII/gsac.src/ so that
arrow keys work for gsac
on OSX
- Modified the Setup script so that the help reads:
Usage: Setup SOL WIN32 SOL-EGCS LINUX LINUX64 CYGWIN OSX
SOL SUN Solaris Compilers
SOL-GNU SUN Solaris with gcc/g77 compilers
CYGWIN CYGWIN 98/NT/2K/XP gcc/g77 Compilers
CYGWIN40 CYGWIN 98/NT/2K/XP gcc/gfortran Compilers
LINUX40 Linux with gcc/gfortran compilers
LINUX6440 Linux 64 bit with gcc/gfortran compilers
OSX40 Apple with gcc/gfortran compilers
OSX40-32 Apple with 32-bit gcc/gfortran compilers
Older systems - not modern
LINUX Linux with gcc/g77 compilers
LINUX64 Linux 64 bit with gcc/g77 compilers
OSX Apple with gcc/g77 compilers
The purpose of the OSX40-32 flag is
to force a 32 bit compile. This is useful on some Mac's where the
gfortran and gcc compilers such that one is 32 bit and the otehr
64 bit
- Fixed hspec96 by
adding a -ZSRC zsrc and -ZREC zerc flag that will define the
maximum depth of the source/receiver location models to patch
onto the teleseismic model. Before this change the
halfspace was replaced by a 1 km thick layer, which messed up
the resulting model. Creating a merged model is not simple
scientifically irrespective of the programming error.
October 3, 2010
-
Upgraded the distributed IRIS programs evalresp from
version 3.2.40 to 3.3.3 and rdseed from 4.8 to
5.0. In order to support the CPS architectures, e.g.,
CYGWIN (Windows), MacOS-X, LINUX and Solaris, changes were
made to these codes. The changes are given in the links in
this section. This upgrade was required in order to continue
to use the most recent SEED volumes.
The two programs are installed in PROGRAMS.330/bin
with the executable nanames evalresp and rdseed.
The IRIS distribution found in PROGRAMS.330/IRIS/rdseedv5.0
provides the executables rdseed.linux (64 bit), rdseed.MAC.x86
(64 bit), rdseed.powerpc, rdseed.sparc (64 bit),
and rdseed.sun.x86 (64 bit). If you wish to use these
programs, you must install them in your system and adjust the
path to find them.
-
evalresp - The
SLU version has the fixes suggested by Eric J. Haug of
SLU in May, 2008 to handle continuation for FIRST
with a large number of coefficients. Basically we need to
get past the tendency of some SEED writers to put the
following phase in the RESP file "No Abbreviation
Referenced" for Response in units lookup and Response out
units lookup. This is a proble with the CMG-5TD
accelerometer which has a very, very FIR stage.
Detailed changes.
September 30, 2010
- Modified gsac to
implement the convolve command.
This required significant effort because positioning and time
delays are a problem. The syntax of the new command differs from
that of Sac and is:
Con Z [B|O|A|T0|T1|T2|T3|T4|T5|T6|T7|T8|T9] File external_sac_file [Suffix suffix]
One
marker
must
be
set
to
define
the
zero
time
of
the
pulse
in
order
to
prevent
incorrect
delays.
The
following
example
creates
an
impulse, sets "O" to the start of the impulse, convolves with the
internal function "Triangle" and then uses the "Convolve" command
to do the same:
gsac << EOF
#####
# define the filter
#####
fg imp delta 0.1 npts 256
w imp.sac
r imp.sac
# set the o marker
ch o 0
w
triangle w 1.0
w tr.sac
#####
# define the impulse with a different sample rate
#####
fg imp delta 0.05 npts 512
w nimp.sac
r nimp.sac
conv z o f tr.sac s ".con"
w
q
EOF
A comparison of the files tr.sac and nimp.sac.con shows that
they are the same and have the same spectral amplitudes, except at
high frequency because of the different sample rates.
- The current version of gsac
is now V1.1.38 30 SEP 2010.
September 20, 2010
- Fixed some declaration errors in sacpsd
September 17, 2010
- Introduced the simple plotting program genplt
- Added a new program wvfdly96 for use in
the source inversion. This program interprets the time shift
used for regional moment tensor inversion to a change in origin
time and location. This was necessary when analyzing events
whose locations are poorly constrained
August 30, 2010
- Surface-wave programs in VOLIII/src were rewritten: sregn96, spulse96, slegn96, sdpegn96, sdpdsp96, slat2d96, sdisp96, sdpder96 and sdpsrf96. The purpose for
the rewrite was to align the code with the development in the
class notes
- Added new programs to VOLIII/src for transverse-isotropic (TI)
media. The new programs provide TI synthetics and phase-velocity
partials with respect to the four vertical and horizontal P- and
S-wave velocitys and the parameter η = F/(A - 2L). These new
programs are called tcomb96,
tdpder96, tdpsrf96, tprep96,
tsregn96, tdisp96, tdpegn96, tlegn96, tpulse96 and tregn96. The programs
work together in the same manner as the original codes for
isotropic models:
tprep96 -M model.TI -L -R -NMOD 500 -d dfile -HS 10 -HR 0
tdisp96
tlegn96
tregn96
tpulse96 -V -p -l 4 -d dfile > tifile96
fprof96 -A < tifile96
Of course isotropic models can be used with these programs, also
though the computations will take longer than for the
corresponding isotropic codes.
August 20, 2010
- Modified gsac in the
following manner. The prs
(plotrecordsection) command supports the use of the
MAG field as an axis. The sort
command now permits a sort on MAG.
- The current version of gsac
is now V1.1.37 20 AUG 2010.
July 15, 2010
- The mt (moment
tensor) command in gsac
was modified to permit point forces with the command syntax FN fn FE fe FD fd where the
force(units of dynes) is described in the north, east and down
directions, respectively. The output will be in
meters. The current version of gsac is now V1.1.36 15 JUL 2010
- For the purposes of documenting the menus, gsac accepts a ~ interactive input under
the ppk and refr commands. This will
create a unique CALPLOT graphics file with name DUMPxxx.PLT
which will show the menu buttons and the waveforms.
June 17, 2010
The mt (moment tensor)
command in gsac was
modified to permit an explosion source, mt iso mw 5.0 FILE fileproto. In addition the
generation of separate Z, R, T, N or E components can also the new
syntax mt TO UZ or
in general one of UR, UZ, UT, UN or UE in place of the
previous R, Z, T, N or E, respectively. The current version of gsac is now V1.1.33 17 JUN 2010.
June 14, 2010
- Fixed login error in the gsac
command sync which
failed when there were P picks in the file. The code is a bit
cleaner now. The current version of gsac is now V1.1.32 13 JUN 2010
June 6, 2010
- Added a new command to gsac.
The shift command
moves
the
position of a trace in time, adjusting the A T0 ... T9
markers. The syntax is shift
Fixed
Amount. If amount
is positive the trace is plotted later in absolute
time. This command was introduced to fupport finite fault
modeling, which requires that subfaults fire a later times as
the rupture front excites the subfault. I previously
manipulated the O (origin time marker) and did some fancy
cutting that was not intuitively obvious. The current version of
gsac is now 1.30 06 JUN
2010.
- Modified output of elocate to
provide
a
command
line
that
can
be
used
with
gsac or sac2000. The
line appears as ch o
gmt YEAR DOY HR MN SEC MSEC
May 20, 2010
- Modified iputmod.f in
PROGRAMS.330/SUBS.
Created
igetmod.c, iputmod.c, imodel.h
to implement 'getmod', 'putmod' in C. The igetmod.c source also
includes a routine getdval which
search the model to get the model parameters at a given depth
within the model.
- Added a conversion to ZNE
in the gsac mt command in order to
make Z N and E components. Previously we could make Z R and T
components. I needed this for short distance finite fault
simulation for which the back azimuth to each sub-fault can
change. This makes the current version number of gsac [V1.1.30 19 MAY 2010]
- Added rot3 to zne to the gsac rot3 command.
- At the suggestion of Carol Bryan, fixed a bug in the
deviatoric constraint in the program wvfmtd96 in PROGRAMS.330/VOLVII/src/wvfmtd96.f
- pltsac
- added -TMIN fmin -TMAX tmax
command line arguments to make better plots.
- shwmod96 - added dashed line option - but be
careful if used with automatic -K -1 command line flag.
- sdpegn96 - added -NOBLACK command line flat
to prevent black outline around the observed dispersion for
compatibility with sdpdsp96.
April 7, 2010
- Slight modifications to plotnps.
When I implemented the background fill to implement the -I flag,
I also had a default forced white backgound when the -I flat was
not invoked. While this was the desired effect for converting an
EPS to graphics format such ad PNG using ImageMagick so
that there would be no transparency, this actually broke the
inclusion of EPS files using groff
since the background fill erased the page. As an
example of the use of the current flags, this is how I now work
with the ImageMagick command convert
and the GraphicsMagick command gm convert:
plotnps -F7 -W10 -EPS -K -BGFILL < SACPSD.PLT > 2.eps
convert -trim 2.eps im2.png
gm convert -trim 2.eps gm2.png
plotnps -F7 -W10 -EPS -K -I < SACPSD.PLT > 3.eps
convert -trim 3.eps im3.png
gm convert -trim 3.eps gm3.png
- The current version of Computer Programs in Seismology is NP330.Apr-07-2010.tgz
March 18, 2010
- Modified the generation of EPS by plotnps to a) paint the
background to avoid transparency in ImageMagrick, and b) to
permit an inverted background. The inverted background is useful
for presenting some colored lines, e.g., yellow, which are
difficult to see against a white background. The following
images show the syntax of the command abd also the result of
converting the EPS to a PNG file using GraphicsMagick. The
ImageMagick command trims to the smallest size
|
|
plotnps
-F7 -W10 -EPS -K < GRAYSC.PLT > t.eps
gm convert -trim t.eps graysc.png
|
plotnps
-F7 -W10 -EPS -K -I < GRAYSC.PLT > t.eps
gm convert -trim t.eps grayscinv.png |
February 28, 2010
- Introduced a new program sacpsd
to computer power spectral density of ground acceleration noise
for use in seismic network quality control. This program is in
PROGRAMS.330/VOLVIII/src. The program uses long data segments in
Sac file and the acceleration sensitivity output of the IRIS evalresp from the RESP files
created by the IRIS rdseed.
Documentation describing the use of this program is given
in sacpsd.pdf
and the shell scripts used in the document are given in the
file SACPSD.tgz
(which is unpacked using the command sequence gunzip -c SACPSD.tgz | tar xvf
- . Note that some errors were
corrected on February 28, 2010.
February 16, 2010
- Made slight changes to two routines in the current version of
rdseed:
- PROGRAMS.330/IRIS/rdseedv4.8.slu/Main/output_seed.c
- PROGRAMS.330/IRIS/rdseedv4.8.slu/Parsers/parse_11.c
to move two declarations to the top
of update_type74 in
output_seed.c
and in parse_type11 in parse_11.c. The Solaris cc compiler
did not like the declarations to be intermixed with code farther
down.
- Changed the Main/makefile, Printers/makefile,
Snoop/makefile and Utilities/makefile to remove a CC=gcc
line that I had entered previously
- Made changes to elocate
in PROGRAMS.330/VOLVIII/src to facilitate the exchange of data
for use in other location programs and also for easier change of
source parameters in gsac.
The extra output (inlooks like
STA IWT ARRIVAL TIME PhIDQL PHASE FM CHAN LAT LON ELEV
OK001 0 20100213053058.250 2 0 i S X Z 35.5611 -97.2895 350.00
OK001 0 20100213053057.219 1 0 i P C Z 35.5611 -97.2895 350.00
OK002 0 20100213053100.168 2 0 i S X Z 35.5493 -97.1966 354.00
... ... ... ... ...
HYPO71 Quality : BA
Gap : 84 deg
ch evla 35.5331 evlo -97.2982 evdp 5.52
ch ocal 2010 02 13 05 30 55 968
The last entries are used by cut
here and pasting into the gsac
command line.
January 11, 2010
- Modified the lp, bp, br and
bp commands of gsac to implement a
Chebyshev Type - I filter using the C1 command argument. Following sac/sac2000 the lowpass
filter prototype is defined by the number of poles, the corner
frequency, fc, a transition bandwidth, transbw, and a desired
attenuation factor at the frequency (1+transbw)*fc.
References on this filter are given in wikipedia, but the
reference providing all information, including the gain
normalization constant is Rorabaugh (1997). Implementation
of the Type-II filter, also known as the Inverse Chebyshev
filter (C2 command argument in sac/sac2000
usage) was not implemented since data structures in the file gsac_filt.c would have to
be completely redesigned.
Digital Filter Designers Handbook with C++ algorithms
C. Britton Rorabaugh 2nd Edition
McGraw Hill, New York, 479 pp 1997 Chapter 5
- While testing the Chebyshev implementation, an error in the
implementation of the bandreject filter for a single-pole
stage changing
} else if (filt_lhp == BANDREJECT){
wc = fwarp(Wc,dt);
wl = fwarp(Wl/ffac,dt);
wu = fwarp(Wu/ffac,dt);
num[0] = wl*wu*dt*dt;
num[1] = 0.0;
to
} else if (filt_lhp == BANDREJECT){
wc = fwarp(Wc,dt);
wl = fwarp(Wl/ffac,dt);
wu = fwarp(Wu/ffac,dt);
num[0] = wc*wc*dt*dt;
num[1] = 0.0;
December 4, 2009
-
Upgraded the distributed IRIS programs evalresp from
version 3.2.38 to 3.2.40 and rdseed from 4.7.5
to 4.8. In order to support the CPS architectures, e.g.,
CYGWIN (Windows), MacOS-X, LINUX and Solaris, changes were
made to these codes. The changes are given in the links in
this section. This upgrade was required in order to continue
to use the most recent SEED volumes.
The two programs are installed in PROGRAMS.330/bin
with the executable nanames evalresp and rdseed.
The IRIS distribution found in PROGRAMS.330/IRIS/rdseedv4.8
provides the executables rdseed.linux, rdseed.ppcMac,
rdseed.intelMac, rdseed.Sun86. rdseed.SunSparc.
If you wish to use these programs, you must install them in
your system and adjust the path to find them.
-
evalresp - The
distributed code has an #include <malloc.h>. This is
deprecated in MacOS-X to <malloc/malloc.h> which
means that the code will not compile. Since modern
compilers provide the prototypes for malloc in
<stdlib.h>. I changed alloc_fctns.c, parse_fctns.c, regexp.c and string_fctns.c by commenting out the
#include <malloc.h> and ensuring the presence of a
#include <stdlib.h>.
Detailed changes.
NOTE THAT evalresp-3.2.40 will configure properly for the
MacOS.
October 8, 2009
Summary of changes made since July
22, 2009
- gsac is now at version
1.1.28 30 SEP 2009. The
following changes were made:
- The command plotpk (ppk)
regional or plotpk
(ppk) teleseism now had an Undo button which removes the last phase
pick made. This was introduced to make phase picks for
location easier to use
- The command merge
had an improper index at line 214 which caused memory
corruption.
- The subroutine gsac_sac.c,
which is used for low level reading and writing of the SAC
files, had an improper test of the file type. It now correctly
reads the files with an IXY format, e.g., the spectral
amplitude files.
- fmtp has a flag -X to output nodal
plane strike dip and rakes to the standard output with a line
that reads
SDRPLANES stk1 dip1 rake1 stk2 dip2
rake2
The reason is that I wanted to
examine the dip variation in the L'Aquila aftershock sequence
- elocate was changed
so that the location time is also written in a format for use by
the gsac command ch ocal, e.g., ch ocal 2009 07 29 10 00
37 123. The new output is indicated in the color red:
RMS Error : 0.150 sec
Travel_Time_Table: nnCIA
Latitude : 35.9288 +- 0.0130 N 1.4501 km
Longitude : 43.2796 +- 0.0173 E 1.5470 km
Depth : 13.65 +- 0.88 km
Epoch Time : 1247949147.952 +- 0.31 sec
Event Time : 20090718203227.952 +- 0.31 sec
Event (OCAL) : 2009 07 18 20 32 27 952
HYPO71 Quality : CD
Gap : 161 deg
In addition two more prototype
velocity models were added. The KOREA
model corresponding to my pre-computed Green's functions
for Korea, and the nnCIA
model corresponding to my pre-computed Green's functions for
waveform of earthquakes in the Central Italian Appenines.
July 22, 2009
- Corrected an inconvenience in the gsac write
command. The program was in the specification of the APPEND suffix and PREPEND prefix option.
Until this modification the shortcuts AP suffix and PR prefix were permitted. The string matching
in gsac_arg.c only
considered the first two characters. This an attempt to execute
the command GSAC>
write APZNZ10HHR APZNZ10HHT APZNZ10HHZ was doomed since
the first AP was
interpreted as APPEND rather than a station name. This is
fied now, but NEVER use APPEND
or PREPEND as a file name in a write command. Implementing this feature
of sac force the use
of a particulatly bad command grammar. [Thanks to Harley Benz]
- Documentation for the Inversion
of Surface-Wave Dispersion and Receiver Functions for
Structure was corrected to show the correct use of the -ZMAX flag for shwmod96 [thanks to Wen-Jie
Jiao]
June 19, 2009
MEMORY | 1 | 2 | 3 | 4 | ...
1-D real array | x(1) | x(2) | x(2) | x(4) | ...
2-D real array | x(1,1) | x(2,1) | x(1,2) | x(2,2) | ...
Complex array | Re[x(1)] | Im[x(1)] | Re[x(2)] | Im[x(2)] | …
This was carefully fixed by
defining a revised routine zfour(z,npts,isign,dt,df) in
all
code. This is a step in defining every variable in a manner
compatible with modern computer languages.
The program files updated
were as follow:
VOLII/src/sacmat96.f
VOLII/src/sacmft96.f
VOLII/src/sacpom96.f
VOLIII/src/spulse96.f
VOLIV/src/rftndr96.f
VOLIV/src/twofft.f
VOLV/src/ffilt96.f
VOLV/src/fspec96.f
VOLVI/src/hpulse96.f
VOLVI/src/hrftn96.f
VOLVI/src/hudson96.f
VOLVI/src/trftn96.f
VOLVIII/src/sacdecon.f
VOLVIII/src/sacevalr.f
VOLVIII/src/sacfilt.f
VOLVIII/src/saciterd.f
VOLVIII/src/sacspc96.f
VOLIX/src/cpulse96.f
The following documentation was
updated:
DOC/OVERVIEW.pdf/cps330o.pdf
DOC/GSAC.pdf/cps330g.pdf
DOC/STRUCT.pdf/cps330c.pdf
DOC/SOURCE.pdf/cps330s.pdf
CALPLOT.pdf/cps330p.pdf
June 11, 2009
-
Hermann Zeyen noticed a problem
in the FORTRAN implementation of the subroutine
arsac in the file SUBS/sacsubf.f. "if the
number of data is multipleof five, the program tries to read
one line too much." This is not a problem in the C
version in SUBS/sacsubc.c
-
Modified the SLU version of rdseed to improve
comments in the sac pole-zero file.
May 23, 2009
-
Corrected moment magnitude
definition in fmech96 to replace the 16.05 by 16.1 in
log Mo (dyne-cm) = 16.1 + 1.5 Mw
-
Added a new command to gsac.
The mt command uses the precompute Green's
functions (e.g., f96tosac -B -G or f96tosac -B -T ) to compute
a Z, R, T seismogram or the 3 ZRT seismograms for a
doubel couple source. The output will be in units of m/s
for a step dislocation. The purpose of this command is
to be able to quickly create forward synthetics for use in
finite-fault inversion.
-
Changed the part sacmft96
that selects the period limits. The previous code looked
like if(p .ge. pmin .and. p .le. pmax.and. p.ge.permin
.and. p.le.permax)then which had roundoff problems as
noted by Renata Schivardi at INGV Bologna. The code was
changed to be slightly more generous by changing the test
to
if(p .ge. 0.99*pmin .and. p .le. 1.01*pmax .and. p.ge.
0.99*permin .and. p.le.1.01*permax)then
-
Changed the part of sacmat96
that selects the period limits. The previous code was if(p
.ge. pmin .and. p .le. pmax)then This was changed to if(p
.ge. 0.99*pmin .and. p .le. 1.01*pmax)then to overcome
roundoff error. This problem was noted by Renata Schivardi at
INGV Bologna.
-
Modified gsac to permit a title to be placed on the
figure generated by the plotsp command.
March 30, 2009
The reason for this is that a field
system had a broadband sensor with internal UVW sensors and also
an accelerometer. As part of seismic network quality control, the
recordings of each were corrected for instrument response and the
derived ground motions were compared for that part of the signal
with high signal-to-noise. These should be identical since the
ground motion is the same. For our field system, this
did not occur, and we returned the broadband and
accelerometer to the university and compared the recordings to a
reference system. A similar comparison showed that the
problem was with the broadband sensor. The rotate3 to uvw was
added to gsac to isolate the internal UVW sensor that was
the cause of the problem. To do this the ZNE motion of the test
and reference systems were rotated to the UVW coordinate system of
the test system and the individual UVW traces were compared.
Because we compared two different 3-component sensors, recording
of the UVW channels of each would not have worked since each was
different.
A description of the coordinate system transformation inplemented
now is gsac given in the PDF document rotate3.pdf
March 19,2009
-
At the request of Chuck Ammon, hudson96 now has a
flag that permits one to override the model computed T*. This
is invoked with the command option -TSTAR tstar .
A valid range of the tstar is >= 0. If one
enters a value of 0.0 the program sets this internally to
0.0001 as a quick way to implement this without a major code
rewrite.
March 18, 2009
-
Incorporated a fix in subroutines
PROGRAMS.330/SUBS/sacsubf.f and PROGRAMS.330/SUBS/sacsubc.c
to use the Enumerated Types for IZTYPE of IUNKN, IB, IDAY,
IO, IA, IT0, IT1, ..., IT9.
-
Modified PROGRAMS.330/VOLVI/src/hudson96.f
to fix logical errors in the time delay correction for the
T* operator, and more importantly to overcome numerical
proplems for the S-wave incident at the receiver. This was
done by replacing an explicit matrix sub-determinant by a
compound matrix value.
-
Modified PROGRAMS.330/VOLV/src/time96.f
to compute the travel time and ray parameter for the sS
arrival.
-
The distribution now includes the source code for the menu
program for terminals, dialog (dialog-1.1-20080819)
March 3, 2009
February 21, 2009
-
Renamed the utility program CAL
in PROGRAMS.330/CALPLOT/Utility to calplt. The
reason for this is that MacOS-X does not distinguish between
upper and lowercase executable names, and CAL
conflicts with the calendar program cal. Since CAL
was used in shell scripts to annotate plots, instances of
CAL will be changed to calplt until all are
corrected. The graphics output of the program is renamed
CALPLT.PLT to be consistent with the naming convention
of other programs. Correct documentation in PROGRAMS.330/DOC
and MOMENT_TENSOR tutorial and others tutorials to reflect
this change.
-
Corrected the MOMENT_TENSOR tutorial scripts to provide more
information on copying files. Also corrected the DOBYTEORDER
in the GREEN directory of precomputed Green's functions.
February 18, 2009
-
Fixed a but in the programs spulse96, hspec96,
hudson96, time96, timmod96 and refmod96
which
was
due to the fact that the parameter refdep was not
defined in subroutine frstar. This was
discovered because of memory dump problems on some computers.
January 5, 2009
-
Introduced a new program fmlpr which is based on the
program lprmech written by Charles J.
Ammon. This program is controlled from the command line
and creates two files. The first is a printer-plot tabulation
of a double-couple focal mechanism solution, and the second
provides those results in a global CMT 'ndk' format. A example
run if given by the line
fmlpr
"TITLE" "DATETIME" "GEOG" EVDP MW strike dip rake "FILE1" "FILE2"
"FILE3"
TITLE DATETIME and GEOG are enclosed in quotes
FILE1 - name of file with stations used
FILE2 - file with data summary: NSTA NCMP MINPER
FILE3 - file with filtering commands
fmlpr "USGS/SLU Moment Tensor
Solution" "${YEAR} ${MO} ${DY} ${HR} ${MN} ${SEC} ${MSEC} ${LAT}
${LON} ${DEPTH} ${MAG}" "${STATE}" ${HS} ${MW} ${STK} ${DIP}
${RAKE} sumstaused.txt sumswavendk.txt sumwfilter.txt
Here the first string
provides a title for the printer plot, the second string provides
information about the earthquake location and the third string
provides a geographical description of the earthquake sourcce
zone. These nest items, e.g., HS, MW, STK, DIP and RAKE are
the source depth, moment magnitude and double-couple source
parameters. The last three items are text files for turther
annotation. The sumstaused.txt lists the stations used
for the source inversion, sumswavendk.txt consists of a
single line with three numbers - the number of stations used, the
number of components and the minimum period used in the inversion.
The last file, sumwfilter.txt, lists the gsac
signal preparation commands so that the filter band is known.
When invoked as
fmlpr "USGS/SLU Moment Tensor Solution" "2008 10 14 03 07 28
000 35.76 -100.704 5.0 3.80" "Texas" 11.0
3.65 285 70 -80 sumstaused.txt sumswavendk.txt sumwfilter.txt
with sumstaused.txt containing
IU.ANMO TA.MSTX
US.CBKS US.JCT US.KSU1 US.OGNE US.SDCO
US.WMOK
and sumswavendk.txt contining
8 19 10
and sumwfilter.txt containing
hp c 0.05 n 3
lp c 0.10 n 3
br c 0.12 0.25 n 4 p 2
two files are created: 20081014030728.msg, which is the
printer plot, and 20081014030728.ndk, which is the
solution information in the Global CMT ndk format (http://www.seismology.harvard.edu/projects/CMT/catalog/allorder.ndk_explained
). The contents of the file 20081014030728.ndk are
ENS 2008/10/14 03:07:28:0 35.76 -100.70 5.0 3.8 0.0 EL SALVADOR
C20081014030728A B: 0 0 0 S: 8 19 10 M: 0 0 0 CMT: 1 TRIHD: 0.2
CENTROID: 0.0 0.0 35.76 0.00 -100.70 0.00 11.0 0.0 FREE S-20090105234025
21 -2.379 0.000 2.526 0.000 -0.147 0.000 2.681 0.000 -0.949 0.000 -0.064 0.000
V10 3.758 24 7 0.000 9 102 -3.758 64 211 3.758 285 70 -80 78 22 -116
January 2, 2008
-
Created a new
program: redodate The purpose of redodate is to change a data/time
specification through an offset, e.g.,
redodate
2009 01 01 00 01 02 456 -3660
returns
2008 12 31 23 00 02 456
I require this in order to
specify window segments for traces that I wish to
download. Previously i started at the origin time, but I
needed more signal for inversion, so I would like to start some
offset (a negative number) before the specified origin
time. This code segment indicates how this would be
used in a shell script, where I would like to start 60 seconds
before the origin time:
YEAR=2009
MONTH=01
DAY=01
HOUR=00
MINUTE=01
SECOND=02
MSEC=456
OFFSET=-60
redodate ${YEAR} ${MONTH} ${DAY} ${HOUR} ${MINUTE} ${SECOND} ${MSEC} ${OFFSET} >
tmp
read YEAR MONTH DAY HOUR MINUTE SECOND MSEC < tmp
echo "YEAR :" ${YEAR}
echo "MONTH :" ${MONTH}
echo "DAY :" ${DAY}
echo "HOUR :" ${HOUR}
echo "MINUTE :" ${MINUTE}
echo "SECOND :" ${SECOND}
echo "MSEC :" ${MSEC}
rm -f tmp
December 22, 2008
November 12. 2008
-
Modified plotxvig to
permit the use of a -C flag. When this is set,
the arrow cursor for display is replaced by a crosshair.
The pupose of this is to be able to read axis numbers for a
point within a plot.
-
Modified do_mft to permit the use of a cross-hair
cursor. Note that this is not perfectly implemented. The
purpose of both of these changes is to be able to compare the
dispersion shown in a regular plot to that of the multiple
filter analysis when plotting both side-by-side. I was
comparing observed dispersion to an independent tomographic
estimate.
September 3, 2008
August 4, 2008
-
Modified the CALPLOT graphics
programs plotxvig, plot4014, plotnps, plotgif
and plotpng to support -X0x_offset and -Y0y_offset
flags. These were always part of the reframe command
for repositioning a plot. I added this feature for use mostly
with plotxvig so that I can display plots that are
too high on the page.
-
Changes #define NCHAR 1000 in VOLVIII/gsac.src/gsac_parse_command.c
to #define NCHAR 2000 to permit more file names on
the command line. I niticed this when trying to read in a lot
of files, but I dod not have this problem with using a wild
card.
June 13, 2008
-
Corrected the gsac_refr.c (refraction
analysis
routine)
of
gsac, so that the refraction velocities are given
correctly for a reverse spread, e.g., negative distances which
would yield negative slopes and velocities.
-
Modified refmod96 so that the line width parameter
(e.g., command line -W width) applies to the reflected
arrivals. For multilayered models, the line width or
pattern is a function of the reflection coefficient of the
deepest reflecting interface to be able to focus on the
important reflection. The reflection times are still the times
to the deepest layer, e.g., if the deepest layer boundary is
3, then the rays computed are 1 1, 1 2 2 1 for the
first reflection, and 1 1 1 1 and 1 2 2 1 1 2 2 1 for
the first multiple; however the mixed multiple 1 2 2 1 1
1 is not computed. A plot showing the refraction
and reflection predictions is given in refmod96 examples
May 25-29, 2008
-
Corrected hudson96 so that the SH arrival times are
computed and ultimately placed in the file96 and sac
file trace headers (May 25, 2008). To properly set the
arrival time, the start of the teleseismic attenuation pulse
is determined empirically using two FFT's. This must be done
using the same DT as for the synthetic because of the
frequency dispersion inherent in the causal T* operator (May
28, 2008). Corrected indexing when merging source and receiver
structure with teleseism structure. The halfspace was
not inserted correctly (May 29, 2008).
May 7, 2008
-
Modified evalresp in PROGRAMS.330/IRIS/evalresp-3.2.38.slu
to handle FIR stages with a large number of
coefficients. Specifically, Change Eric J. Haug SLU May,
2008 to handle continuation for FIRST with a large number of
coefficients. Basically we need to get past the tendency of
some SEED writers to put the following phase in the RESP file
"No Abbreviation Referenced" for Response in units lookup and
Response out units lookup. The
modified code provides the same output as the current version
of the IRIS jevalresp
April 9, 2008
March 31, 2008
-
Created a new program hudson96 which makes
approximate teleseismic synthetics following the work of
Hudson(1969). This approach focuses on the direct teleseismic
P or S arrival through the use of a stationary phase
approximation for the ray parameter connecting the direct
arrival from the source to the receiver. The interesting
feature of this approach is the use of propagator matrices for
the source and reciever crusts, so that the depth pahses are
automatically included as well as the details fo the source
and receiver crustal models. The program is self
contained in that the travel time, T* and geometrical
spreading are comptued within the program. A tutorial on
the use of the program is on the tutorial page.
Hudson, J. A. (1969). A quantitative
evaluation of seismic signals at teleseismic distances - II. Body
waves and surface waves from an extended source, Geophys. J.
Roy. astr. Soc. 18, 353-370.
-
The programs time96, timmod86, refmod96,
spulse96, and hspec96 were
upgraded with the same travel time routine used in hudson96,
to now compute the geometrical spreading for teleseismic
arrivals.
March 14, 2008
-
Added a new flag to hprep96. The -ML length_mul
is typically used in increase the length for wavenumber
integration,L , estimated using the Bouchon criteria, which is
that dk = 2 pi / L . There are circumstances when the
default value of L is a little too small, and
significant noise signals occur at later time. This
option is easieer than trying to guess the correct value of L
for the command line override, e.g., -XL xleng
February 8, 2008
-
Corrected an error in the prs command of gsac.
Because of an incorrect clipping command in the
graphics, the trace annotation was not done correctly
when using the prs ann sta command.
-
Tested the Earth flattening used in hspec96, tspec96,
surf96, joint96, sdisp96, sregn96 and slegn96
January 28, 2008
-
Created a new program ttime96
-
Installed identical travel time
routines for TI models in the following programs: ttime96,
ttimmod96, tspec96 and hprep96.
-
The travel time validation is
given at the links Isotropic Model Test
and Transverse Isotropic Model Test
-
Fixes errors in tspec96 that affected synthetics with
Q. For an isotropic model, the output of tspec96 agrees
with that of hspec96
January 22, 2008
-
Installed identical travel time routines for isotropic models
in the following programs: time96, timmod86, refmod96,
spulse96, hspec96. The new routines provide the
correct results for flat and spherical velocity models for
travel times, ray parameter, T* of the P, S, pP and sP
phases. They also correctly handle the artifice of
representing the atmosphere by a stack of layers with negative
layer thickness so that a depth of 0 corresponds to the
surface of the solid.
January 17, 2008
January 10, 2008
January 3, 2008
A detailed tutorial on teleseismic
synthetics and source inversion must be prepared.
This update changes the following programs: sdisp96, sregn96
and slegn96 in PROGRAMS.330/VOLIII/src, srfdis96,
srfdrr96 and srfdrl96 in PROGRAMS.330/VOLIV/src,
and hspec96 in PROGRAMS.330/VOLVI/src. Note hspec96
was modified on January 14, 2008 to correctly handle the source
and receiver depth insertion for the flattening. The use fo the
reference depth still needs work to handle atmospheric layers
about the surface.
November 17, 2007
-
Upgraded the distributed IRIS programs evalresp from
version 3.2.25 to 3.2.38 and rdseed from 4.5 to
4.7.5. In order to support the CPS architectures, e.g.,
CYGWIN (Windows), MacOS-X, LINUX and Solaris, changes were
made to these codes. The changes are given in the links in
this section. This upgrade was required in order to continue
to use the most recent SEED volumes.
The two programs are installed in PROGRAMS.330/bin
with the executable nanames evalresp and rdseed.
The IRIS distribution of rdseed provides the
executables rdseed.linux, rdseed.ppc-mac, rdseed.intelmac
and rdseed.solaris. These binaries are available in
the PROGRAMS.330/IRIS/rdseed4.7.5 directory of CPS.
If you wish to use these programs, you must install them in
your system and adjust the path to find them.
-
evalresp - The
distributed code has an #include <malloc.h>. This is
deprecated in MacOS-X to <malloc/malloc.h> which
means that the code will not compile. Since modern
compilers provide the prototypes for malloc in
<stdlib.h>. I changed alloc_fctns.c, parse_fctns.c,
regexp.c and string_fctns.c by commenting out the #include
<malloc.h> and ensuring the presence of a #include
<stdlib.h>.
Detailed changes.
NOTE THAT evalresp-3.2.40 will configure properly for the
MacOS. When this is officially released, it will be included
in CPS!
November 9, 2007
-
Modified the code of wvfmch96 so that there is no
internal limit on the number of waveforms for which forward
synthetics are required. In the present distribution, the
maximum number of waveforms permitted for use by wvfmch96,
wvfgrd96, wvfmt96 and wvfmtd96 is 2000, with
each trace permitted to consist of 4096 points. Because
wvfmch96 creates a forward synthetic for a given
moment tensor description, this program can be used as part of
a QC process to evaluate all recorded waveforms in a data
center. Since there can easily be more than 2000 such
waveforms today, hard dimension limits are a major
inconvenience. The logic of wvfmch96 was
modified so that the forward synthetic is created separately
for each input file, rather than reading in all files and
corresponding Green's functions into memory.
October 21. 2007
Hacked a fix for the distance computation used in gsac (gsac_subs.c)
and udelaz. The problem was that if the event -
station latitude/longitude pair differed in coordinates by 0.0001
degree or less, then the computed distance in kilometers was
incorrect because of roundoff error of the angles used for the
elliptic integrals. The artifice used for distances less than 1.0
km, is to use the spherical relation delkm = 6371.003 *
delradian for the short distances.
October 19. 2007
October 04, 2007
-
Modified sacmft96 and sacpom96 to take
ABS(DIST) where DIST is from the SAC header. This is
because I sometimes require negative distances when using
shallow refraction data. For a split spread, the gsac
refr command will display the split spread. The refl
command will also work now. The only bad thing is
that the refraction velocities may now appear to be
negative. These changes were made as part of an effort
to develop a procedure for determining shallow S-wave velocity
structure.
A problem with the distance/azimuth
computations in gsac was noticed in the process of
computing epicentral distances less than 1 km. This problem
will also affect udelaz. This is currently under
investigation.
August 30, 2007
-
gsac - implemented
two new commands in prs (plotrecordsection): ScaleRelative
(SR) and ScaleAbsolute power (SA Power) to control the
amplitude of the traces in the record section.
ScaleRelative is the old behavior which just scales each trace
to the same maximum amplitude (AMP value), except that the
scaling now applies only to the actual trace window and not
the whole trace. ScaleAbsolute 0.0 adjusts all traces to
plot true amplitude. If the non-time axis is distance, then
ScaleAbsolute 1.0 adjusts for geometrical spreading by
multiplying all traces by distpower . For a true
refracted arrival (e.g., with no velocity gradients),
power=1.5 would show the refractions at the same level. The
current gsac version is now[V1.1.21 30
AUG 2007]
-
wvfgrd96 is now modified so that the output traces,
e.g., TRACE.obs and TRACE.pre have the time shift applied.
This means that if pltsac is used with the flag -USER9
the time shift will be displayed even though the traces
align. The objective here is to present the observed and
predicted in such a way to focus on amplitude and phase
differences and not time shifts.
August 19, 2007
|
> sac2000
SEISMIC ANALYSIS CODE [8/8/2001 (Version
00.59.44)]
Copyright 1995 Regents of the University of
California
SAC> r ../*.sac
../imp.sac ...nimp.sac
SAC> w append .s
../imp.sac.s ...nimp.sac.s
SAC> ls
SAC> r
../imp.sac ...nimp.sac
SAC> w prepend p.
p.../imp.sac ...nimp.sac
ERROR 101: opening file
SAC>
|
GSAC - Computer Programs in Seismology [V1.1.19 01 AUG
2007]
Copyright 2004,
2005, 2006, 2007 R. B. Herrmann
GSAC> r ../*.sac
../imp.sac ../nimp.sac
GSAC> w append .s
imp.sac.s nimp.sac.s [Note
write in local directory not original ]
GSAC> r
../imp.sac ../nimp.sac
GSAC> w prepend p. [Note write in local directory not
original - note Sac2000 ]
p.imp.sac p.nimp.sac [cannot prepend in parent directory ]
GSAC> w append .S prepend P. [Note write in local directory not
original ]
P.imp.sac.S P.nimp.sac.S
GSAC>
ls
[show what was created in the
current directory]
imp.sac.s nimp.sac.s p.imp.sac
P.imp.sac.S p.nimp.sac P.nimp.sac.S
|
One major difference is that
although a 'w' under both sac2000 and gsac
will overwrite the original file in the upper directory, the gsac
append and prepend will create files in the current
directory. Because gsac strips off the leading
directory path prior to implementing the prepend command,
this operation will work.
August 16, 2007
August 14, 2007
-
Initial development of new source model for surface-wave and
waveform inversion - the crack. Following Nakano, M., and H.
Kumagi (2005). Waveform inversion of volcano-seismic signals
assuming possible source geometries, Geophys. Res. Letters 32,
L12302, doi:10.1029/2005GL022666, a new command line option -CRACK
is permitted for the programs srfgrd96, sdprad96,
sdpspc96 and wvfmch96. The crack is specified
in terms of the parameters -DIP dip -STK -RAKE rake -MW mw
or -M0 mo where dip is the dip of the
crack, stk is the dip-direction of the crack
(dips down to the right when looking in the direction of
strike). The rake is used to indicate an
expanding crack (positive) or closing crack (negative).
The moment magnitude or moment are positive numbers.
At present the program fmplot is not modified, nor
is the output of sdprad96 to indicate that the crack
model was used instead of the earthquake double couple.
The reason for this introduction of a new source was to study
the Utah mine collapse of 2007 08 06 08 48 40!
August 1, 2007
July 16, 2007
-
gsac - corrected error in the remove trend rtr command.
Roundoff
error
messed
up
the
determination
of
the
regression
equation
a
+
bx
.
Previously
x
was
the
time
index,
e.g.,
i=0,
i
<
npts. For long time seeries this let to large errors in the
determinant. I replaced this by defining x as i/npts with
varies between 0 and 1 for the time series.
July 1, 2007
ppk pq
or ppk pquality
This is
functionally similar to the previously introduced 'ppk
quality' except that the purpose is to simultaneously define
a useful trace and to reset the P-time. For my source inversion, I
set the theoretical P-time using the AK135 predicted times.
This is OK for source inversion since the inversion code permits a
simple waveform time shift. However for broadband depth or energy
determination, it is necessary to start at the P time. To
speed up the review, I added this new option. So now move the
crosshair to the P-wave first arrival, and then click any mouse
button. Simultaneously the P-arrival time is reset and the trace
is marked as one to be used (IHDR20 is set to +1).
April 23, 2007
March 31, 2007
March 19, 2007
-
gsac - Added an option to ppk (plotpk), e.g.,
ppk q
For source inversion, one must examine many traces to decide
which to use since the inversion results are strongly
influenced by bad data. For teleseismic and regional
data, one may review hundreds of traces. So gsac was
modified to permit a rapid assessment of the trace quality.
The assessment is placed in the integer header value IHDR20
(gsac notation) and a subsequent shell script examines the
value there to decide if the trace should be moved to the
final directory for source inversion.
The purpose is to use the plotpk display to quickly
define whether to use a trace for a source inversion.
For teleseismic inversion one must examine the quality of the
P-wave arrivals, so one reads the traces, bandpasses, sorts
with distance, applies a suitable xlim and executes the
command
GSAC> ppk relative perplot 4
q
The result is the image
The default behavior is that a mouse click says accept - of
course one can use the 'a' key to accept and the 'r' key to
reject. The default behavior of the mouse click can be changed
using the menu at the bottom. When a trace is accepted or
rejected, a red '+' or a blue 'X' is drawn at the lower left
corner of the trace. The effect of these operations is
to set the integer header position 20, which I call IHDR20 to
+1 for accept and 0 for reject. The default value of this
header field is -12345 This changes added about 300
lines of trivial code to gsac_plotpk.c. (In the image above,
the P times are from the velocity model. The slight offset is
no problem for the inversion since the inversion programs
implement a time shift). This process of performing QC for P,
SH and SV is shown the shell script IDOQC
The value of this approach to assess trace quality was
immediately apparent by the resulting improved source
inversions. The current version of gsac is V1.1.18 31
MAR 2007
March 15, 2007
-
time96 - corrected the
on-line help to reflect the -P -SV and -SH flags
-
wvfmch96, wvfmtd96, wvfmt96 and wvfgrd96 - added
error
checking
on
the
input
files
-
reject
a
data
set
if
the
number
of
sample
points
is
<=
4
or
if
DEPMAX=DEPMIN,
which
indicates
a faulty data set. Current version of wvfmtd96 and
wvfmt96 were modified to use both the external a
priori and internal variance weighting schemes.
February 22. 2007
for i in *sac
do
GCARC=`saclhdr -GCARC $i`
EVDP=`saclhdr -EVDP $i`
A=`time96 -M ${MODEL} -GCARC ${GCARC} -T -EVDP ${EVDP}`
T0=`time96 -M ${MODEL} -GCARC ${GCARC} -T -EVDP ${EVDP} -SV`
echo ${GCARC} ${EVDP} ${A} ${T0}
gsac << EOF
rh $i
synchronize O
ch A ${A} T0 ${T0} T1 ${T0}
wh
quit
EOF
done
The 'synchronize O' is used to reset
the time stamp so that 'O' , the offset of the origin time with
respect to the reference time, is zero. In this case then the 'A'
will be the travel time of the P first arrival.
February 20. 2007
February 8, 2007
-
The computation of epicentral
distance in kilometers was changed in gsac, udelaz
and edabac. Prior to this it was estimated from the
great circle arc, thus assuming a spherical Earth. It is now
computed numerically by direct evaluation of an incomplete
elliptic integral of the second kind. Sac uses Rudoe's reverse
formula (Geodesy, G. Bomford, Carlendon Press, 4th ed, 1980,
Section 2.15(b), page 121. Rather than getting confused
about inverse tangents, the idea of Rudoe was used, e.g., the
plane through the center of the Earth, the station and
epicenter intersects the spheroid as an ellipse, and the
linear distance is computed along the ellipse. However, this
derivation uses vector analysis more that it does
trigonometric relations. The computation of the SAC
header value dist agrees with the values of sac2000
(based on a comparison of 19554 different combinations of EVLA,
EVLO,
STLA and STLO.
-
Output formats of sacmft96 were changed to
avoid format overflow (***** in FORTRAN) when AZ,
BAZ and GCARC had the undefined values of -12345.
January 24, 2007
timmod96 -TIMEP pfile
timmod96 -TIMES sfile
where pfile or
sfile are ASCII files with a one distance – time pair per line,
e.g.,
124.454140 20.464310
238.482346 35.048897
145.274033 24.015045
350.404083 48.721909
which can be
created from the DIST and A (for P) sac header values through a
loop
for i in *Z
do
saclhdr -NL -DIST -A >> ptime
done
This option
permits plotting measured first arrivals on the same plot as the
model predicted values
-
gsac – the merge command
now
has
the
command
line
option
GAP
gap_value.
Before
this
change,
a
merge
of
SAC
files
having
gaps
resulted
in
a
zero
fill
of
the gaps. Now the user value is inserted. This may permit a
user program to detect the gap. For example if the SAC trace
consists of integers from the ray 24 bit digitizer, the
expected extreme values are ±8388608. A “MERGE GAP 1.0e+9” would fill
the trace with an impossible number which can then be
flagged.
-
gsac – new command outcsv This
command
causes all traces in memory to be output in ASCII as comma
separated variables (CSV) in the file name f001.csv.
Each row of output consists of the values at one time sample
with entries “Time,amp_trace_1,amp_trace2 ...” The reason
for this command was to create a simple output format that
could be read by Excel or OpenOffice spreadsheet programs so
that students can perform simple time series analysis and
plots. Note that OpenOffice does not permit more than 65536
to be read in
January 13, 2007
-
Modifications to gsac:
Introduced new commands grid, xgrid, ygrid, background. The
grid options permit setting a grid on plots and also too
define the color and type of the grid lines. The background
command permits setting a background color in the plotting
frames.
-
GSAC>
GSAC - Computer Programs in Seismology [V1.1.15 17 JAN 2007]
Copyright 2004, 2005,
2006, 2007 R. B. Herrmann
Also corrected appearance of the help screen and got the
grids to work on the plotsp command. My excuse is that I only
had battery power in a house without electricity for two days
because of the ices storm
January 3, 2007
-
Fixed fmech96, which did not set the back azimuth
correctly. Now one can create Z R T and Z N E three-component
synthetics using fmech96, convert them to SAC format
using f96tosac and used gsac to read in the Z
N E to make the Z R T and everything agrees.
January 3, 2007
January 3, 2007
-
Changed all dimensions for
layering to now have 200 layers instead of 100. Note this does
not apply to the Cerveny programs cprep96, cpulse96, cseis96.
-
changed time96 so that it now can output the ray
parameter. This gets around the problem of using udtdd which
is
an implementation of the Jeffreys-Bullen curves. There is now
a consistency between the model96 synthetics and the
travel time prediction. time96 is used to predict ray
parameter and first arrival for teleseismic P-wave synthetics.
See the tutorial.
December 28, 2006
-
Reformatted help entries in gsac
to be easier to read.
-
Implemented new commands in gsac: version,
triangle, boxcar, trapezoid. version lists the version
number. triangle, boxcar and trapezoid are
time domain filters that convolve the traces in memory with an
isosceles triangle defined by its halfwidth, a boxcar defined
by its width, and a trapezoid defined by the length of its
three segments. All of these pulses are designed to have
unit area and function as a lowpass filter. There is some
granularity in the implementation that means that the sequence
of commands
triangle width 1.0
is not the
same as two commands
boxcar width 1.0
boxcar width 1.0
even though
this should be true by virtue of the convolution theorem.
December 15, 2006
December 12, 2006
November 8, 2006
October 18, 2006
October 14, 2006
October 1, 2006
September 16, 2006
-
gsac Version 1.1.11 – Commands ignore a leading #.
Also the pole-zero and frequency-amplitude-phase files ignore
leading blanks. Finally the ABS macro in gsac.h was correctly
written, meaning that the x -x works properly in plotpk
whether the first x is the earliest or the latest time. The
ignore leading blanks in the response files was required to
overcome some laxness in some pole-zero files. Finally the
ignore # serves the same purpose as in the sh or bash shells –
e.g., to have an inline comment. This means that gsac
commands in a shell script can now have comments. For example,
we can now
#!/bin/sh
#####
# generate a pulse and low pass filter it
#####
gsac << EOF
#####
# generate a pulse
#####
fg triangle delta 0.05 npts 1024
#####
# lowpass at 1 Hz
#####
lp c 1 n 2
#####
# save as the file LP.sac
#####
w LP .sac # a comment can be placed here too
quit
EOF
August 20, 2006
August 8, 2006
-
Corrected a major error in the first arrival computations in
gsac in the interpolation routine. This was
caused by a time series with more than 10,000,000 and involved
floating point roundoff problems in the time array used for
interpolation.
August 8, 2006
-
Began adapting codes to the gcc/gfortran
(4.2) compiler suite. For the most part this just
required adhering to the strict 72 character line for the
FORTRAN 77 standard.
-
The gfortran 4.2 uses buffered
IO, thus overcoming slow IO of the gfortran 4.0 suite
-
The new makefiles force a 4 byte header for the unformatted
sequential IO instead of the default 8 byte.
August 4, 2006
-
Corrected a major error in the first arrival computations in
time96, timmod96, refmod96, hprep96,
hspec96, tspec96, spulse96. Faulty
logic gave too early an arrival time if the first arrival was
the direct wave. This was fixed by computing the distance at
which the critical ray appears; if this is greater than the
desired distance, then use the direct arrival time. In
addition we found that one of the refraction times was not
comptued correctly which velocity decreases with depth. The
error means that the pole-zero option was not correctly
implemented. (Thanks to C. J. Ammon).
June 16, 2006
June 10, 2006
-
Added map command to gsac. This command
creates a well commented GMT script that plots stations,
epicenters, raypaths and topography. The shell script can be
easily edited to change the default plotting parameters.
May 27, 2006
April 26, 2006
-
Corrected gsac_lh.c and gsac_ch.c to properly handle the
IEVREG header variable. sac2000 actually has internal list of
the Flinn-Engdahl regions that print out a nice string for the
integer variable. gsac now just outputs the integer value of
this field.
-
prfmod96 is available. This was written to display
regional velocity models from the joint inversion of receiver
functions and surface waves in Korea. The user defines a
profile in terms of the latitude and longitude at the ends of
the profile. The coordinates of each model (station) and
station model file are read in. This program plots a contoured
velocity map from stations within +-W of the profile. The
contours are labeled. The use has control over the contour
intervals and the labeling intervals.
-
Major changes to the VOLVII waveform inversion
programs. wvfmtd96 and wvfmt96 are now
multi-pass programs that perform and inversion, determine an
alignment time shift and weighting, and then perform final
inversion. The logic of the code follows that developed by
Harley Benz at USGS for teleseismic inversion. As part of the
code modifications, wvfgrd96 and wvfmch96
were also modified.
March 19, 2006
-
Error in gsac/ gsac_prs.c that
tried to flush X11 when in plot mode and not X11 mode. This
caused the program to crash. Also prs relative works.
I noticed this when I wants to plot observed versus predicted
teleseismic P-wave traces, windowed about the P arrival as a
function of distance. This was done to support the newly
modified wvfmtd96 which permits time shifting
February 27, 2006
-
Fixed major error in PROGRAMS.330/CALPLOT/src/XVIG/src/window.c
The execl( ...,init_height, 0) was WRONG!. It should have
been execl(...,init_height, NULL) where NUL is a (char *)0.
This error appeared using a 64 bit LINUX. Thanks to Domenico
Di Giacomo, INGV - Istituto Nazionale di Geofisica e
Vulcanologia, Sezione di Milano.
February 27, 2006
-
Added a backup or return to
previous display option to the plot command of gsac.
This is useful when previewing many traces using the perplot
option to skip back to re-examing missed features.
-
Changed internal routines to prevent a mis-setting of USER1
and USER2. USER1 and USER2 are set to be the minimum and
maximum periods of the filtered trace. When picking P- and
S-arrival times, one may wish to filter the trace with a
zero-phase bandpass filter to see the arrival. The pick is
then saved using the wh command. Until this fix the
USER1 and USER2 were also set. Then then led to an
inconsistency that the header information would not reflect
the trace content. Now the minimum and maximum periods are
preserved, but only written out when the w command
rewrites the trace. This fix was introduced in the source code
for all bandpass filters, the transfer and whiten commands.
January 3, 2006
-
Began adapting codes to the gcc/gfortran
(4.0) compiler suite. For the most part this just
required adhereing to the strict 72 character line for the
FORTRAN 77 standard. There are still problems with the
surf96/rftn96/joint96 suites
-
Introduced an O or o option to gsac
to permit synchronize to set the reference time as the origin
time.
December 14, 2005
September 25, 2005
August 10, 2005
August 3, 2005
June 30, 2005
-
Fixed cut option in gsac so
that
cut CAL YEAR MONTH DAY HOUR MINUTE SECOND MILLISECOND
works correctly.
-
time96 created to predict P first arrival times for
flat and spherical earth velocity models. This is used to
define the starting time window for teleseismic synthetics to
avoid unnecessary white space before the first arrival. The
use of this program in a shell script to compute the
teleseismic P-wave signal but to apply a phase-velocity filter
to focus on teleseismic P for speed
for GCARC in \
20 21 22 23 24 25 26 27 28 29 \
30 31 32 33 34 35 36 37 38 39 \
40 41 42 43 44 45 46 47 48 49 \
50 51 52 53 54 55 56 57 58 59 \
60 61 62 63 64 65 66 67 68 69 \
70 71 72 73 74 75 76 77 78 79 \
80 81 82 83 84 85 86 87 88 89 \
90 91 92 93 94 95 96 97 98 99 \
100 101 102 103 104 105
do
#####
# convert epicentral distance in degrees to kilometers
# for the Earth
#####
dist=`echo ${GCARC} | awk '{print 111.195*$1}' `
#####
# for this distance and source depth get the ray parameter
# and define the phase velocity limits
#####
RAYP=`udtdd -EVDP ${HS} -GCARC ${GCARC}`
CVAL=`echo ${RAYP} | awk '{print 1.0/$1}' `
CMAX=`echo ${RAYP} | awk '{print 4.0/$1}' `
C1=`echo ${RAYP} | awk '{print 3.2/$1}' `
C2=`echo ${RAYP} | awk '{print 0.49/$1}' `
CMIN=`echo ${RAYP} | awk '{print 0.46/$1}' `
#####
# get P travel time == arrival time
February 27, 2006
-
Fixed major error in PROGRAMS.330/CALPLOT/src/XVIG/src/window.c
The execl( ...,init_height, 0) was WRONG!. It should have
been execl(...,init_height, NULL) where NUL is a (char *)0.
This error appeared using a 64 bit LINUX. Thanks to Domenico
Di Giacomo, INGV - Istituto Nazionale di Geofisica e
Vulcanologia, Sezione di Milano.
March 19, 2006
-
#####
A=`time96 -M AK135sph.mod -GCARC ${GCARC} -T -EVDP ${HS}`
T0=`echo $A | awk '{printf "%f", $1 - 100}' `
cat > dfile << EOF
${dist} 1.0 1024 ${T0} 0.0
EOF
hprep96 -M AK135sph.mod -HS ${HS} -HR 0.0 -EQEX -TF -BH -d dfile \
-CMAX ${CMAX} -C1 ${C1} -C2 ${C2} -CMIN ${CMIN} \
-NDEC 1
hspec96 -H
hpulse96 -p -V -l 1 >> file96v
hpulse96 -p -D -l 1 >> file96d
-
Modified sacmft96 and do_mft
to create a temporary file called MFT96.disp instead
of MFT96.DSP to avoid problems when using a caseless
file system, e.g., Cygwin under Windows
-
Modified sacmft96 so that color contouring ande the
spectral amplitude plots reflect the envelope amplitudes
within in the chosen group velocity window. Prior to this,
colors and amplitudes were set by the amplitudes of the entire
filtered trace.
June 16, 2005
-
Fixed typos in gsac which
had
Pp instead of pP in the Teleseismic phase menu
for plotpk. Also ensured that the reset in prs
actually reset all parameters
-
hspec96 now has an earth flattening routine for
spherical earth models. Note hprep96 and tspec96 must be
upgraded to handle a spherical model.
March 23, 2005
-
Implemented B option to plotpk
command of gsac to permit moving back in a
multipage display. (requested by Risheng Chu of SLU)
-
Implemented fileid command of gsac to permit
trace labeling. The syntax was extended to permit placement at
upper (UC) and lower (LC) center of each plot. (Requested by
Luis Rivera and Jeroen Ritsema)
February 11, 2005
-
Added boxcar to gsac
funcgen command - PROGRAMS.330/XOLVIII/gsac.src/gsac_fg.c
-
Added debug statements to gsac
transfer
command to catch types in response input - PROGRAMS.330/XOLVIII/gsac.src/gsac_trans.c
-
Modified PROGRAMS.330/CALPLOT/src/cmd/plotxvig.c
to introduce a new window behavior. If the window is
resized before an erase, the next frame will then use the
window. This means that a non-default window size does nto
ahve to be changed. The gframe calplot command has
been changed to perform an automatic erase but to prohibit a
resize.
-
Corrected gsac xlim commmand to correctly handle GMT
and CAL arguments and to forbid xlim o -10 o 10 if o
is not set withing the header
February 5, 2005
-
Modified PROGRAMS.330/CALPLOT/src/XVIG/src/xvig.c
to unlock the keyboard for the terminal window. Successful,
except that the first character is lost because of the event
catching mechanism. More low level X11 work must be done. On
Solaris with the CDE window manager, the trace window can only
be moved use the Move under the menu option and not by
dragging the window - I will not worry about this in the
future
-
Modified VOLIII/src/spulse96.f VOLVI/src/hpulse96.f
VOLIX/src/cpulse96.f VOLV/src/gpulse96.f to
add a -Z flag to indicate that the internal parabolic or
triangular pulses are to be zero phase - this is required if
we attempt to model the interstation Green's functions
obtained by cross-correlation of background noise, otherwise
there will be a delay causing a mismatch of synthetic and
observed responses. This also necessitated a change in the
documentation DOC/OVERVIEW.pdf/cps330o.pdf
January 31, 2005
-
SUBS/SUBS
-
sacsubc.c, sacsubc.h,
sacsubf.c Added brsach and arsach
only header information. Previously brsac was used with
the overhead of reading all the data. This fixed a problem
in saclhdr with very large files
-
grphsubc.c fixed logic
in clipping routine (baker@usgs.gov)
-
VOLI/src
-
VOLIII/src
-
sio.f corrected minor
typos
-
sdprad96.f corrected
minor typos
-
srfgrd96.f corrected
minor typos
-
sdpspc96.f corrected
minor typos
-
VOLV/src
-
genray96.f subroutine
chkdep, hm1 not defined changed to dm1,
baker@usgs.gov
-
gpulse96.f LER not
defined in subroutine source - baker@usgs.gov
-
ttimmod96.f subroutine
fstarr
-
vlmm undefined - changed to vlmn, baker@usgs.gov
-
fmdfit.f extra long
lines split
-
fmplot changed default
so that no nodal planes are plotted unless requested by
strike,. dip, rake, force or moment tensor flags. In this
simnple mode the command fmplot -F fmplot.tmp will
just plot first motion data. It is then is then up to the
analyst to use the stereonet producted by stereo
to determine the nodal planes and then to rerun fmplot
for the final plot
-
VOLVI/src
-
tspec96.f vlmm not
defined in subroutine fstarr relace with vlmn,
baker@usgs.gov
-
hprep96.f vlmm not
defined in subroutine fstarr relace with vlmn,
baker@usgs.gov
-
VOLVII/src
-
VOLVIII/src
-
saclhdr.c use brsach
instead of brsac - do not need data but want header of
potentially large trace file
-
saciterd.f nbumps
undefined in subroutine tdomain, changed to nshift. tshift
undefined in subroutine tdomain changed the theshift,
baker@usgs.gov
-
elocate.f emares not
defined in subroutine sumup - now set to -9.99
-VELMOD flag cretes the file VEL.MOD. It no longer loists
the sample velocity model fileN
-
sacfilt.f LER not
defined in subroutine gcmdln baker@usgs.gov
-
sacdecon.f thshift
undefined in main routine - change dto the shift
baker@usgs.gov
-
VOLVIII/gsac.src
-
Many changes
-
Improvements:
-
plot, plotpk - low
level
plot
routines changed to let qdp decimation be
controlled by screen resolution
-
plot, plotpk - low
level plot routines changed to reduce number of
non-drawing penup moves
-
prs
(plotrecordsection) - low level plot routines
changed to reduce number of non-drawing penup moves
-
cut, xlim - now
also take GMT 2004 002 01 02 03 445 and CAL 2004
01 02 01 02 03 455 to define the time window
-
GNU readline implemented
- the user may use the 4 arrow keys to move around the
command line and to view history of commands
-
New commands
-
reverse - reverse
a time series
-
correlate - cross-correlate
time
series
-
sign - one bit
sgn() operator
-
stack - stack
traces in memory
-
whiten - force
flat amplitude spectrum
-
markt - put up
velocity tics at 1, 2, 3, 4, 5, 6, 7 and 8 km/sec (note
that these values are pre-assigned. The user
specification is not yet implemented)
-
taper - trace
taper implemented
-
DOC/GSAC.pdf/cps330g.pdf
TO BE DONE
