ProAZ Guide
ProAZ Guide
Guide to ProAZ
Introduction
ProAZ is the program from Hampson-Russell which you can use to evaluate Amplitude Versus
Azimuth (AVAz) anomalies. The AVAz process allows one to detect anisotropy and then from
this infer information about fractures and the stress state.
This tutorial takes you through a typical AVAz project consisting of the following steps:
Starting Geoview
Geoview is the name of the Hampson-Russell program which contains all available options,
including AVAz analysis.
To start this tutorial, first start the Geoview program. On a Unix workstation, do this by going to
a command window and typing:
geoview
On a PC, start Geoview by clicking the Start button and selecting the Geoview option on the
Programs > HRS applications menu, or by clicking the Geoview icon on your desktop if one
has been set up.
December 2013
2 Guide to AVAz Analysis
When you launch Geoview, the first window that you see contains a list of any projects
previously opened in Geoview. For example, the figure below shows a series of pre-existing
projects, which you can open by clicking on the relevant project. Your list will be blank if this is
the first time you have ever run Geoview.
For this tutorial, start a new project. Before doing that, you can set all the data paths to point to
the location where you have stored the tutorial data. To do that, click the Settings tab:
December 2013
Guide to AVAz Analysis 3
Now you can see a series of default locations for the Data Directory, Project Directory, and
Database Directory. You can change all of these to point to the directory where the tutorial
data is stored.
To change all of the directories to the same location, click on the option Set all default
directories and then click the button to the right:
Then, in the File Selection Dialog, select the folder which contains the tutorial data:
After setting all three paths, the Geoview window will now show the selected directories (note
that yours may be different from what we show below). When you have finished setting all the
paths, click Apply to store these paths:
December 2013
4 Guide to AVAz Analysis
Now click the Projects tab and choose the option to Create New Project:
A dialog appears to set the project name. Call it Erksine_Azimuthal, as shown below. Enter the
project name and click OK on that dialog:
December 2013
Guide to AVAz Analysis 5
Now a dialog appears, asking you the name of the database to use for this project. The database
is used to store all the wells used in this project. By default, Geoview creates a new database,
with the same name as the project and located in the same directory. For example, this project is
called Erskine_Azimuthal.prj, so the default database would be called Erskine_Azimuthal.wdb.
That would be desirable if we were starting a new project, intending to read in well logs from
external files. For this tutorial, to save time, we have already created a database, which has the
well already loaded. To use that database, click Specify database. You will see a pop-up menu
appear below this button. Choose Open:
Open and then choose the erskineWells.wdb file from the dialog. The dialog then shows the
selected database and the new project name. Click OK to accept this:
December 2013
6 Guide to AVAz Analysis
The Geoview Window now has the new Project Erskine_Azimuthal as the first line in the
Projects window, similar to shown below:
The next step is to load the seismic volume. On the far left side of the Geoview window, click
the Seismic tab:
December 2013
Guide to AVAz Analysis 7
The window to the right of this tab shows all seismic data loaded so far. This is empty. Go to the
bottom of the window (on the left of the workspace) and click the Import Seismic button:
On the dialog that appears, select the file erskineGath.sgy and then click Next at the base of the
dialog:
December 2013
8 Guide to AVAz Analysis
Set the Geometry Type to 3D (this is the default) and click Next:
On the third page, tell the program what information it can use from the trace headers. In this
data set, we have the inline and crossline numbers; and the source and receiver coordinates, so
leave all the defaults intact, and just click next.
This brings up the next dialog in which the headers are defined. By default, this page assumes
that the seismic data is a SEG-Y file with all header values filled in as per the standard SEG-Y
convention. In this case, you must redefine byte locations for the inline and crossline numbers as
shown. If you are not sure that is true, click Header Editor to see what is in the trace headers.
December 2013
Guide to AVAz Analysis 9
For this dataset the Inline header starts at byte 1, the Xline header at byte 13 and the CDP header
at byte 9. Further, you need to specify that the X & Y coordinates are calculated from the source
and receiver x & y coordinates and that in this case the Receiver X starts at byte 81 and Receiver
Y at byte 85. The Azimuth can be calculated or read directly from the headers. For data which
has not been prestack migrated the azimuth can be calculated from the source and receiver
coordinates. After prestack migration these coordinates lose their meaning. In this case it is
better to read the azimuth directly from the headers. The azimuth can be stored using two
conventions; Map North (azimuth angle is measured clockwise from North) or Cartesian
(azimuth angle is measured counterclockwise from the x-axis (east)). The azimuth for the
Erskine dataset was stored as an integer starting at Byte 185 using the Map North convention.
After specifying all these fields, click Next to move to the next page.
December 2013
10 Guide to AVAz Analysis
Now the following warning message appears because the program is about to scan the entire
SEG-Y file:
Click Yes to begin the scanning process. When the scanning has finished, the Geometry Grid
page appears:
December 2013
Guide to AVAz Analysis 11
After building the geometry files, a new window appears, showing how the well is mapped into
this seismic volume:
Since the mapping is correct, click OK to accept the locations shown on this window.
Now the seismic data appears within the Geoview window (click the Unzoom button and move
the slider down to get the same display);
December 2013
12 Guide to AVAz Analysis
The workspace currently shows CDP gathers from a single inline in this dataset. To see other
parts of the line, slide the scroll bar at the base of the display.
To see the display positioned at the well location, click the Well icon on the right side of the
menu bar above the seismic data as shown below.
The drop-down menu shows the wells in the project. Select the 6-34-38-22 well and the
Geoview window shows the seismic data in the vicinity of that well location.
December 2013
Guide to AVAz Analysis 13
You can also modify other plotting parameters by using the Seismic View Parameters window.
To make that window appear, click the “eyeball” icon and select View 1 Attributes, as shown
below:
The Seismic View Parameters window contains a series of pages which control various aspects
of the plotting. To see the parameters for a specific item, select that item from the list at the left
side. Select Scale Annotation and turn on the Azimuth option as shown below.
By clicking OK, the azimuths will now be annotated as shown in the next figure.
December 2013
14 Guide to AVAz Analysis
Create a stack.
Now that you have read in all the data necessary for the AVAz Analysis, you are ready to
proceed to the ProAZ workflow. First, look at the tabs to the left of the Geoview window. You
will see that one of these tabs is called Workflows. Click that tab to see a list of pre-defined
workflows for various processes, including AVAz Attribute Analysis:
After you click the AVAz Attribute workflow, the workflow expands, itemizing individual
processes that are typically run as part of an AVAz workflow.
These processes can also be accessed under the Processes menu, but the workflow organizes
the relevant processes together in a flow. The first step in this flow is to Select Pre-stack
December 2013
Guide to AVAz Analysis 15
Seismic, double-click on this to bring up the Select Pre-Prestack Seismic dialog where the
input seismic is specified. In this case, specify this to the SEGY-Y gathers read in the previous
step, erskineGath.
After clicking erskineGath and then Select, you are ready to perform the CDP stack.
Normally in this workflow we might select horizons that have previously been read in from some
external source but these are not available in this case. You must pick a horizon after creating the
stack. It is generally a good idea to pick a mute before stacking but again skip this step for the
sake of brevity.
December 2013
16 Guide to AVAz Analysis
There are some features of this dialog which are common to all Process Parameter dialogs.
For example, there is a location to specify the Input and Output files names:
The input file must be one that has already been loaded. In this case, there is only one file
available, erskineGath. The output file can have any name. The program has suggested
cdp_stack.
There is a section to specify the data range to process. By default, it is the entire volume:
By default, only the most critical parameters for this process are specified on this page. To see
the more advanced options, click the button at the base of the dialog:
December 2013
Guide to AVAz Analysis 17
This produces a series of extra tabs, which control in detail the process to be performed. Note
that these will differ from one process to the next.
Click that icon and the Parameter Dialog detaches from the Geoview window to allow it to be
moved aside, making the data more visible. Clicking the “airplane” again re-attaches the dialog.:
If you click the Run Batch button, that will start the process in the background, leaving the main
Geoview program free to carry on with other work. That is often helpful for long, computer-
intensive processes. For now, click Run to start the CDP stack process as usual.
When the process finishes, the Geoview window looks like this: a split window showing both
the input and output volumes:
December 2013
18 Guide to AVAz Analysis
The stacked section is an important quality control tool. At this point you can verify that the
input gathers result in the same stacked section as might have been earlier interpreted. You can
also verify that any imported horizons are consistent with the stacked section. If no horizons
have been imported, you can pick horizons on it. Further, you can overlay the well control and
examine the correlations.
The default split-screen display is very useful for looking at the results, but there are many
modifications possible.
For example, you can increase the available plot space by clicking the “x” on the Project
Manager window, as shown, to temporarily hide that window:
To restore the Project Manager window, click its name to the left:
You can also temporarily hide one of the views. For example, click on the first icon shown
below to temporarily hide View 1, which shows the input data:
December 2013
Guide to AVAz Analysis 19
December 2013
20 Guide to AVAz Analysis
Finally, to see the most complete control of the seismic display, right-click on either of the
seismic windows. A pop-up menu appears:
If you click this item, a dialog appears, allowing complete control of the display:
December 2013
Guide to AVAz Analysis 21
Picking Horizons
Next, pick a horizon for later use. Before picking the horizon let’s go to the 14-13-38-22 well
location by clicking on the well icon and selecting the appropriate well.
To pick horizons, first select Pick Horizons from the Horizon tab at the top of the seismic
display.
This brings up the Picking Horizon dialog. The horizon can either be picked on the prestack
gathers (view 1) or the stacked section (view 2). In this case, pick the stacked section so select
View 2 and click OK.
When picking horizons, you can quality control the picks by displaying them on a map. The next
dialog asks if you want this quality control. Choose Yes.
December 2013
22 Guide to AVAz Analysis
Second, the picking dialog appears below the seismic window. Change the Mode to Left & Right
Repeat and keep the search as Peak. Then click the event just above 1100 ms, as seen below:
This event on the stack is auto-picked with orange plus signs. The yellow dashed line that
appears on the gathers to the left shows the constant picks from the stack but indicates that pre-
stack picking has not been done. On this particular dataset, you need to pick at least two inlines
to avoid cycle skipping in the auto-picker. Use the well icon to select the 6-13-28-22 well to
move to inline 27.
December 2013
Guide to AVAz Analysis 23
Rather than picking the whole volume manually, you can automatically pick the horizon. From
the Options menu, select Automatic Picking.
On the next dialog the defaults are all fine so just click OK. The auto-picked time structure map
is then displayed as shown below.
December 2013
24 Guide to AVAz Analysis
The map may be closed by clicking the x in the upper right corner of the map.
The picking session can then be closed by clicking OK in the Picking dialog.
The azimuthal amplitude analysis is performed on constant time slices. It is important align the
events prior to performing the azimuthal AVO Analysis. Go back to the workflow and double-
click Trim Statics.
Trim Statics is the process which attempts to correct for residual moveout errors and align the
December 2013
Guide to AVAz Analysis 25
events on the gathers. There are a number of reasons why this process is necessary, including
residual NMO due to azimuthal traveltime variations, and non-hyperbolic moveout. Whatever
the cause, these residual moveout errors can cause significant distortions to the calculated AVAz
attributes, so they need to be corrected before proceeding. In the Trim Statics process, a pilot
trace is formed by stacking each CDP gather. Then, each gather trace is correlated with the pilot
trace, using a series of sliding windows. The cross correlations are used to calculate an optimal
time shift for that window. Finally, the shifts for the windows are interpolated to produce a time-
variant stretch of the trace. The result is to align events with the pilot trace.
The dialog which appears shows that the default is to apply the process to the entire volume and
this will create a new output volume called trim_statics:
Specify the input as erskineGath. Rather than working with a single window to calculate the
trim statics, use a series of windows of Size 80 ms, Window Step of 40 ms, with a Maximum
Shift of 10 ms as shown in the following figure:
When you have filled in the menu as shown above, click Run to start the process. The result
December 2013
26 Guide to AVAz Analysis
shows a much improved alignment at the target zone around 900 ms. The following display has
been displayed horizontally.
Having run the time variant trim statics, now create the Common Offset / Azimuth Gathers using
the trim_statics as the input to the process.
In order to view the azimuthal behavior of the data, next create a Common Offset / Azimuth
(COCA) Gather. The COCA gather:
• Improves the S/N ratio in a similar fashion to supergathers. The interpreter defines an M
inline by N crossline running mix, thus effectively increasing the fold.
• Regularizes the data in offset and azimuth by performing partial offset/azimuth stacks.
This makes it easier to see offset and azimuthal variations. A potential issue with this is
that for large bins, there can be smearing between the amplitude versus offset and
azimuth responses. Thus, these are typically only used for quality control purposes on a
couple of inlines and cross-lines through key wells. The actual AVAz analysis is
performed on the input gathers.
• Can be generated for gathers with or without azimuthal traveltime variations.
December 2013
Guide to AVAz Analysis 27
This brings up the Common Offset / Azimuth Gather dialog. The first combo box specifies the
name of the input gathers to the process. Change this to the SEG-Y gathers previously created,
trim_statics. The next line, a text box, specifies the name of the output. This first run primarily
sorts by azimuth, then offset, so call the output CazCoff rather than CoffCaz. The output can be
quite large, with every output gather consisting of N azimuths by M offsets, thus typically only a
subset of the dataset is run. In this case, create Common Offset Azimuth Gathers for a single
inline 27 through the 6-13-28-22 well.
Again to reduce the output size, only consider offsets to 1800 m. The selection of 12 offset bins
results in offset bins of 150 m. This dataset is quite old and poorly sampled in azimuth, so only
generate 4 azimuths starting at 0° of azimuth. In order to try and improve the offset and azimuth
distribution and fold, 5 crosslines and 5 inlines are read in to generate the supergather as shown
in the dialog below.
December 2013
28 Guide to AVAz Analysis
Run the process to generate a display similar to the one shown below. The yellow “banding” is
for display purposes only, and makes the individual azimuths easier to visualize.
December 2013
Guide to AVAz Analysis 29
Note the two wells have been aligned by selecting the Well icon and displaying around the 6-13-
38-22 well.
December 2013
30 Guide to AVAz Analysis
The padded traces show the missing offsets and azimuths, highlighting some of the problems
with this dataset.
An alternative way to organize the data is to sort first by offset, then azimuth. This emphasizes
the azimuthal nature of the data. Rerun the Common Offset / Azimuth gathers, this time
selecting Primarily sort by offset, secondary sort by azimuth. Change the Output name to
COCA to reflect the sort order and then click Run.
Regularizing and improving the fold helps emphasize the azimuthal amplitude and traveltime
variations. The azimuthal amplitude variations are quite subtle at this stage. They will become
more obvious in the interactive analyzer.
December 2013
Guide to AVAz Analysis 31
The two Common Offset/Azimuth Gathers can be displayed one on top of the other. Do this by
going to the Project Data tab, then to the Seismic window and dragging CazCoff to View 1 as
shown below:
December 2013
32 Guide to AVAz Analysis
Now display the range of incident angles as a color display. In order to do this, you need to
specify a P-wave velocity. The easiest way is to specify this is from a well log. In the workflow
double-click Setup Velocity Model.
The dialog which appears specifies the velocity field used to transform from offset to angle:
December 2013
Guide to AVAz Analysis 33
You have three choices for specifying this velocity field: as a table of control points, using a
single log or providing a volume in SEG-Y format. This SEG-Y velocity field can be constructed
in a similar fashion to constructing a background velocity field for a Strata post-stack inversion
project. For this small volume, just use a velocity log from one of the well logs. In this case use
the 6-13-38-22 well and in particular P-wave_1 velocity curve. Click Save to save the velocity
model velfield1 and exit the dialog.
Note that by default a 500 m vertical smoother has been applied to the velocity field. A certain
amount of smoothing is desired since the high frequency variations present in the velocity field
are only correct at the well location and yet this velocity field will be used for the entire seismic
dataset. The amount of smoothing should reflect the dominant frequency of the wavelet. You
can turn off the smoothing option or change the size of the smoother. In this case, leave the
default values.
To show the offset-to-angle mapping, right-click the mouse anywhere in View 2 and select
Incidence Angle.
The COCA gather now shows the incident angle as a color display:
December 2013
34 Guide to AVAz Analysis
We can see from this display that the maximum incident angle at the zone of interest (1100 ms)
is around 40°. That information will be used in a later step.
After flattening the gather and specifying the offset to angle mapping, you are now in a position
to perform the Amplitude versus Azimuth Analysis. This typically involves fitting the data to
some linear approximation of the Zoeppritz equation.
There are two key steps in performing Azimuthal AVO in this workflow:
Azimuthal AVO Analysis: This interactive analysis allows us to analyze one gather (or
supergather) at a time. We can easily observe the amplitude data as a function of
azimuth or offset or both offset and azimuth. We can perform an AVAz inversion on a
single event and interactively see how changes in the parameterization influence the
results. We can test parameters such as the size of the supergather and the range of the
angle of incidence. Further, we can directly observe what the distribution of offsets and
azimuths are at a particular location and see how this might influence the result. Lastly,
we can study the noise by observing the scatter of the data about the model constructed
based on the results of the inversion.
Azimuthal AVO Volume: Once we are satisfied with the selected parameters in the
interactive analysis, we can perform the AVAz inversion on the full dataset and create
volume attributes. These attribute volumes can then be interpreted to identify
anomalies. This is typically done using visualization software so both the magnitude
and azimuth information can be interpreted simultaneously.
In addition to these two key processes, you can run the AVAz inversion with constraints.
Constraints help stabilize the parameter estimates in areas with poor fold or azimuth distributions
such as the edge of surveys. If Density, P-wave and S-wave velocity well log information are
December 2013
Guide to AVAz Analysis 35
available, you can use the statistics from these curves to calculate the constraints. Otherwise,
default values can be used.
The Azimuthal AVO Analysis process is once again initiated from the Workflows menu as
shown below.
The Azimuthal AVO Analysis window then appears. This is a large window so you should
collapse the Project Manager window to create more space. The input for this process can be
the original flattened data or the partially stacked data created by the Common Offset / Azimuth
Gathers process. The partially stacked data has fewer traces, requires less memory and hence is
quicker to load into the AVAz Analyzer. The potential negative of using partially stacked data is
the bin smearing associated with this process. If the bin is large, it can have large amplitude
versus offset variations within the bin. With irregularly sampled data in azimuth, this can appear
as Amplitude versus Azimuth variations. For this reason, in this guide use the trim static
corrected data: trim_statics.
The interactive analysis is performed at a single location. Perform the analysis at the 6-13-38-22
well log location. Note that the analysis location can be moved and updated within the program.
We also specify the size of the spatial rolling window for the input gather. Typical values for this
are 5X5 or 7X7. With today’s modern acquisition and prestack 5D interpolation, it might not be
necessary to use a rolling window. Use the 5 by 5 spatial window used previously in the flow
when we generated the COCA gathers. The default velocity field is the same as the one we
specified earlier to generate an angle of incidence display.
December 2013
36 Guide to AVAz Analysis
Similarly to AVO, the AVAz can be approximated using linear approximations to the anisotropic
Zoeppritz equations such as the near-offset Rüger equation or azimuthal Fourier Coefficients
(FCs). The near-offset Rüger equation considers how the amplitude changes as a function of
offset and azimuth and uses all the data to estimate the attribute volumes. The Azimuthal FCs
method describes how the Amplitude varies as a function of azimuth at one specific angle of
incidence. It tries to reconstruct the AVAz as a sum of cosine functions of different periodicities.
For each cosine function the magnitude and phase are estimated. The magnitude and phase of
the 2nd FC conveys similar information to the anisotropic gradient and the Isotropy plane azimuth
from the near-offset Rüger equation.
The near-offset Rüger equation can be thought of as an extension of the standard ABC AVO
analysis to include azimuthal terms. The inversion of the near-offset approximation solves for
four values at every time sample for every CDP:
The anisotropic gradient B iso and the isotropy plane azimuth φiso are additional information
derived from the azimuthal analysis. The intercept A and the isotropic Gradient B iso are
approximately equivalent to the similarly named attributes coming from the 2-term isotropic
AVO inversion. However, they are biased by the presence of the anisotropy. The anisotropic
gradient B iso and the isotropy plane azimuth φiso are vector quantities conveying magnitude and
angle information and ideally should be interpreted together.
December 2013
Guide to AVAz Analysis 37
In this guide, use the near-offset Rüger equation as the Type of Analysis as shown in the
previous figure. The near-offset Rüger equation was originally derived in terms of the Intercept
and Gradient but it can be transformed to P-impedance and S-impedance reflectivity, similarly to
the derivation of the Fatti equation from the Aki & Richards equation. The AVO
parameterization allows the user to choose between these two types of output. In this exercise
we use the default A,B parameterization. The Number of AVO parameters parameter is related
to the AVO parameterization describing how many isotropic terms to solve for. If there is a
sufficient range of angles of incidence, we can perform a 3-term AVO inversion either solving
for the Density, P-wave and S-wave impedance reflectivity or the A, B, C parameterization. In
this case, the data is too noisy to justify performing a 3-term inversion so we leave the default
value for the Number of AVO parameters to 2.
It is possible to run the AVAz inversion either as Simple Least Squares (SLS) or Iterative
Reweighted Least squares (IRLS). The IRLS option is first solved with SLS from which a
model is created and the misfit is calculated. Data with large misfit errors are then weighted
smaller. New parameters are then estimated using weighted least squares. This is done several
times to accurately identify and suppress large outliers. The default option is to run the inversion
using IRLS. The only time it might be appropriate to run SLS is for model data with no noise.
The inversion can be run with constraints. This will be discussed later. For now leave the
default as No.
The Azimuthal Reciprocity flag allows you to display azimuths from -90° to 90° or -180° to
180°. For P-wave seismic data, it should not matter if the azimuth is measured from the source
to the receiver or vice versa, hence an azimuthal range of 180° should be adequate. The default is
the more limited azimuth range, as implied by reciprocity.
Associated with the Azimuth Fourier Coefficients is a range of angles. The interactive program
uses this information to specify the angles for which to calculate and display AVAz curves. The
interpreter specifies a minimum and maximum value and a step size. The amplitude versus
Azimuth is calculated and displayed for each bin. The curve is modeled using the parameter
estimates from the AVAz inversion.
After you run the process, a display similar to the one below appears. By default the event which
is analyzed is at 1000 ms. In the display below, the red event has been moved to the zone of
interest. You can do this by clicking the cursor on the gather at the desired time. In this case the
event is specified to be 1058 ms. The time of Event 1 can also be specified in the text box on the
left hand side of the program.
December 2013
38 Guide to AVAz Analysis
The left side of the display shows the supergather organized by increasing offset and azimuth.
This is sometimes called a snail plot. The data is organized so it monotonically increases in both
offset and azimuth as shown in the figure to the right. The right side displays a graph where the
data is plotted as a function of Amplitude versus Azimuth. The individual data points are color
coded by the angle of incidence. As described on the previous page, the predicted AVAz is
displayed as colored curves superimposed on top of the data. The color corresponds to the angle
of incidence as indicated by the color bar.
We can update the color data values through the menu of the AVAz Curves tab. Specify the
Minimum and Maximum Value, in this case 0 and 40, and the Step Size of 10. If the window is
too small, a ">>" icon appears which brings up these controls.
After you change the parameters, the AVAz Curves plot changes to the one below;
December 2013
Guide to AVAz Analysis 39
Note that we can resize the area of the AVAz window by dragging the divider to the left.
December 2013
40 Guide to AVAz Analysis
At the top of the window are the parameters estimated as part of the AVAz inversion for this
event.
So far we have concentrated on the AVAz Curves Display. We can display the Amplitude
versus Offset by clicking the AVO Curves tab.
In the AVO Curves display, the amplitude is plotted as a function of angle or offset. Note that
there is a combo box that allows you to switch the horizontal axis between angle of incidence
and offset. The data itself is color coded by its azimuth. The best fit AVO curve is shown in red.
Ideally, we would like to observe a noticeable azimuthal variation about the AVO trend line.
Alternatively, it is possible to view the amplitude as a pseudo time or depth slice where the x-
axis is the azimuth and the y-axis is the angle of incidence. The data is plotted as a series of open
boxes whose color corresponds to the amplitude value as shown in the color bar to the right. The
background color is the modeled data reconstructed from the AVAz inversion using the same
color scale. This view is called the Domain Slice.
December 2013
Guide to AVAz Analysis 41
This display gives valuable information about how well sampled the data is at this location. The
near angle data is only acquired with a very narrow range of azimuths. This is not too
problematic since the Zoeppritz equation predicts little or no azimuthal amplitude variation at the
near offsets. All that is needed is enough near offset data to accurately predict the intercept.
There seems to be a reasonable distribution of azimuths between 20° and 30° but there is poor
sampling beyond this.
We can also access the parameters and rerun the azimuthal AVO analysis, by clicking on the
Parameters Tab as shown below;
December 2013
42 Guide to AVAz Analysis
For example one could increase the number of Inlines and Xlines going into the Spatial Rolling
Window, rerun the analysis and see how that impacts the distribution of offsets and azimuths.
Alternatively, we could exclude the poorly sampled azimuthal data beyond 30 degrees of angle
of incidence and see how this impacts the inversion. Let’s keep the current parameters and now
run the AVAz inversion on the whole dataset.
To start the azimuthal AVO analysis, select the Workflow tab. At this point you might
need to click on the Project Manager tab to restore the workflow space as shown to
the right. Initially, we will run the inversion without constraints so we skip the
Calculate AVAz Constraints step. Double-click the Azimuthal AVO Volume process
as shown below:
December 2013
Guide to AVAz Analysis 43
This brings up the Azimuthal AVO Volume Process dialog. Most of the parameters should be
updated based on choices made in the Azimuthal AVO Analysis process. The parameters and
the layout are the same with a few exceptions. There is no reciprocity flag since the actual
azimuths are always used in the calculation. In addition, the Azimuthal Fourier Coefficients
Calculation Angles parameters are only displayed if you select Azimuthal Fourier
Coefficients as the Type of Analysis.
Change the input to trim_statics and call the output Ruger. For this initial run, let’s only run 1
inline to save time, so specify inline 27, then click Run. When the process is completed, the
calculated attributes appear in a split screen: View 1 shows the input while View 2 shows the
intercept.
December 2013
44 Guide to AVAz Analysis
The program adds the name of the attribute as a suffix to the output. In our case the output is
named Ruger, so the process outputs the Intercept Ruger_A, the Isotropic Gradient
Ruger_Biso, the Anisotropic Gradient Ruger_Bani and the Isotropy Plane Azimuth Ruger_Az.
In order to view the anisotropic gradient, right-click anywhere in View 2, and choose Color Data
Volume > Ruger_Bani as shown below.
The default color range is too large. To modify the range, right-click over the color bar and
select Modify Range.
December 2013
Guide to AVAz Analysis 45
Then modify the Upper Value of the range to 10000 and click OK.
Lastly, we change the horizontal scales so the plot fits nicely in the screen. First click the Eye
icon.
This brings up the attribute selection dialog. In this case, we want to change the plot scales for
View 2 so click Modify Attributes for View 2.
December 2013
46 Guide to AVAz Analysis
This brings up the Seismic View Attributes dialog. Select the Scale option and change the
Time Axis to 7.5 in/s and the Horizontal Scale to 18 traces per inch. Note that, depending on
the size of each individual's screen, these parameters might need to be changed.
After entering all the parameters, click OK. This results in a display similar to the one below.
December 2013
Guide to AVAz Analysis 47
Note that the amplitudes at the edges of the survey are abnormally large, indicating a possible
instability with the inversion. This is due to the poor distribution of offsets and azimuths at the
edge of the 3D. You can mitigate this somewhat be specifying a larger value for the Minimum
Acceptable Fold (on the Basic Parameters tab) or a greater value for the Minimum Azimuth
Range on the Advanced Parameters tab. If the Minimum Fold or Minimum Azimuth Range
Criteria is not met, the inversion is not run. This is the case with the zero traces at the edge of the
survey in the figure above.
You can use constraints to further mitigate these instabilities at the edge.
Calculate Constraints
In order to run a constrained inversion we first need to specify the constraints. If density, P-
wave and S-wave velocity logs are available, you can calculate the constraints from the statistics
of the well logs using the process Calculate Constraints. In our case we do not have all this
information so we will use the default parameters. For those interested, there is an appendix at
the end of the guide describing the theory, calculation and application of the constraints.
We now rerun the azimuthal inversion, this time using constraints. Once again, from the
Workflows select Azimuthal AVO Volume as shown in the figure below:
This brings up the Azimuthal AVO Volume dialog. Most of the parameters should have been
retained from the previous unconstrained run. Change the output to constrRuger and answer Yes
to Use Constraints. Modify the Trace Range to run the Entire Volume.
December 2013
48 Guide to AVAz Analysis
You can modify the constraints on the Advanced Parameter tab (see Appendix). Upon clicking
Run, you should get a display which looks like the one below:
December 2013
Guide to AVAz Analysis 49
In order to display the anisotropic gradient, right click within the View 2 area and select Color
Data Volume > constrRuger_Bani.
This brings up the anisotropic gradient generated using the constraints. In order to compare this
to the unconstrained inversion, go to the Project data tab, select the Seismic tab, select
Ruger_Bani and drag it to View 1. This should result in a display which looks similar to the one
below:
December 2013
50 Guide to AVAz Analysis
Note that the constraints have addressed the instabilities at the edges. There are still anomalies at
980 and 1060 ms at the 6-13-38-22 well location that we previously saw with the Azimuthal
AVO Analyzer.
In order to clearly see the anomalies, you might need to turn off the horizons. First click the Eye
icon:
This brings up the attribute selection dialog. In this case we want to change the plot scales for
View 2 so click Modify Attributes for View 2.
This brings up the Seismic View Attributes dialog. Select Horizons, select No Horizons and
then click Apply. You should now be able to see the anomaly under the horizon. You can turn
the horizon back on by selecting All Horizons in project and clicking OK.
December 2013
Guide to AVAz Analysis 51
December 2013
52 Guide to AVAz Analysis
Azimuthal attributes contain both magnitude and phase (azimuth) information. So far we have
only plotted the anisotropic gradient or magnitude information. Let’s now display the Isotropy
Plane Azimuth. Recall that the Isotropy Plane Azimuth attribute is created with the suffix _Az
so for the output constrRuger, the Isotropy plane azimuth attribute is called constrRuger_Az.
Drag this to View 1 as shown in the figure below. It is displayed with a circular phase color scale
going from -90° to 90°.
As explained in the theory section, there is a 90° ambiguity with the estimate of the isotropy
plane. This ambiguity arises due to the fact we cannot estimate the sign of the anisotropic
gradient. The calculation of the anisotropic gradient involves calculating a quantity similar to
magnitude and then takes the square root of it, assuming the positive root. If the root is actually
negative this introduces a 90° rotation. The Anisotropic Gradient B ani reflectivity series has
both positive and negative values in it. Assuming the anisotropic gradient is positive entering a
fractured layer, it will be a negative when it leaves the fractured layer. The azimuth at the top of
the fractured layer will be correct but at the bottom it will be rotated by 90°. This explains the
cyclicity seen in the above figure.
Note the Isotropy Plane Azimuth only makes sense for anisotropic media. Typically, it is only
shown when the anisotropic gradient exceeds a certain threshold. Thus both the anisotropic
gradient and isotropy plane azimuth really need to be interpreted together. This will be shown in
a future section on visualization.
December 2013
Guide to AVAz Analysis 53
The AVAz attributes are best interpreted in map view so let’s now create horizon slices at the
Horizon 1 level.
Data Slices
To create a data slice of the anisotropic gradient around Horizon 1, return to the workflows and
double-click Create Data Slice.
This brings up the Create Data Slice dialog. Choose as input constrRuger_Bani. You need to
extract the amplitude of the anisotropic gradient about the event Horizon 1. There are a number
of different options for extracting the amplitude. In our case the anomaly is centered on the
horizon so choose using a Window centered on target of 10 ms. Lastly we need to specify the
name of the output. Call it slice_Bani and then click Run.
December 2013
54 Guide to AVAz Analysis
This generates a map and brings up the map window. Unfortunately, the default color scale is
plotting everything in white, so you need to modify the color key range. Right-click over the
color bar and change the Upper Value to 10,000:
Note the anomaly close to the 6-13-38-22 well. We could also create a similar display for the
azimuth. Instead, use the visualization software, View 3D, to display the Anisotropic Gradient
and Isotropy Plane Azimuth at the Horizon 1 surface simultaneously.
December 2013
Guide to AVAz Analysis 55
View 3D
To visualize the magnitude and azimuth directions simultaneously, select the View 3D tab. This
is the visualization software in Hampson Russell. From the Project Data tab, select the
constrRuger_Bani volume and drag it over to the View3D window;
This brings up a data loading dialog for the visualization center. All the default values are okay
so just press OK.
December 2013
56 Guide to AVAz Analysis
Next, select the constrRguer_az volume and drag it over to the View3D tab, again clicking OK
on the next dialog. Now there are two volumes in the viewing window, but only the last is being
displayed in color, as seen at the top.
Next, open the Horizon tab in the Project Manager and select Horizon 1. Drag it over to the
View3D window.
December 2013
Guide to AVAz Analysis 57
Then, select the slice horiz1_Bani from the Slice tab. Drag it over to the View3D window.
Finally, select well 6-13-38-22 from the Well tab. Drag it over to the View3D window.
December 2013
58 Guide to AVAz Analysis
Then, make more space by closing one or both of the side bars:
Now, turn on more viewing planes. This allows you to slide through the volume a slice at a time
in the X, Y and Z directions. You can also switch between viewing ConstrRuger_Az and
constrRuger_Bani using the Color Volume button above the view.
December 2013
Guide to AVAz Analysis 59
However, there is another option for visualization, in which you can view both the
ConstrRuger_Az and constrRuger_Bani simultaneously as a vector-type display. To use this
option, right-click in the black space and select Data Config from the Vector option.
In the data configuration option, select constrRuger_Az in the upper box and constrRuger_Bani
in the lower box.
December 2013
60 Guide to AVAz Analysis
Set the Min. Threshold to 0 and the Max. Threshold to 10000. Make sure the Size Ratio to Min
Threshold is unchecked, then click OK. Note these parameters can be updated later.
This brings up “glyphs” whose direction indicates the azimuth and color and whose size
represents the gradient. In order to see them better zoom in a couple times.
December 2013
Guide to AVAz Analysis 61
At this point, the glyphs are displayed on the time slice defined by the z-value. In order to
display them on the surface associated with Horizon 1, turn off the x, y and z checkboxes so that
only the surface is left. Optimize the surface display by right-clicking in the View 3D window
and selecting Surface>View Parameters as shown in the figure below:
This brings up a dialog to control the parameters associated with the surface. It is possible to
select color bar associated with the surface and the set the minimum and maximum values. In
this case, select the AVO Envelope Color Bar. Set the Minimum Value to 0 and the Maximum
Value to 10000, similar to the seismic displays we performed earlier. The dialog should look
similar to the display below:
December 2013
62 Guide to AVAz Analysis
Having modified the colors, you should now associate the glyphs with the surface. This can be
done by modifying the Vector Parameter Display. In order to do this, right-click in the black
space and select Vector > Vis Parm as shown in the figure below:
December 2013
Guide to AVAz Analysis 63
This once again brings up the Fractures Vis. Parameters dialog. To associate the glyphs with
Horizon 1, select the Horizon Based checkbox. In addition, let’s modify the Minimum and
Maximum Thresholds. In the absence of anisotropy, the Isotropy Plane orientation is not
meaningful. Isotropic media do not have a preferential orientation. The Minimum Threshold
specifies the minimum value of the anisotropic gradient to display a glyph. In this case, specify
2000. Set the Maximum Threshold to a value larger than the maximum you want displayed,
such as 10,000, and they click Apply.
December 2013
64 Guide to AVAz Analysis
Lastly, the color of the glyphs is controlled by the color volume's Color Bar at the top of the
View 3D display. Select the AVO envelope color bar as shown in the following figure. Zoom
in several times to make the image the glyphs larger. The figure can also be rotated about
different axis by dragging the figure. Alternatively, the icons on the left size of the window can
be used to display a number of predefined views or perform rotations.
This surface display shows both the magnitude and orientation of the P-wave amplitude
anisotropy. The size and color of the glyphs convey the magnitude information while the
orientation shows the azimuth of the principal axis of the gradient ellipse. If it is assumed that
the local anisotropy is due to a single set of vertical fractures, then the magnitude can be
interpreted in terms of fracture density and strike.
There is a certain amount of redundancy in the above display. The background color of the
surface, the color and the size of the glyph are all conveying the same information. It is possible
to display different attributes and hence convey higher dimensional data.
December 2013
Guide to AVAz Analysis 65
This is the end of the guide. There is no need to save the project. The project is incrementally
updated as you work on it. To close down the Geoview program, select File > Exit on the upper
left corner as shown below;
In least squares inversion, the optimal solution is determined by minimizing the squared misfit
between the model and the data. Sometimes the problem is ill-conditioned (poor S/N for the
amount of data) or underdetermined (not enough data to find a unique solution) so that the
inversion becomes unstable or unsolvable. In these cases we can use constraints to find an
optimal solution which is consistent with our a priori knowledge. A priori information is our
knowledge of what the solution should look like. For example, the mudrock relationship states
that the S-wave velocity should be linearly track the P-wave velocity. Since our knowledge is
uncertain, it is best to describe this using probability theory and distributions.
Constraints have been implemented for the 3-term AVO and near-offset Rüger equations.
Constraints should help in areas with low fold and poor azimuthal distributions. The constraints
are calculated from well log statistics using the process Calculate AVAz Constraints. The
process Azimuthal AVO Volume then optionally reads these constraints and uses them to run
constrained inversion. The constrained inversion can be run either using Simple Least Squares
(SLS) or Iteratively Reweighted Least Squares (IRLS). IRLS identifies and weights down
outliers to improve the solution.
Constraints describe probabilistic relationships between the different parameters. The P-wave
velocity reflectivity can be related to the S-wave velocity reflectivity through the mudrock
relationship. The mudrock slope, the Vs/Vp ratio and the correlation coefficient parameters are
sufficient to describe a Bivariate Gaussian Probability distribution, which can be used to
constrain the AVO problem.
December 2013
66 Guide to AVAz Analysis
Gardner like relationships can be used to constrain the density to track the P-wave velocity
reflectivity and S-wave velocity reflectivity leading to a multivariate Gaussian distribution.
An equiprobability surface of this looks like an ellipsoid in 3D parameter space.
1 C
−1
P (m | I ) ∝ exp − m T m2 m,
2 η
where C m is the covariance matrix
σ 2 RVp σ R R σ RVp Rden
Vp Vs
C m = σ RVp RVs σ 2 RVs σ RVs Rden ,
σ
RVp Rden σ RVs Rden σ Rden
2
and η is the global scalar. The process Calculate Constraints calculates the covariance matrix
C m from the statistics of the well logs.
December 2013
Guide to AVAz Analysis 67
The optimal solution occurs at the maximum probability of the prior probability (blue) and the
seismic misfit probability (black) as shown in the figure below;
The potential negative of using constraints is that they bias the solution towards the priors.
Ideally the solution should be driven by the data. To get around this problem, the relative
weighting of the misfit and prior functions changes depending on the S/N ratio.
The 3-term AVO constraints are extended to the near offset Rüger equation by assuming the
anisotropic gradient is proportional to the crack density and the crack density is uncorrelated
with the density, fast P-wave and S-wave velocities. Under these assumptions the anisotropic
parameter B ani is uncorrelated with the isotropic parameters. It is then possible to introduce a
scaling factor Lambda that relates the variance of the anisotropic gradient to the variance of the
P-wave impedance reflectivity. Using this scale factor, and the fact that B ani is uncorrelated with
the other parameters the covariance matrix can be extended to four terms.
From the workflow, click Calculate AVAz Constraints to launch the Constraints Plug-in.
This brings up the Calculate AVAz Constraints which calculates constraints for the azimuthal
AVO Volume process. It calculates a covariance matrix from the well control. First, the well
December 2013
68 Guide to AVAz Analysis
that the constraints are going to be calculated at needs to be specified, in this case select 16-13-
38-22. Then the P-wave velocity curve, S-wave velocity curve and a Density curves need to be
specified. Choose P-wave_corr, S-wave_transform and Density_trans. We default the time
window to 0 to 99999 ms at 2 ms intervals. The time interval controls the blocking. Larger block
sizes reduce the variance. Lastly, the name under which the constraint file is going to be stored
and accessed needs to be specified. The default is the name of the well log.
After clicking Apply, the constraint matrix is calculated. The actual calculation involves these
steps:
1) The curves are converted from depth to time using the default time to depth mapping.
2) The reflectivity for each of these curves is calculated.
3) Then the covariance matrix is calculated over the time window specified from the Start
Time to the End Time.
The resulting covariance matrix is then display along with the curves generated in the calculation
as shown below:
December 2013
Guide to AVAz Analysis 69
The program actually calculates the normalized parameter covariance matrix for the scaled
velocity AVO parameterization. The covariance matrix is normalized so that the C 11 term is 1.
This C 11 term is the variance of the P-wave reflectivity and acts as a scaling factor. The standard
deviation of the P-wave reflectivity is shown as a parameter below the covariance matrix and can
be used as a scale factor for the Azimuthal AVO volume attributes under the advanced
parameters. In addition, the average V S /V P ratio calculated from the well logs is displayed.
The program also calculates key rock physics relationships from the reflectivity including the:
• Mudrock slope
• Mudrock slope correlation coefficient
• P-wave Gardner coefficient
• P-wave Gardner correlation coefficient
• S-wave Gardner coefficient
Note it is possible to observe that the S-wave and Density curves for this well log were
constructed using empirical relationships. The mudrock slope and P-wave Gardner coefficient
correspond to published values, and the correlation coefficients are close to 1. This implies that
S-wave velocity is almost perfectly tracking the P-wave velocity and likewise for density.
December 2013
70 Guide to AVAz Analysis
The above parameters are transferred to Azimuthal AVO volume through the output file 6-13-
38-22 and from these the parameter covariance matrix is reconstructed. The advantage of using
these parameters is that even in the case where there is no well control, it is possible to construct
the parameter covariance matrix from reasonable guesses of the parameters. The covariance
matrix generated thus is guaranteed to be invertible. This is a requirement of the inversion.
In order to run the Azimuthal AVO with constraints, from the Workflows select Azimuthal AVO
volume as shown in the figure below:
December 2013
Guide to AVAz Analysis 71
This brings up the Azimuthal AVO Volume dialog. Most of the parameters should have been
retained from the previous unconstrained run. Change the output to constrRuger and answer Yes
to use Constraints?
The constraints are read under the Advanced tab as shown below.
December 2013
72 Guide to AVAz Analysis
This then reads in the parameters generated from the Calculate Constraints Process.
December 2013