Park Seismic LLC, Shelton, Connecticut, Tel: 347-860-1223, Fax: 203-513-2056, Email: parkseis@parkseismic.com
ParkSEIS | Generation of Back Scattering Analysis (BSA) Section - "Working with Sample Data"

1.  Importing Input Data

Seismic Data with Source/Receiver (SR) Setup [*(SR).dat]
In the main menu, go to "Process" → "Back-Scattering Analysis (BSA)" and then select the sample data in "...\Sample
Data\VOID\VOID(SR).DAT."
2.  Generating BSA Section

Click "Run BSA" button.  "Dispersion" and "Output" will be automatically executed in the order.
Import the dispersion curve file ("VOID.DC") that was used to generate the sample data set.  You can also use your own
dispersion curve file processed from the sample data by following the normal procedure of dispersion analysis as explained in
section 4 below.  Then, specify output file name.
After this, the BSA operation will be executed and the final BSA section will be displayed.  Default display mode for BSA
section is a variable-area display as shown below.  The target back scattering feature originating from the surface location of
the anomaly (44-m) is indicated in the figure below.  All other less conspicuous slant features are computational artifacts.
3.  Depth Evaluation
Now, successive narrow-band-pass (NBP) filtering will be applied to this data set to see when the back-scattering feature
completely disappears.  Choose "Process" tab, and then click "Filter" button.  Double click anyplace in the display.  The
filtering dialog will appear.
Specify four (4) filter parameters as shown below to apply NBP filtering centered at 20 Hz.
The output from the NBP filtering below shows the feature still remaining.  Then, another NBP filtering with a higher center
frequency now has to be applied.  Currently, the "PROCESSED" (i.e., "NBP filtering applied") data is being displayed.
First, the original data before filtering has to be displayed in the window.  For that, click the toggle button (the one marked on
the right below) to display the data without "(PROCESSED)" title on top.  Then, click the "Filter" button and double click the
display to launch the filtering dialog again.
This time, specify four (4) filter parameters as shown below to apply NBP filtering centered at 30 Hz.
The output from the NBP filtering still shows the back-scattering feature, but with much less amplitude this time.  Therefore, it
may be necessary to apply, at least, one more filtering with even higher center frequency.
Now, specify four (4) filter parameters as shown below to apply NBP filtering centered at 40 Hz.
This output shows the back-scattering feature almost disappeared.  Now, try to apply another round of filtering to see if it
removes the feature completely.  
Now, specify four (4) filter parameters as shown below to apply NBP filtering centered at 50 Hz.
Now, the feature seems to be completely removed.  So, the phase velocity at 50-Hz is about 150 m/sec as shown in Figure 12
in Overview section.  Then, its wavelength is 3 m.  So, the top of the anomaly is deeper than 3 m.  Actual depth is 4 m.  
4.  Generating Average Dispersion Curve
This section describes how to prepare one dispersion curve that represent the average dispersion property over the entire
surveyed distance.  Although you can import all the dispersion curves processed from all the records in the input seismic data
set, the most optimal BSA output is often generated when this "representative" dispersion curve is used for the process.  This
section will demonstrate how to generate all individual dispersion images from the sample data set (of 50 records).  Then,
they will be stacked to produce one "stacked" dispersion image, from which the representative dispersion curve will be
extracted and saved to be imported at the beginning of the BSA process.

In the main menu, go to "Dispersion" → "Make Dispersion Image From S/R Coded Data" and import the same data set [VOID
(SR).DAT].
Then, click "Run Dispersion Image" button in the dialog.
In the main menu, go to "Display" → "Dispersion Data" → "Dispersion Image" and then select the previously processed data
set [VOID(SR)(ActiveOT).DAT].
Choose "Multi-OT" tab, and click "Stack" button to specify the output file name [*(VStack).DAT].
It will ask how the surface coordinate of the stacked dispersion image should follow.  It does not matter for the purpose of
getting only one dispersion curve.  Click "Yes" and then "No" to the option to display the stacked dispersion image.
Then, in the main menu, go to "Dispersion" → "Get Dispersion Curve(s) From Image Data" and import the previously saved
stacked image [*(VStack).DAT].
The following display window in which you can extract the dispersion curve from the image will appear.  Although detailed
instructions about how to extract a dispersion curve from the dispersion image is explained in the user guide of "Dispersion
Curve Extraction (Vs1D)" [or "*(Vs2D)"], the simplest sequence would be, first, click the "Bounds" button and then click several
points (e.g., five or more) on the dispersion trend to mark the lower and upper bounds within which the curve will be extracted
based on the maximum amplitude at each frequency.  Then, click the "Extract" button to extract the curve shown in the image
below.  Click "Save" button to save the extracted curve (*.DC).  This is the "average" dispersion curve that is representative of
the entire survey area and you can import it during the BSA operation.

In the main menu, go to "Display" → "Seismic Data."  Then, open the file previously saved [*(BSA)(HE=5).DAT], which has
been filtered for horizontal events as explained in section 3 of "
Common-Offset Sections."