Modelling Tomography
Modelling Tomography
In the following the different modelling/tomography tools within the modelling menu of REFLEXW
are described. The main tools are:
2D and 3D Finite difference modelling for seismic and electromagnetic wave propagation (chap. II)
2D ray tracing (chap. III)
2D- and 3D- transmission and refraction tomography (chap. IV)
Chap. I includes the generation of a model. This chapter is a more general one because all tools need a
model to be generated first.
Please use in addition to this user’s guide the handbook instructions and the online help.
I. model generation
6. You may set or change the parameters either by editing the corresponding fields of the table or by
using the general fields situated in the right upper corner.
Using the second possibility first you have to enter the parameters within these fields. These
parameters can be overtaken to selected points of the current layer. The selection is done by clicking on
the fields in the first column of the table (the fields, which indicate the number of the points) and is
indicated by a cross or by activating the option take over all. Clicking the take over button in the
ControlBox leads to the updating of the parameters at the selected points or at each point (option take
over all activated).
9. Return to the main modelling menu by simply activating it or by closing the input of model
parameters menu.
10. activate the option new for the next layer. The layer nr. is changed to 2.
11. For that layer you have to define all layer points and the corresponding model parameters - see step
6 to step 9.
13. All layers must be closed - this means they must start or end at the model border or at any other
layer boundary. It is not necessary to do this manually but the options extrapolate and hor.extrapol.
may be used for that purpose. The interfaces don't have to be entered exactly at the edge, intersection
with another interface, respectively, but in the vicinity because the option extrapolate makes the
automatic interpolation between two interfaces as well as the extrapolation of the interface to the
boundary possible. Therefore the program searches automatically the nearest border at the exposed end
and extrapolates in that direction (in the case of an edge, extrapolation in x-, z-direction, respectively,
in the case of an interface to the nearest layer point).
PLEASE NOTE: The layer points are sorted automatically from lower x-distance to higher x-distance,
i.e. the interface function is unequivocal. Should for example a convolution be given it has to occur by
giving several interfaces, like it is done by the predefined symbols circle and rectangle.
14. Enter a filename and save the model using the speedoption or the option file/save model.
15. For additional features like using predefined symbols, combining existing layers or adding a
topography please refer to the handbook instructions under the individual chapters or to the online
help.
The Finite Difference (FD) modelling tool allows the simulation of electromagnetic or seismic wave
propagation, respectively by means of FD-method for different sources (plane wave, point source as
well as "Exploding-Reflector"-source). As a result a single line or the complete wavefield are stored
and can be displayed after.
In the following we are describing the GPR-simulation for a 2D zero-offset section (standard GPR-data
acquisition).
1. First a new model must be generated (see chap. I) or an already existing model must be loaded using
the option file/load model.
6. Enter the wanted DeltaX increment in space direction (equal in x- and z-direction). For the FD-
computation the layer model has to be rastered with a given increment in x- and z-direction (option
DeltaX). Just so, a time increment has to be given (option DeltaT). The size of the space- and time-
increment corresponds to the minimal wave length as well as the velocity. The program determines
automatically the critical value of the space-increment (1/8 of the min. wave length for FD-scheme 4-
space, 1/12 of the min. wave length for FD-scheme 2-space, respectively) and shows this one in the
calc. critical values box down right. This value should not be passed over. A too big chosen DeltaX
increment results in numerical dispersion of the wavelet. Therefore if the result shows such a
dispersion you have to decrease the DeltaX increment.
7. Enter the wanted timeincrement DeltaT. The max. time-increment depends on the max. velocity as
well as on the given space-increment DeltaX (approximately: ∆t <= 1./(%2V) for the el.magnetic
propagation and ∆t <= 1./(Vp+Vs) for the elastic propagation). The critical time-increment of the
current set DeltaX increment is shown in the calc. critical values box at the bottom. A too big chosen
DeltaT increment results in an instability, this means the amplitude increases exponentially with time.
Therefore if the result shows such an amplitude increase you have to decrease the DeltaT increment.
9. Enter the boundary conditions, e.g. lin.absorbing range for the GPR-simulation.
11. Choose the output type - for this example single line.
12. Enter the output parameters for the single line (rec x start, rec x end, rec z start, rec z end). For
example: rec x start: 0; rec x end: 5; rec z start: 0; rec z end: 0 (registration line at the surface)..
13. Activating the option StartFD starts first the rastering of model (if the option raster is activated)
and then starts the external program FDEMSEIS.EXE which will normally be executed in the
background (option background activated), so that it is possible to go on with the work in REFLEXW.
If the option background is deselected, the computation is faster but there is no possibility to work with
REFLEXW until the FD-computation is finished.
14. After having finished the FD-computation you may display the simulation result within the 2D-
dataanalysis.
The ray tracing modelling tool allows the traveltime simulation of electromagnetic or seismic wave
propagation, respectively by means of a finite difference approximation of the eikonal equation. The
calculation of the synthetic traveltimes is restricted to the first arrivals for an arbitrary 2-dimensional
medium. No reflections and secondary arrivals can be calculated. This can be done using the FD-
simulation (see chap. II).
The main application is the seismic refraction but it is also possible to simulate any transmission data.
The raytracing may be used for
- the control of an inverted model
- an iterative adaptation of the calculated and real data by stepwise changing the underground model
1. first a new model must be generated (see chap. I) or an already existing model must be loaded using
the option file/load model.
7. the calculated traveltimes are shown in the lower picture in addition. Now you may check for the
mean traveltime difference using the option Analyse/calculate traveltime differences
In the case of the 2D refraction vertical tomography all sources and receivers are located within one
line at the surface. In order to allow for a high data coverage within the medium vertical velocity
gradients should be present and a curved raytracing for the calculation of the traveltimes must be used.
Chapter IV.1 includes the format and the picking of the traveltime data.
In chapter IV.2 the 2D tomographic interpretation of borehole-borehole transmission tomography is
described.
In chapter IV.3 the 2D tomographic interpretation of a refraction tomography is described.
Before performing the tomography the traveltime data to be inverted must be present.
2D-data format:
travel time, code, transmitter_X, transmitter_Y(Z), receiver_X, receiver_Y(Z)
F8.2 I8 F8.2 F8.2 F8.2 F8.2
example (2 travel times):
800.00 1 0.00 2.00 100.00 2.00
801.60 1 0.00 2.00 100.00 6.00
3D-data format:
travel time, code, transmitter_X, transmitter_Y, transmitter_Z, receiver_X, receiver_Y, receiver_Z
F8.2 I8 F8.2 F8.2 F8.2 F8.2 F8.2 F8.2
example (2 travel times):
800.00 1 0.00 2.00 1.00 100.00 2.00 5.00
801.60 1 0.00 2.00 1.00 100.00 6.00 5.00
These data can be created externally or within REFLEXW when picking the original wavedata.
In the following the picking of the original wavedata within REFLEXW is described.
The most important part is the definition of the source- and receiver-coordinates. These coordinates are
stored within the traceheader of each trace. There are two different possibilities for both the refraction
and the transmission data.
Either each shot is imported (single shot analysis), filtered and picked separately (standard procedure
for “normal” refraction interpretation - see also refraction guide, Import the data and pick the first
onsets (done within the module 2D-dataanalysis).
Or one datafile contains all shots (several shots analysis - standard procedure for the reflection seismic
To be considered: In any case the traceheader coordinates must be correctly defined before picking! If
the 2D-data format shall be used the x- and y-traceheadercoordinates will be considered. The z-
traceheader coordinates will be ignored. In the following the definition for this 2D-data format for
equidistant receivers will be described.
Depending on the shot/receiver geometry three different data types can be distinguished:
single shot: surface/surface geometry - shots and receivers are located on the surface, e.g. seismic
refraction or reflection data
single shot/boreholes: borehole/borehole geometry - shots and receivers are located within two
parallel boreholes
single shot/VSP: borehole/surface geometry - the receivers are location within a borehole, the shots
are located at the surface
The shotfiles can be imported and handled completely separately or using a conversion sequence.
Enter the import menu and choose the correct data type and enter the overall 2D-fileheader geometry of
your shot and the receivers- see the tabella below. Depending on your original data it might be useful
to deactivate the import option read traceincr. Otherwise the end receiver coordinates might be
changed.
The equidistant receivers are The equidistant receivers are The equidistant receivers are
located at the surface from 5 located within borehole 1 (25 located within the borehole (at
to27 m. The shot is placed at 22 m along the surface) from 5 to 25 m along the surface) from 5
m. The lateral offsets of the 27 m depth. The shot is to 27 m depth. The shot is
receivers and the shot are 0. situated within borehole 2 (4 m situated along the surface at 4
along the surface) at 5 m depth. m.
In all cases the conversion sequence inline lines can be used in order to import the different shot files
within one step with an automatic actualization of the overall fileheader geometry and of the
traceheader coordinates. For the latter case choose
fileheader for update traceheaders within the update
traceheader/gps cordinates box. After having activated the
option convert to Reflexw you must select all original files
which fall in the selected geometry layout using the shft or
ctrl key. In the case of the converison sequence inline lines
the edit traceheader coordinates menu does not open after
having converted the files.
Surface/surface layout 1 with moving receivers and shots (e.g. seismic refraction or reflection data).
For a fixed line with moving shots simply deactivate move receivers.
After having imported all files it is possible to redefine the geometry of the single shots within the edit
several fileheaders menu (to be found under file):
! Select the wanted filepath and open files.
! Choose all wanted files and make any changes for the shot and receiver positions if necessary.
Rec-offs. and shot offs. describe the location along the surface. Rec-start and rec-end as well as
shot pos. describe the location within the boreholes.
! Choose "update trace headers" -> fileheader and click on save. You may check your geometry
After having imported all files it is possible to redefine the geometry of the single shots within the edit
several fileheaders menu (to be found under file):
! Select the wanted filepath and open files.
! Choose all wanted files and make any changes for the shot and receiver positions if necessary.
Rec-offs. and shot pos. describe the location along the surface. Rec-start and rec-end describe
the location within the boreholes. The shot is always situated at the surface at a depth of 0m.
! Choose "update trace headers" -> fileheader and click on save. You may check your geometry
using the option file/edit traceheader.
It is only possible to redefine the geometry within one step for the same shot/receiver layout. Therefore
if you have two different layouts as described within example 1 and 2 you must redefine the geometry
for these two layouts separately.
You can put together the shots during the import (conversion
sequence combine lines/shots) or afterwards using a special
processing option. The shots must have the same sample
number if using the import option conversion sequence
“combine lines/shots”.
2. load the first shot and put together all shots using the option
combine files f. CMP under processing/edit traces. Click on
load and choose all the other shots except the actually loaded
file (multiple file choice using the ctr or shft key). To be
considered: the sorting of the files is done automatically with
ascending alphabetic order of the filenames. Therefore a
renaming of the files e.g. 1.dat,...,9.dat to 01.dat,...,09.dat may
be necessary.
4. enter the geometry using the option CMP (see also chap.
1.12.4.1). Click on geometry and activate fixed line for standard
geometry. Enter the geometry within the standard geometry box.
The radio box standard line direction allows to define the direction of the standard geometry.
x-direction activated: the line (shots and receivers) is assumed to be orientated in x-direction
(surface/surface layout). This is the case for a seismic refraction dataset. Shot offset and receiver offset
define the constant offset in y-direction and should be set to 0 for seismic refraction data.
y-direction shots/rec. activated: the total line (receivers and shots) is assumed to be orientated in y-
direction (borehole/borehole layout). Use this option for example to define the geometry of a two
boreholes transmission measurement. Shot offset and receiver offset define the positions of the 2
boreholes along the x-axis (surface).
y-direction shots activated: the shots are assumed to be orientated in y-direction (surface/borehole
The parameters for the surface/surface configuration (layout 1 of chap. IV.1.1 with 12 receivers) are
the following:
nr. channels: 12
first trace: 73
last trace: 108
shot start: 12
shot increment: 5
shot offset: 0
receiver increment: 2
receiver offset: 0
First receiver: -7
left shot receiver: -1
right shot receiver: 1
last receiver: 15
x-direction shots/rec. activated
The parameters for the borehole/borehole configuration (layout 2 of chap. IV.1.1 with 12 receivers)
are the following:
nr. channels: 12
first trace: 1
last trace: 36
shot start: 27
shot increment: -5
shot offset: 4
receiver increment: 2
receiver offset: 25
First receiver: 5
last receiver: 27
y-direction shots/rec. activated
nr. channels: 12
first trace: 37
last trace: 72
shot start: 4
shot increment: 5
shot offset: 0
receiver increment: 2
receiver offset: 25
First receiver: 5
last receiver: 27
y-direction rec. activated
After having imported the data and having defined the traceheader geometry the traveltime data must
be picked.
For that purpose activate the option pick and pick the data using one of the picking options. Open the
pick save menu using the option save. The save picks menu opens (see figure on the right). In any case
the picks also should be saved using the Reflex Win format in order to have the possibility to load
them again in a later stage.
Use the format ASCII-2D tomography or ASCII-3D
tomography in order to generate the ASCII-file for a
subsequent tomography. With the option “export
several existing picks into 1 ASCII-file” activated you
may export several existing pickfiles into 1 ASCII-file.
The pick-file will have the extension TOM and will be
stored under the path ASCII under the current
projectpath.
1. First a starting model must be generated (see chap. I) or an already existing model must be loaded
using the option file/load model. Normally the starting model may be a simple homogeneous model
whereby the velocity should be within the expected range.
- For a first tomographic result you may use the other default parameters. There are no general rules for
these parameters but you have to adapt the parameters to your data in order to get the best result.
- Enter a name for the final model. Please use not the same
name as for the starting model because this may lead to
problems.
In the case of the 2D refraction vertical tomography all sources and receivers are located within one
line at the surface. In order to allow for a high data coverage within the medium vertical velocity
gradients should be present and a curved raytracing for the calculation of the traveltimes must be used.
The curved rays are calculated using a finite difference approximation of the Eikonal equation (see
raytracing). A start model must be defined. No assignment to layers is necessary.
The start model should contain a quite strong vertical velocity gradient and the max. velocity variations
should be large enough (e.g. 200 % of the original values) in order to enable strong vertical gradients at
those positions where an interface is assumed. A smoothing in horizontal direction is often useful
because of the normally quite large receiver increments.
- Load the data using the option load data (see also item 1). If the 3D-data format is used for the 2D-
tomography you have to deactivate the option use 2D-data and you have to specify the second
coordinate (y or z) within the radiobox sec.coord. The first coordinate is always x. The third
coordinate is neglected.
- Enter the wanted space increment (equal in x- and z-direction). Normally this increment should be
small enough in order to allow small scale variations with depth. It may be significantly smaller than
the receiver increment.
- Often it is useful to force the first iteration (option force 1.iter. activated) to generate a new model
even if the resulting residuals are larger than for the starting model.
- For a first tomographic result you may use the other default parameters. There are no general rules
for these parameters but you have to adapt the parameters to your data in order to get the best result.
- Enter a name for the final model. Note: do not use the same name like for the starting model.
- Start the tomography. The tomographic result is stored using the “normal” REFLEXW format. You
may display the result within the 2D-dataanalysis.
4. Control the tomographic result by a forward raytracing. For that purpose activate the option ray.
The raytracing menu opens in addition.
Load the traveltime data using the option File/load data traveltimes. Then the screen is split vertically
showing in the upper window the model together with the tomographic result and in the lower the data.
Now the ray tracing parameters have to be chosen:
- enter the wanted raytracing type FD-Vidale.
- enter the gridding increment DeltaX - this increment must be equal to the increment used for the
tomography.
- enter the output-scale, e.g. 1
- enter the calculate type - in this case data traveltimes because we want to simulate all loaded
observed traveltimes
- enter the outputfile name
- deactivate the option raster
- start the raytracing using the option start. As the option raster is deactivated you are asked for the
raster file. Choose the tomography raster file.
- the calculated traveltimes are shown in the lower picture in addition. Now you may check for the
mean traveltime difference using the option Analyse/calculate traveltime differences