Introduction

This tutorial will guide you through a complete modeling session to become familiar with the main
module functionalities. The tutorial starts with the subsampling of a displacement map associated with
an Mw 6.6 earthquake that occurred near the city of Bam (Iran) on December 26 th, 2003. After importing
the geodetic data, the coseismic displacement is modeled via Non-Linear and Linear inversions, to
retrieve the geometric parameters of the fault and its related slip distribution. The latter is then used
to calculate the change in stress generated by the earthquake on the fault itself; finally, the retrieved
model is used in a direct (forward) modeling to generate the raster map of the modeled (predicted)
displacement, projected onto the SAR Line-Of-Sight.

Though this tutorial covers the main modeling features, it cannot be considered exhaustive of all the
software potentialities. In most cases, only one of the possible processing configurations is presented:
we strongly suggest the user to refer to the Help manual for a complete description of all the available
options. At the beginning of every paragraph, the keyword to retrieve a topic from the help index is
reported.
Since the characterization of a geophysical source is generally non-unique, different solutions can lead
to comparable results, equally able to reproduce the observed data. Only experience and the combined
analysis of all the available data (e.g. geodetic, geological, seismological) can help to find the most
realistic solution.
Tips and notes are also provided to improve the modeling strategy and discover further functionalities.

The Bam event data

This tutorial is based on the displacement map generated with a pair of SAR images acquired by the
Envisat satellite from the European Space Agency. This map has been obtained with an interferometric
processing carried out with the SARscape interferometric module; the images have been acquired on
December 3rd, 2003 and February 11th, 2004 and have a spatial baseline of about 4 m. In this tutorial
the displacement is modeled with a single fault with distributed slip.
Refer to the cited literature (Wang et al., 2004; Funning et al, 2005) for a review of the event.

REFER TO YOUR LOCAL DISTRIBUTOR TO GET THE DATA FOR THIS TUTORIAL.

To run the tutorial, you need the following files:

 Bam_envisat_dsc_disp: displacement map (m) obtained from a descending Envisat image
pair;
 Bam_envisat_dsc_ALOS: map of the Line-Of-Sight azimuth angle (deg);
 Bam_envisat_dsc_ILOS: map of the Line-Of-Sight incident angle (deg);
 Bam_SRTM_dem: SRTM digital elevation model with 90 m resolution;
 Bam_project_backup.xml: XML Project File (see next paragraph);
 Subsampling_areas.shp: shapefile containing the areas to sample the displacement map.
All the images are in a UTM-WGS84, Zone 40 North, projection.

The XML Project File

Help Index: “XML_Modeling_Project_file”

1
In the main modeling panel, an XML Project File must be set (see the related page in the help). This
file allows to save the input configuration of any processing for a quick restore in any moment. Most of
the processing outputs are stored in the XML Project File as well, thus allowing also an efficient project
portability.
In the tutorial, an XML Project File created from scratch and named Bam_project.xml will be used.
Together with the input data, we also provide a project with all the sections already present (the
Bam_project_backup.xml file). This permits the user to avoid the setting of parameters, which can
be directly restored into the panels; however, we suggest the use of a new project file in order to get
confident with them.

Image subsampling

Menu: “SARscape”-> “Interferometry”->“Displacement Modeling”->“Tools”->”Image Subsampling”

Help Index: Image subsampling, InSAR_dataset

Image subsampling can be carried out with two different approaches: over areas with equally spaced
points or with the Quadtree algorithm. In this tutorial we use the first approach, which needs the
definition of the areas to sample and, for every area, the distance between points.
We define the sampling areas through the Subsampling_areas.shp file, where two zones are
defined, respectively with a 500 m sample spacing close to the fault and 2000 m one far from it (Fig.
1). This file is provided together with the input data.
Tip You can define your own subsampling areas, creating a new shapefile and setting the resolution
values in the “.dbf” file. See the “Image_subsampling” help page to know the rules to create a
new shapefile for subsampling.

Figure 1. Sampling areas (left) and resulting points (right)

Once the file with the areas to sample has been defined, the subsampling can be carried out:
1. Open the “Image Subsampling” panel;
2. Set the following files:
 “Subsampling Image”: Bam_envisat_dsc_disp;
 “Azimuth LOS image”: Bam_envisat_dsc_ALOS;
 “Incidence LOS image”: Bam_envisat_dsc_ILOS;

2
  “DEM”: Bam_SRTM_dem;
3. Set the “Select subsampling method” to “Mesh from vector file”;
4. Set in “Sampling Areas” the Subsampling_areas.shp file (or a newly created one);
5. Set in “Output Shapefile” the Bam_sampled_points.shp name;
6. Click “Start” and wait for the “END” message.

If you sample the image using the Subsampling_areas.shp file provided with this tutorial, the
Bam_sampled_points.shp will contain 7448 points, shown in Fig. 2 with the related attributes.
To view the image in colors based on displacement, drag the file from the folder in which it was saved
into the ENVI environment. Right click and click on ‘Properties’. In the pop-up window, under ‘Select
Attribute’ choose ‘Observed’, and under ‘Color Table’ choose ‘Rainbow’.

Figure 2. Subsampled points with the observed displacement and related attributes

3
In addition to the usual shapefile components (.shp, .dbf, .shx, .prj) a further file has been created,
the Bam_sampled_points.shp.xml file, which contains ancillary information required by the
inversion; see the “InSAR_dataset” help page for further details.

Non-Linear Inversion

Menu: “SARscape”->”Interferometry”->“Displacement Modeling”->“Non-Linear Inversion”.

Help Index: NL-Inversion

With the Non-Linear Inversion we try to reproduce the observed displacement by means of a
geophysical source. We assume here that nothing is known about the source and all its parameters
must be inferred from InSAR data. We model the data by means of a single fault, accounting also for
a possible offset that might affect the data, since the reference point in the interferogram could have
been set in a deforming area.

Before adding the dataset to invert and initializing the source, the XML Project File must be set:

1. Open the “Non-Linear Inversion” panel;

2. Set the “XML Project File” to Bam_project.xml, selecting the “New…” button.

Dataset setting
Help Index: InSAR_dataset
The dataset to invert, obtained with a subsampling of the displacement map, is added as follows:

1. Click on the “Add…” button in the “INPUT Datasets” list and select the
Bam_sampled_points.shp;
2. Open the Parameter setting panel through the “Set…” button;
3. Set “Constant offset” as “Account for an orbital surface” option;
4. Click on “Commit”.

By setting “Constant offset”, we let the inversion to assess also a possible offset that might affect the
data.

Source setting

Help Index: “Initialize_Values_from_CMT_Cat”, “Elastic_dislocation_Okada”

Though the source parameters are assumed to be completely unknown, we must set, for each
parameter, a range of allowed values between a minimum and a maximum. Since the elastic dislocation
source is characterized by a high number of parameters, we developed a specific tool to initialize these
values from the Global Centroid Moment Tensor catalogue. The only thing to know is the CMT event
identifier, which for the Bam earthquake can be obtained from the Global CMT website as follows:
1. Go to the http://www.globalcmt.org/CMTsearch.html site;
2. Set the following parameters for the:
 Starting Date - Year: 2003
 Starting Date - Month: 12”
 Starting Date - Day: 26”
 Ending Date - Number of days: 1
 Moment magnitude range with 6 <= Mw <= 7;
3. Click on “Done”;
4. Two events should match these constraints: copy the identifier of the “SOUTHERN IRAN” event
(122603B).

4
Once the CMT event identifier is known, initialize the source to invert as follows:

1. Click on “New…” button from the “INPUT Sources” tab;

2. Select “Elastic dislocation (Okada)” from the pull down menu;
3. Click on the “Initialize source from Global CMT catalog…” button;
4. Set the event identifier in the “Insert the CMT id” field and click the “Search” button (it could
take a while to get the result);
5. From the pull down menu select “Plane 1” then click “Commit”;
6. Set “Bam_fault” as “Source name”, then click “Commit” to return to the main Non-Linear
inversion panel;
7. Click on “Save” button to save the XML Project File updates;
8. Click on the “Start” button to run the inversion;
9. Wait until the “END!” message (it could take few minutes);
10. Click “No” to the “Focal Mechanism” dialog box.

Figure 3. Source Parameters initialized with the CMT solution

Tip The CMT catalogue always provides two alternative solutions, corresponding to the real slipped
plane and the auxiliary one. In the case of Bam, the displacement pattern allows to state that the
slipped fault is oriented North-South, therefore the “Plane 1” must be selected.
Note When the XML file is updated, the “NLInput” and ”NLOptions” sections are created under the
“ModelingRoot”-”NonLinearInversion” XML section. If they already exist from a previous
processing, they are overwritten. In order to maintain the alignment between input and output,
when the “NLInput” and ”NLOptions” sections are updated, the “NLOutput” section, if existing
from a previous processing, is canceled. The new “NLOutput” section is written at the end of the
inversion.
When the Non-Linear Inversion ends, the Bam_project.xml file is updated to store the results: source
parameters, data offset, RMS of the residuals between observed and modele data, geodetic moment,

5
moment magnitude, etc… All the information are written in the “ModelingRoot”-”NonLinearInversion”-
”NLOutput” section.
Note the Non-Linear Inversion algorithm is based on multiple restarts where every starting
configuration is randomly picked within the allowed ranges: this is intended to avoid, as far as
possible, to trap the cost function into local minima. At the same time, this can lead to different
results for different runs.

Tip We strongly suggest to run the Non-Linear inversion several times and inspect every time the
inferred source, in order to understand the reliability and the stability of a result.
Here we show a result obtained with the source initialized as in Fig. 3. The following files have also
been created:

 Bam_fault_nonlinear.shp, containing the best fit source with its parameters.

 Bam_sampled_points_nonlinear.shp, containing the observed and modeled data.

The results of Fig. 4 show that the source automatic initialization with CMT values already leads to
reasonable results. However, at this stage, further improvements are still possible.

Tip When the inverted parameter reaches the upper or the lower limit of the allowed range, this range
should be modified to let it find the best fit value. In general, a good result is obtained when all
the parameters fall within the specified range without assuming a limit value. Check the “Report”
panel, after the inversion, to see which parameters need to be better tuned (they have three
asterisks “***”).
Tip When a parameter is already known and doesn’t need to be inverted, set minimum and maximum
equal to the desired value.

6
 From the XML Project File

Overall residual RMS 0.049 m

Geodetic Moment 1.26 ∙1019 N∙m

Moment Magnitude 6.70

Assessed orbital contribution -0.0139 m

Figure 4. Modeled points, fault parameters and other general parameters after the Non-Linear
Inversion automatically initialized with the CMT values.

In Fig. 4 the fault dip corresponds to the upper limit, as also shown in the Report panel, suggesting
that higher values could be preferred; however, since most of the signal is on the eastern side of the
fault (see Fig. 1), the fault is likely to dip eastward and not westward, as proposed by the CMT solution.
Because of the convention on the strike and dip angles (see “Elastic_dislocation_Okada” in the help),
to let the source dipping eastward, also the strike range must be changed. Through the “Edit…” button
on the “INPUT Sources” section, change the “Strike” and “Dip” ranges according to Fig. 5
Note: It is a good idea to delete the already created datasets from ENVI not to confuse the program.
This can be done via the “Data Manager” (top left corner of the ENVI environment).

7
 Figure 5. Source parameters modified after the first inversion

After that, re-run the Non-Linear inversion and verify that all the inverted parameters fall within the
allowed ranges. You can check the inverted parameters from the tab “OUTPUT Sources” selecting your
source and the “View…” button (Figure 6, this tab appears when the Non-Linear inversion step is
completed), or through the Report tab, shown below.

Figure 6. Source parameters obtained from a second trial of the Non-Linear inversion
After any inversion, a panel with a synthetic report is created, to immediately assess the quality of the
results, as shown in Figure 7. Information are provided about the modelled sources (source parameters,
geodetic moment, event magnitude, etc.) and datasets (rms of the residuals, rms of the original dataset,
suggested dataset weight, orbital parameters assessed for InSAR, etc.). Parameters to fix after a non-
linear inversion are highlighted with an “***”: this suggests the change of the parameter searching
range.

8
 Figure 7. Report tab containing the main parameters of the source obtained from the Non Linear
inversion step and Output dataset RMS, Offset, etc.
Parameters highlighted with an ”***” in Fig. 7 can be adjusted by changing their limits as shown in
Fig.8 and the Linear Inversion step can be re-run obtained the new results shown on the report on Fig.
9, where no parameters are requiring additional fixing.

Figure 8. Source parameters modified after the second inversion

9
 Figure 9. Report tab containing the main parameters of the source obtained from the Non Linear
inversion step and Output dataset RMS, Offset,…

Once a satisfactory set of source parameters are obtained, a functionality allows the assessment of the
modelled source parameter uncertainties and all the parameter trade-offs, as shown in Figure 10.

Figure 10. “Calculate Statistics” button

10
This is carried out by adding some ad hoc noise (correlated in space, for InSAR) and performing the
inversion for an arbitrary number of times, set by the user. The output is a grid of scatter plots and
probability distributions involving all the parameters retrieved by inversion, as shown in Figure 11.

Figure 11. Modelled Source parameter uncertainties and parameter trade-offs.

11
 From the XML Project File

Overall residual RMS 0.049 m

Geodetic Moment 4.66 ∙1018 N∙m

Moment Magnitude 6.41

Data offset -0.0176 m

Figure 12. Modeled points, fault parameters and other general parameters after the Non-Linear
Inversion with revised ranges for Strike and Dip.
Note Different solutions can fit almost equally the observed data, as visible by comparing the results
of Fig. 4 and Fig. 12, which have almost the same RMS for different sources. Only external
information and experience allow to judge whether a solution is reliable.

Linear inversion

Menu: “SARscape”->“Interferometry”->“Displacement Modeling”->“Linear Inversion”.

Help Index: “Linear_Inversion”
With the Non-Linear inversion we modeled a source characterized by a unique slip value, which is a
mean fault dislocation of 2.42 m (Fig. 12 and Fig. 9). With the Linear Inversion we retrieve a realistic
distribution of slip over the fault plane.
As for the Non-Linear Inversion, this task requires the setting of the XML Project File;

12
 1. Open the “Linear Inversion” panel
2. Open the Bam_project.xml as “XML Project File”
Note Through the tutorial, we add every task to the same XML Project File. However, every task is
independent from the others and it can be stored in a different XML Project File.

Dataset setting
To load the dataset to invert, follow the same steps of the Non-Linear case:

1. Click on the “Add…” button in the “INPUT Datasets” list and select the
Bam_sampled_points.shp;
2. Select the dataset and open the Parameter setting panel through the “Set…” button;
3. Set “Constant offset” as “Account for an orbital surface” option;
4. Click on “Commit”.

Source setting

Help Index: “Fault_patches_Okada”, “XML_Modeling_Project_file”

The starting point for the Linear Inversion is the source calculated by the Non-Linear inversion: we do
not create a new source with the “New…” button, but we load the Non-Linear inversion results. This
can be done in two ways: loading the Bam_fault_nonlinear.shp through the “Add from file…” button
or retrieving the fault parameters with the “Add from XML…” button, which allow exploring the content
of any XML Project File.
Note The “Add from XML…” button allows accessing any XML Project File, to inspect and retrieve any
stored source (input/output of Non-Linear Inversion, Linear Inversion, CFF Stress Transfer,
Forward Modeling). In this tutorial, we access the same Bam_project.xml file to get the output
of the Non-Linear inversion.
To load the source stored in the XML Project File after the Non-Linear inversion as input for the Linear
Inversion, follow these steps:

1. Click on the “Add from XML…” button under the “INPUT Sources” list;
2. Select the Bam_project.xml file;
3. In the XML section, select “Non-Linear Inversion Output”;
4. Select the “Bam_fault” item;
5. Click “Commit”.
Tip Since the slip retrieved through Non-Linear Inversion is a mean value of the real one, a good
strategy in the Linear Inversion setting is to extend that source to let the slip vanish at the fault
limits (Fig. 13).

13
 Figure 13. Fault from Non-Linear Inversion (gray rectangle) extended and subdivided into 30x15
patches; blue points represent the observed displacement in the Line-Of-Sight

The source retrieved from the Non-Linear inversion is added to the “INPUT Sources” list but more
information must be provided before running the inversion. To check the missing parameters and to
change the source dimension:

1. Select the source in the “INPUT Sources” list and click the “Edit…” button;
2. Set the “Length (m)” field to “30000.”;
3. Set the “Width (m)” field to “15000.”;
4. Set the “Depth (m)” field to “0.”;
5. Change the “Reference Point” to “(4) Fault Trace Center – Along Dip Top Edge”;
6. Set the “Patches Along Strike” field to “30” and “Patches Along Dip” field to “15”;
7. Set “Distributed Slip with fixed rake” under the “Linear Inversion Input Options”;
8. Set “Non-Negative Least Squares” under the “Linear Inversion Input Options”;
9. Only for ENVI 5.0 or higher, click the “Draw Source in ENVI 5.x” button to see the source in
the ENVI view (a reference layer must be already present);
10. Click the “Commit” button;
11. Set the “Damping value” to 0.005 in the “Options”;
12. Save the XML Project File;
13. Start the inversion;
14. Answer “No” to the “Focal Mechanism” dialog box after the inversion ends.
We switched to the Point Reference 4 (see “Elastic_dislocation_Okada” in the help) to change the
source length and width without affecting the East and North coordinates.
In this way, we extended the initial fault to a 30x15 km source, subdivided into patches of 1 km. After
clicking on the “Start” button, the Bam_project.xml file is updated by creating, or updating, the
”LinInput” and ”LinOptions” sections in “ModelingRoot”-”LinearInversion”. In Fig. 14 we show the
results of the Linear Inversion.

14
 From the XML Project File

Overall residual RMS 0.008 m

Geodetic Moment 0.53 ∙1019 N∙m

Moment Magnitude 6.45

Data offset 0.0007 m

Figure 14. Modeled points, slip distribution and other parameters after the Linear Inversion

Tip The slip distribution strongly depends on the Damping Value: a low damping leads to scattered
slip values while a high values to an over-smoothed solution. The best value can be obtained only
with a trial-and-error approach. Unless you need a complete damping vs. RMS curve, we suggest
a divide-and-conquer algorithm: start from a low value, then move to a high value, then test a
value in between, and so on… checking at every step the result. This is quite fast, if the output
shapefile is loaded into a GIS: since the output shapefile, at every step, is overwritten, you just
need to refresh the view to check the result for every new damping value.
The following output files have been created:

 Bam_fault_linear.shp: containing the best fit source with its parameters;

 Bam_sampled_points_linear.shp: containing the observed and modelled data.
The main parameters obtained as output of the Linear Inversion can be visualize from the Report tab
(this tab appears when the Linear inversion step is completed). Moreover, selecting the Output Source
tab (this tab appears when the Linear inversion step is completed) and the button “View…” the slip
value of each patch can be read, selecting the corresponding “Feature#” value, as shown in Figure 15
where Feature# 25 has been selected.

15
 Figure 15. Source parameters obtained from the Linear Inversion. Slip result for each patch can be
seen selecting the corresponding “Feature#” number.

Calculate the moment tensor and draw the “beach ball”

Menu: “SARscape”-> “Interferometry”->“Displacement Modeling”->“Modeling Tools”->”Calculate and

Draw Focal Mechanism”.

Help Index: “Calculate and Draw Focal Mechanism”

After any Non-Linear or Linear Inversion, it is possible to automatically calculate the seismic moment,
the moment tensor and draw the beach ball of the focal mechanism, which can be exported to a .pdf
file.
This can be also done in any moment, through the “Calculate and Draw Focal Mechanism” panel,
loading in the “INPUT Sources” list the sources for which you want to do the calculation. A source can
be created from scratch or loaded from a shapefile or an XML Project File. In this tutorial, we load the
results of the Linear Inversion from the Bam_project.xml file, where they have been stored after the
inversion:
1. Open the “Calculate and Draw Focal Mechanism” panel, under “Modeling Tools”;
2. Click on the “Add from XML…” button;
3. Select the Bam_project.xml file;
4. Form the “XML section” pull down menu, select “Linear Inversion Output”;
5. Select in the source list the “Bam_fault [Fault patches(Okada)]” item and click “Add”;
6. Select the imported source and click “Start” (Fig. 16).

16
 Figure 16. Moment tensor and focal mechanism of the Bam earthquake
Tip You can test, for instance, the effect of changing the shear modulus  just editing the source,
through the “Edit…” button, then setting a different value in the “’mu’ Lame’s constant” field.

CFF Stress Transfer

Menu: “SARscape”-> “Interferometry”->“Displacement Modeling”->”CFF Stress Transfer”

Help Index: “CalculateCFF”

This tool can be used to calculate the stress change induced by any source, or a list of sources (input
sources), into another source, or a list of sources (receiver sources). In the case of Bam, only one
source is involved, we therefore calculate here the stress change induced by the Bam fault on the fault
itself.
Tip The self-induced stress change can be used to verify if the aftershock distribution over a fault
plane is in agreement with that expected, i.e. whether they are located where the stress change
has the highest values.

As for the Non-Linear and Linear Inversions, this task requires the setting of the XML Project File,
which is the same Bam_project.xml file:
1. Open the “CFF Stress Transfer” panel;
2. Set the “XML Project File” to Bam_project.xml.
The input and receiver sources are retrieved, as in previous tasks, from an XML file (which is the
same Bam_project.xml):
1. Click on the “Add from XML…” button under the “INPUT Sources” list;
2. Select the Bam_project.xml file;
3. Form the “XML section” pull down menu, select “Linear Inversion Output”;
4. Select in the source list the “Bam_fault [Elastic dislocation (Okada)]” item and click “Add”;
5. Click on the “Add from XML…” button under the “Receiver Sources” list;
6. Repeat the steps from 2-4;
7. Save the XML Project and click on the “Start” button in the main panel; wait for the “END”
message.
The output containing the CFF value is stored in the XML Project File, and a shapefile is generated:

17
  Bam_fault_cff.shp;
To inspect the result the shape file is loaded in ENVI or it can be loaded into a GIS software, as in Fig.
17.

Figure 17. CFF stress change induced by the slip of Fig. 14.

Forward Modeling

Menu: “SARscape”->“Interferometry”->“Displacement Modeling”->”Forward Modeling”

Help Index: “ForwardModeling”

This task requires the setting of the XML Project File:

1. Open the “Forward Modeling” panel;

2. Set the “XML Project File” to Bam_project.xml.

In this step, we use the results of the Linear Inversion, the distributed slip on the Bam fault, to generate
a raster map of the surface displacement. The Forward Modeling can be carried out for a set of points
stored in a shapefile or creating three raster files containing the East, North and Up components. We
create here a raster output, using the Bam_envisat_dsc_disp as a reference for the extent and
resolution (i.e. the output of the interferometric processing, used as input for the image subsampling),
but with a reduced resolution of 250 m instead 25 m.

Note As explicitly written in the Help, any processing performed with the modeling tools must be
carried out in a projected cartographic system; in the case of Bam, data are in the UTM-WGS84,
Zone 40 North, projection.

To create the raster maps of the surface displacement:

1. Click on the “Add from XML…” button under the “INPUTSources” list;
2. Select the Bam_project.xml file;
3. From the “XML section” pull down menu, select “Linear Inversion Output”;
4. Select in the source list the “Bam_fault [Elastic dislocation (Okada)]” item and click “Add”;
5. In the “Forward Model Output” pull down menu of the Option tab, select “Raster”;
6. Upload the reference DEM (in UTM Projection, Hemisphere Nord, zone 40, Ellipsoid WGS84);

18
 7. Click on “Reference raster…” and select the Bam_envisat_dsc_disp file;
8. Change, if necessary, the “Output File” name and path automatically created;
9. Check on the “Set raster info…” button;
10. Change to 250 the “Cell size”;
11. Click on “Commit” to close the “Set cartographic system” panel;
12. Click on “Commit” to close the “Raster Parameters”;
13. Save the XML Project File;
14. Click on “Start” and wait for the “END” message.
Unless you changed the output name, automatically generated, the processing will create the following
files (Fig. 18):

 Bam_envisat_dsc_disp_forward_east: with the displacement East component;

 Bam_envisat_dsc_disp_forward_north: with the displacement North component;
 Bam_envisat_dsc_disp_forward_up: with the displacement Up component.

Figure 18. East, North and Up components of the modeled displacement.

Note The displacement measurement unit is always coherent with the input source: when the slip is
provided in meters, the output map will be in meters.
As a further information, the Bam_project.xml file has been automatically updated and the section
“ModelingRoot”-“ForwardModeling” has been created. You can navigate through the XML tags to
inspect its content.

Project raster to LOS

Menu: “SARscape”->“Interferometry”->“Displacement Modeling”->“Tools”->”Project raster to LOS”

Help Index: “Project_Raster_to_LOS”
In the last step, the three maps generated with the Forward Modeling panel are combined together to
project them into a single Line-Of-Sight. This simple panel just needs the three raster maps and the
ALOS (Azimuth Line-Of-Sight) and the ILOS (Incidence Line-Of-Sight) images.

Note The group of the three displacement maps and the ALOS/ILOS pair have in general a different
extent and resolution; the extent and resolution of the output raster can be independently set
according to any of the two groups.
Tip When several LOSs are available (perhaps Ascending and Descending) the Forward Modeling tool
can be used just once, to create the East/North/Up displacement maps; then they can be projected
into as many LOSs available.

To run the LOS projection:

19
 1. Open the “Project raster to LOS” panel;
2. Set the following files:
 “East component”: Bam_envisat_dsc_disp_forward_east
 “North component”: Bam_envisat_dsc_disp_forward_north
 “Up component”: Bam_envisat_dsc_disp_forward_up
 “Azimuth LOS image”: Bam_envisat_dsc_ALOS
 “Incidence LOS image”: Bam_envisat_dsc_ILOS
3. Set the “Output image” to Bam_envisat_dsc_disp_forward_los;
4. Set the “Raster extent” to “Same as the ALOS/ILOS components”;
5. Set the “Raster resolution” to “Same as the ALOS/ILOS components”;
6. Define the “output image” name and folder;
7. Click on the “Start” button and wait for the “END” message.
After that, it is possible to compare the Forward Modeling results projected into the LOS with the original
displacement map, as in Fig. 19.

Figure 19. Comparison between the observed (left) and the modelled (right) displacement maps in
the Line-Of-Sight

Note To make a correct comparison of the observed and the modeled displacement maps, it must be
taken into account that in the inversion an orbital surface might have been calculated.
In this case, we calculated in the Linear Inversion an offset of 0.017 m (see Fig. 14), which must be
added to the modeled displacement of Fig. 19 (right) to obtain the complete modeled displacement.

References

Funning, G.J., B. Parsons, T.J. Wright, J.A. Jackson and Fielding, E. J. (2005), Surface displacements
and source parameters of the 2003 Bam (Iran) earthquake from Envisat advanced synthetic aperture
radar imagery, J. Geophys. Res., 110, B09406, doi:10.1029/2004JB003338.

Wang, R., Xia, Y., Grosser, H., Wetzel, H.-U., Kaufmann, H. and Zschau, J. (2004), The 2003 Bam (SE
Iran) earthquake: precise source parameters from satellite radar interferometr, Geophy. J. Int.,
159: 917–922. doi: 10.1111/j.1365-246X.2004.02476.x

20