Commit f9c1aa8f authored by thomas.forbriger's avatar thomas.forbriger

[WP] (i20_gremlin_example): provide essentil building blocks

Provide all essential building blocks for the toy example.
Still missing: proper step-by-step description.
parent 5724ee7f
......@@ -19,7 +19,8 @@ FILETYPE=su
flist: Makefile $(wildcard *.cfg *.sh) README COPYING
flist: Makefile $(wildcard *.cfg *.sh) COPYING \
$(filter-out %.bak,$(wildcard README*))
echo $^ | tr ' ' '\n' | sort > $@
.PHONY: edit
......@@ -197,22 +198,11 @@ layer.VP.ini.pmod.cfg
# start gremlin
# -------------
# prepare gremlin parameter file layer.gpa.cfg
/bin/cp -vpd $< $@
# start gremlin
# initial model
# %.mute.norm.grn Fourier-Bessel expansion coefficients
# P-wave arrival times
# gremlin control parameters
#.PHONY: layerLQ.gremlin
#.PHONY: layerHQ.gremlin
#.PHONY: %.gremlin
%.gremlin: %.mute.norm.grn
cat | gremlin -M $(word 1,$^) \
-G $(word 2,$^) \
-T $(word 3,$^) \
-P $(word 4,$^)
# running gremlin from make is not appropriate, since gremlin is an
# interactive program requiring keyboard input
# ----- END OF Makefile -----
......@@ -18,12 +18,12 @@ reveivers. Noise is added here only for the purpose to demonstrate the effect
of tapering.
Simulated data are available in files
Plots of the data are available in files
Plots of synthetic Fourier-Bessel expansion coefficients are available in
......@@ -69,6 +69,15 @@ clean
completely purges the directory, removing seismogram files too
Shell scripts
a driver scrtipt for gremlin; it calls make to prepare the necessary
parameter and data files and it is able to pass an initial command sequence
to gremlin
this is <README.step-by-step>
step-by-step example
This is a description of a step-by-step use of this toy example. Several cases
may be provided. Currently provided cases are:
layerHQ: One homogeneous layer on top of a homogeneous halfspace. The model is
the one discussed by Forbriger (2011, Figure III.11 on page 88).
layerHQ: One homogeneous layer on top of a homogeneous halfspace. The model
identical with the model for case layerHQ, just quality factors are smaller,
which produces more realistic data, which is more suitable for inversion.
Filename commonly use the case name in a pattern. Simulated data for example
uses the pattern
such that are the simulated seismogram data for case layerLQ.
The description below is provided for case layerLQ. You can transfer the
filenames and commands to a different case by just replacing layerLQ by the
name of the other case.
The file type (here "su") can be changed to (see README).
Forbriger, T., 2001. Inversion flachseismischer Wellenfeldspektren.
Dissertation. Universität Stuttgart. urn:nbn:de:bsz:93-opus-8610
----- END OF README.step-by-step -----
......@@ -24,7 +24,6 @@ make $MODEL $COEFF $ARRIV $PARAM || { echo "aborting..."; exit 2; }
initialize() {
cat << HERE
Berkheim Festplatz Moden anpassen
gremlin parameter file for test-case "layer on top of halfspace"
fmin, fmax, smin, smax, xmax
8., 130., 0.5, 6.5, 0.110
8., 60., 0.5, 4.0, 0.075
balance, tt-error, greenerror, greenmode, logstretch
0.5 , 0.002, 0.2 , 1, 0.1
prefit, ttprefit
0, 4
0, 10
halfspace, stepsize, master, finemodel
30., 0.10, 2, 80
......@@ -22,6 +22,5 @@ min/max Vp, Vs, rho , Qp, Qs
relative stepsize for partial derivative approximation
03/05/2002 changed greenerror
ttprefit is set unusually large by intention since arrival times picked from
near-source traces in synthetic data typically suffer from bias.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment