Chapter 3: Euler 3D Tutorials 21

Tutorial 2: Processing derivative grids

Following data preparation and gridding, the next step in the deconvolution process is
to compute a set of X, Y, and Z-derivative grids, which are required for both the
Standard and Located Euler Deconvolution methods.

T O C REATE D ERIVATIVE G RIDS :

1. On the Euler3D menu, select Process grids. The E3PREP.GX dialog is displayed.

2. Using the Input grid [Browse] button, select the grid file (eulermag.grd).
3. For the X, Y and Z derivative output grids you can specify new names or accept
the default names (dx.grd, dy.grd, dz.grd). We will accept the default names.
4. If you want to reduce short-wavelength noise by upward-continuing the grid
before calculating derivatives, specify a continuation distance and output grid
name. This is parameter is optional, so we will leave these fields blank. For more
information on Upward continuing, click the [Help] button
5. Specify a grid expansion method (square or rectangular). Since the x and y
dimensions of this grid are approximately equal, we will use the square option.
6. Click the [OK] button. The system generates the derivative grids.

Tutorial 3: Displaying derivative grids

After generating derivative grids, you may want to display them on your screen for
comparison with the original total field grid. You have a choice of displaying grids
either as standard or shaded grid images.

T O D ISPLAY D ERIVATIVE G RIDS :

1. On the Grid menu, click Display Grid|Single grid. The Place a grid on a map
dialog is displayed. :
22 Chapter 3: Euler 3D Tutorials

2. Using the [Browse] button, select the Grid name (dx.grd) to display. The
remaining parameters can be left to the intelligent defaults. To plot the grid to a
new map, click the [New Map] button. The dx.grd is plotted to the map, dx.map.
3. Repeat this process until all grids are displayed (dx.grd, dy.grd and dz.grd).
Your grids should look similar to the ones shown below.

Tutorial 4: Standard Euler deconvolution

To perform the Standard Euler deconvolution, you must specify certain processing
parameters, including:
• Structural Index (SI)
 Chapter 3: Euler 3D Tutorials 23

• Maximum % depth tolerance

• Window size
• Maximum distance to accept
• Flying height or Survey elevation
The Structural Index (SI) is an exponential factor corresponding to the rate at which
the field falls off with distance, for a source of a given geometry. The value of the SI
parameter depends on the type of source body you are looking for and the type of
potential field data you are using (magnetic or gravity).

The following table shows some indices for magnetic data. For a complete
description of indices, please refer to Chapter 6: Performing Euler Deconvolution
Analysis page 55.

Magnetic Structural Index

Sill 1
Dyke 1
Cylinder 2
Pipe 2
Sphere 3

The Maximum % depth tolerance determines which solutions are accepted (i.e.
accepts solutions with error estimate smaller than the specified tolerance). The default
is 15 percent — typically a good starting value for a first pass at analyzing the data. A
smaller tolerance will result in fewer but more reliable solutions.

The Window size determines the area (in grid cells) used to calculate the Euler
solutions. All points in the window are used to solve Euler's equation for a source
position. The Standard Euler Deconvolution works best when the search window is
large enough to include the entire anomaly being analyzed, but not so large that it
contains multiple anomalies. Since a typical survey contains anomalies of various
sizes, it may be necessary to run Euler Deconvolution a few times with different
window sizes.

The Maximum distance to accept parameter specifies the maximum distance offset
from the centre of the search window to the location of the source solution. The
deconvolution result is most accurate when the source is centred within the search
window, so therefore source solutions located far from the centre of the search
window may be rejected. Generally, solutions located outside the boundary of the
search window may be rejected as erroneous.

The last two parameters (Flying height and Survey elevation) determine the manner
in which the results are displayed. The units in which the solution z (depth) values are
given are taken from the grids. For drape airborne surveys, enter the flying height,
24 Chapter 3: Euler 3D Tutorials

and the dialog will return z values as depth below the ground surface. For barometric
airborne surveys, enter the survey elevation, and the dialog will give z values as
elevations (i.e. height above sea level).

TO PERFORM S TANDARD E ULER D ECONVOLUTION :

1. On the Euler3D menu, click the Standard Euler decon menu. The Euler
Deconvolution dialog is displayed.

2. Using the [Browse] buttons, select the Magnetic/Gravity grid as (eulermag.grd),

the X, Y and Z derivative grids as (dx.grd, dy.grd and dz.grd).
3. Specify (type) the new Solution database as (Standard_Euler) and specify a
name for the Solution list as (Solutions). The “Solution list” is a group (line) in
the output database to which the solutions are to be written. Up to 20 different
sets of solutions can be written to a solution database.
4. Examination of our test grid shows some long, narrow anomalies probably caused
by magnetized dykes. Assuming these are possible magnetic dykes, specify a
Structural index of (1.0).
5. The Max. % depth tolerance can be specified; however the default of 15 is fine
for our purposes and will produce a large number of solutions, covering most of
the map.
6. Specify the Window size (>=3) in “grid cells”. In our example, the anomalies of
interest are mostly 3-6 km in width therefore a good estimate for the search
window size is about 6 km. Since the grid cell size is 320 m, this corresponds to
about 20 grid cells, therefore, specify (20) here.
 Chapter 3: Euler 3D Tutorials 25

7. Specify the Max dist. to accept. Our search window is 6400 m wide, so set the
maximum distance parameter to half (i.e. 3200 m).
8. Specify the Survey elevation. This survey was flown at a barometric height of
2500 m, so leave the flying height field blank and enter 2500 for the Survey
elevation.
9. Click the [OK] button. The system creates a new group/line (Solutions) in the
(Standard_Euler.gdb) database of Euler solutions with the following channels:
Channel Description

X_Euler Solution x coordinate

Y_Euler Solution y coordinate
Depth (or Elevation) Solution depth (z coordinate). If a survey
elevation is entered, the channel name
becomes "Elevation"
Backgrnd Solution base level
WndSize Window size. This is converted to ground
units by multiplying the number of cells
in the window size by the grid cell size..
dZ Percentage depth uncertainty
dXY Location uncertainty as a percentage of
solution depth Z
Window_X X coordinate of the center of the search
window
Window_Y Y coordinate of the center of the search
window
X_offset X_Euler – X_Window
Y_offset Y_Euler – Y_Window
MASK A flag indicating whether this solution is
to be plotted or not (1 for plot, * for don't
plot). Initially all Mask values are 1. The
E3WINDOW GX edits the Mask channel
to select only significant solutions for
plotting.

Tutorial 5: Located Euler deconvolution

With the new Located Euler deconvolution method, the Blakely test is used to find
peaks in a grid. Unlike the method used by the Standard Euler, where all grid
locations are tested and only those locations with valid solutions passing the input
criteria are retained, the Located method begins by calculating the analytic signal,
finding peaks in the analytic signal, then using these locations for Euler
deconvolution.
26 Chapter 3: Euler 3D Tutorials

This Located solution method produces far fewer solutions than the Standard Euler
and is the method used for UXO detection (for example, used in UX-Detect.)

Calculate and display analytic signal grid

To perform the Located Euler Deconvolution you must first calculate and display an
analytic signal grid. The analytic signal grid is computed from the existing derivative
grids (see Processing Derivative Grids page 21).

T O C ALCULATE AND D ISPLAY A NLYTIC S IGNAL G RID :

1. On the Euler3D menu, click Located Euler decon|Calculate and display analytic
signal grid. The Calculate analytic signal dialog is displayed.

2. Using the [Browse] button, select the Input X, Y and Z Derivative grids (dx.grd,
dy.grd and dz.grd) and specify the name of the Output analytic signal grid
(accept the default: as.grd). From the Display AS grid in new map dropdown list,
select (Create new map).
3. Click the [OK] button and the Data range to map dialog will be displayed.

4. Click the [Scan data] button to scan the selected database for X and Y coordinate
information and then click the [Next>] button. The Create a new map dialog is
displayed.
 Chapter 3: Euler 3D Tutorials 27

5. Specify a Map name as (as) and click the [Scale] button to calculate the scale.
Note that, you can change the scale to a more appropriate value, i.e. 400000.
6. Click the [Finish] button and the Analytic signal grid will be calculated and
displayed on a new map.

Locate solutions

The next step in the Located Euler method is to “locate” the Euler solutions. This step
includes an analysis of the peaks in the analytic signal grid, and then the peak-picking
algorithm is applied to the grid to locate the peaks and write the locations to the
solutions database, and display the locations as symbols on the current map.

The following channels are created and displayed in the Solutions group/line:
28 Chapter 3: Euler 3D Tutorials

• X_Window, Y_Window - window centre locations

• WndSize - an estimate of the peaks' sizes, calculated directly from the grid
• Grid_value - analytic signal grid value at each location
• Mask - used for selecting solutions for plotting etc.
T O L OCATE S OLUTIONS :

1. On the Euler3D menu, click Located Euler decon|Locate solutions. The Locate
solutions dialog is displayed.

2. Using the Grid to locate solutions [Browse] button, select (as.grd).

3. Specify the name of Output solution database as (Located_Euler). Note that, up

to 20 different sets of solutions can be written to a solution database.
4. Specify a name for the Located solution list as (Solutions). The Solution list is a
group (line) in the output database to which the solutions are to be written.
5. Specify a Channel to save grid values as (Grid_Value) and leave the remaining
parameters as the default values. For detailed information on these or other dialog
parameters, click the dialog [Help] button.
6. Click the [Symbol] button to display the Solution location symbol definition
dialog.
 Chapter 3: Euler 3D Tutorials 29

7. You can use this dialog to specify the solution symbol parameters. Click inside
the Symbol outline colour and Symbol fill colour boxes, to display the Color tool.
Click the [<Back] button to return to the Locate Solutions dialog.
8. Click the [OK] button and the Located Solutions will be added to the Solutions
group/line in the Located_Euler database (*.gdb) and displayed as symbols on
the as map (*.map).

Dynamic
data links

Dynamic linking of data point in the spreadsheet (Located_Euler.gdb), profile

(Grid_value column) and map (as.map) windows.

Located Euler deconvolution

The Standard Euler Deconvolution moves a window of a fixed size over a grid of data
and calculates Euler Deconvolution solutions for each window. There are typically
many solutions, virtually one for every window location, which approaches the
number of cells in the grid.

The Located Euler Deconvolution modifies this procedure by first locating only those
windows which encompass peak-like structures in the data. A peak-finding routine is
first run which locates peaks and estimates a window size using the locations of
adjacent inflection points. These locations and window sizes are then used to define
the windows for Euler Deconvolution, using much the same algorithm as used in the
Standard Euler method. The Located Euler method typically produces far fewer
solutions than the Standard Euler method because only a small subset of the grid cells
will be the centers of "peaks" in the data.

The following channels are created and displayed in the Solutions group/line:
• Depth - solution depth
• Backgrnd - background field
• DZ - estimated error in depth
• DXY - estimated location error in solution
30 Chapter 3: Euler 3D Tutorials

• X_Euler, Y_Euler - actual location of the Euler solution

• X_Offset, Y_Offset - difference between the window center and the Euler
solution.
To perform the Located Euler Deconvolution, you must specify certain processing
parameters including:
• Structural Index (SI)
• Maximum % depth tolerance
• Maximum distance to accept
• Flying height or Survey elevation
For detailed information on these parameters, see page 22 or click the [Help] button.

TO PERFORM L OCATED E ULER D ECONVOLUTION :

1. Make sure that the Located_Euler.gdb is the current database (open and
selected).
2. On the Euler3D menu, click Located Euler decon|Located Euler Decon. The
Located Euler Deconvolution dialog is displayed.

3. Using the [Browse] buttons, select the Magnetic/Gravity grid (eulermag.grd)

and the X, Y and Z derivative grids (dx.grd, dy.grd and dz.grd).
4. In the Located solution list box, specify (type) the name of the new solution
group/list (Solutions)
5. Specify the Structural index (1.0), Max. % depth tolerance (15.0), Max. dist. to
accept (3200), and Survey elevation (2500).
 Chapter 3: Euler 3D Tutorials 31

6. Click the [OK] button to apply the Euler Located Deconvolution and display the
results in the Solutions group/line of the Located_Euler.gdb.

Tutorial 6: Displaying solution statistics

Typically, there will be thousands of solutions in the Standard Euler solutions
database or hundreds of solutions in the Located Euler solution database so it is not
practical to examine them individually. Instead we can calculate channel statistics to
get an idea of how many solutions there are, what their depth range is, what the
typical position accuracy is, and so on.
Note: From this point on, we will be using the Standard_Euler.gdb to demonstrate
the Euler 3D capabilities. However, the Located Euler 3D solutions can also
be used in the following tutorials by replacing the Standard_Euler.gdb with
the Located_Euler.gdb.
T O C ALCULATE THE S TATISTICS OF THE D EPTH (E LEVATION ) C HANNEL :

1. Select (highlight) the Standard_Euler.gdb.

2. Click the left mouse button three times on the Elevation channel header cell. The
entire column will be highlighted.
3. Click the right mouse button and from the popup menu, select Statistics. The Stat
Report dialog is displayed.
The following chart summarizes the significance of each value reported for the
Elevation column.

Reported Value Significance

Num. of items Number of items. This represents the total number of Depth (or
Elevation) solutions in the database.
Minimum Lowest depth solution. Since we computed solutions using the
survey elevation, the value is reported in absolute terms (i.e. -687
means 687 units below 0 elevation). If we computed solutions using
the survey height then this value would be reported in relative terms
(i.e. a solution of -100 would be 100 units below the survey height)
Maximum Lowest depth solution. Since we specified a value for the survey
height, this value is reported in absolute terms (i.e. 1902 means
1902 units above 0 elevation).
Mean Statistical mean of all values.
Standard deviation Spread of data around the mean in depth units.
Arithmetic Sum Arithmetic sum of all depth values.

4. Click the [OK] button to close the Stat Report dialog box or click the [Save
Stats] button to save the report to a file.
32 Chapter 3: Euler 3D Tutorials

Tutorial 7: Creating a map to visualize your results

After you have calculated your Euler solutions, you are ready to start creating maps
to visualize your results.

The conventional way to present your results is to draw a circle centred on each Euler
result with a diameter representative of the depth to the top of the model, the larger
the circle, the greater the depth.

T O C REATE A N EW M AP :

1. Make sure that the Standard_Euler.gdb is the current database (open and
selected).
2. On the Euler3D menu, click New map. The Data range to map dialog is
displayed.

3. Click the [Scan data] button to query the database for the data ranges and report
the Minimum X, Y and Maximum X,Y coordinates. Click the [Next>] button and
the Create a New Map dialog is displayed.

4. Enter a Map name (Standard Euler). From the Map template dropdown list,
select (portrait letter) and click the [Scale] button to calculate the scale based on
the defined data range and the specified map template and then adjust the scale to
a more appropriate value (400,000).
5. Click the [Finish] button. The system opens a new blank map window.
 Chapter 3: Euler 3D Tutorials 33

T O A DD A B ASEMAP :

1. Make sure your blank Standard Euler.map is open and selected in the project.
2. On the Euler3D menu, click Draw base map. The Basemap layout dialog, the first
dialog in the basemap layout wizard, is displayed.

3. Using the Map style dropdown list, select (figure). Note that, the type of map
style that you choose here will reflect the type of dialog boxes to follow.
4. Leave the remaining parameters to the default values and click the [Next>]
button. The Figure style base map dialog is displayed.
34 Chapter 3: Euler 3D Tutorials

5. You can specify different types of display styles for the Reference grid via the
dropdown list. Choose (dotted lines) and click the [Next>] button to continue.
The Figure titles dialog is displayed.
6. Specify a Title (Standard Euler Solutions) and a Sub-Title if you wish.

7. Click the [Finish] button to plot the basemap to the current map (Standard
Euler.map). Your map should look similar to the one shown below.

Tutorial 8: Plotting solution symbols, colour bars and

legends
Oasis montaj provides the ability to create a variety of symbol type plots for
spatially located Earth Science data. This functionality is ideally suited for the
presentation of sparse data sets, such as Euler 3D Deconvolution and geochemical
survey results.
Note: The E3PEAKS GX makes the peak locations X_WINDOW, Y_WINDOW the
current X, Y channels while the E3XYEULER GX makes X_EULER,
Y_EULER the current X, Y channels. Therefore, depending on whether you
want to link between the peak locations and the database (for E3PEAKS GX)
or between the Euler solutions and the database (for E3XYEULER GX) you
should be aware of the current XY channel settings. For more information on
Current X, Y (and Z) channels see the Current X and Y Channels topic in the
online help system.
 Chapter 3: Euler 3D Tutorials 35

The topics discussed in this tutorial include:

• Plotting zone coloured symbols (page 35)
• Adding a horizontal colour bar (page 36)
• Using the colour symbol tool (page 38)
• Plotting classified symbols (page 39)
• Adding legend symbols (page 42)

Plotting zone coloured symbols

After creating your first set of database solutions, you may want to plot these
solutions in order to look at the results.

One method of plotting your solutions is with zone coloured symbols. Please note
that the database MUST have current X and Y channels to generate the coloured
symbols.

T O P LOT Z ONE C OLOURED S YMBOLS :

1. Make sure the Standard Euler.map is open and selected in the project.
2. On the Euler3D menu, click Plot Solution Symbols|Zone Coloured. The Coloured
symbols dialog is displayed.

3. Specify the parameters as shown in the dialog above.

36 Chapter 3: Euler 3D Tutorials

4. In choosing the Proportioned symbol base value we specified (2500) to size the
symbols to the value of the original survey elevation (i.e. depth below the plane of
observation) and for the Proportioned scaling (units/mm) we used (500) this
means 500 data units per millimetre.
5. Click the [OK] button. The Euler solutions are plotted on the current map and the
Colour Symbol Tool is displayed.
6. The Colour Symbol Tool enables you to interactively edit/modify the colour of
your coloured symbols, store your custom colour configurations in specialized
colour palette files (*.ITR, *.ZON, *.TBL and *.LUT) and apply the colour
palette files to any of your coloured symbols. Click the [OK] button to close the
Colour Symbol tool.

When viewed with the original grid (eulermag.grd) you can see the distribution of values
along linear features (i.e. SI value of 1.0).

Adding a horizontal colour bar

To help in interpretation of depth values, you may want to add a horizontal colour
bar.

TO ADD A H ORIZONTAL C OLOUR B AR

1. On the Euler3D menu, select Plot solution symbols and then Horizontal colour
legend bar. The Horizontal Colour Bar dialog is displayed. :
 Chapter 3: Euler 3D Tutorials 37

2. Specify the parameters as shown in the dialog above, and click the [Locate]
button.
3. The system will return you to your map. Click the left mouse button where you
wish the horizontal colour bar to be located. The Horizontal Colour Bar dialog
box is redisplayed with Map location X (mm) and Y (mm) values.
4. Click the [OK] button and the system plots a colour legend bar on the map. Your
map should look similar to the map below.
38 Chapter 3: Euler 3D Tutorials

Using the colour symbol tool

To determine the data range for a given colour you can also use the Colour Symbol
tool.

T O U SE THE C OLOUR S YMBOL T OOL

1. To access the Colour Symbol tool, click the Select a Group button ( ) on the
Map editing toolbar. :
2. Move the cursor to the map window and click the left mouse button to select the
symbol group. Then, click the right mouse button and from the popup menu,
select the Colour Symbol Tool. The Colour Symbol Tool will be displayed.
 Chapter 3: Euler 3D Tutorials 39

3. Move the cursor to a colour and read the corresponding data range (e.g. as shown
below yellow = 462.827 to 508.18.

Plotting classified symbols

If you wish to assign specific symbol colours and sizes to certain depth ranges, you
may do so by plotting Coloured Classified Symbols.

T O P LOT C OLOURED C LASSIFIED S YMBOLS :

1. Create a new map and call it Classified Symbol Plot Map.

2. On the Euler3D menu, select Plot solution symbols|Colour range symbols. The
Classified symbol plot dialog is displayed.
40 Chapter 3: Euler 3D Tutorials

3. Specify the parameters as shown in the dialog box above.

4. Click the [Ranges] button. The Specify Ranges dialog is displayed. Specify the
Maximum values as shown below.

5. Click the [Symbols] button. The Specify Symbols dialog is displayed.

6. Select (circle) for the symbol on all levels and click the [OK] button to return to
the Specify Ranges dialog.
7. Click the [Sizes] button. The Specify Sizes dialog is displayed. Specify the Sizes
as shown below.
 Chapter 3: Euler 3D Tutorials 41

8. Click the [OK] button. The system returns to the Specify Ranges dialog.

9. Click the [Colours] button. The Specify Fill Colours dialog is displayed.

10. Specify the colours for all the levels as shown above and click the [OK] button.
Click the [OK] button again on the Specify Ranges dialog and the Classified
Symbol Plot dialog will again be displayed.
11. Click the [Plot] button. The coloured classified symbols are plotted and displayed
on the current map. Your map should look similar to the following.
42 Chapter 3: Euler 3D Tutorials

In this case, we have defined a set of ranges that emphasize deeper sources (i.e. larger
symbols) at the expense of shallower sources.

Adding legend symbols

When you use classified symbols, you will most likely create a legend showing the
symbol colours and size ranges used on the map. When you select this option, the
system automatically reads the settings you specified when creating the classified
symbols initially.

T O P LOT A C LASSIFIED S YMBOL L EGEND

1. On the Euler3D menu, select Plot solution symbols|Colour range symbol legend.
The Classified Symbol Legend dialog is displayed.
 Chapter 3: Euler 3D Tutorials 43

2. Specify a Title and Subtitle for Legend as shown above and click the [Locate]
button.
3. The system displays the Locate Legend dialog box. Click the [OK] button, the
cursor is now displayed as a crosshair, click the left mouse button on the map
where you want the lower left corner of the legend to be.
4. The system redisplays the Classified Symbol Legend dialog box with the X and Y
positions in map units. Click the [Plot] button and the legend will be plotted on
the map.
5. Your classified symbol plot and legend should look similar to the one below.
44 Chapter 3: Euler 3D Tutorials

Tutorial 9: Windowing results

From the previous example, we can see that the Euler Deconvolution has detected the
dyke structures, but the plot also contains many spurious solutions. The next steps are
to window the solutions (removal of certain solutions if they exceed acceptable limits
of depth, weight, or other parameters) and obtain a more focused plot.

There are many ways to window the data, and in normal practice, a lot of trial and
error is involved in making a useful solution plot. Discussion of these considerations
is beyond the scope of this tutorial — here, we discuss a set of windowing parameters
that works well for this dataset.

The Mask channel is used for selecting a subset of solutions from the solution list.
This channel is initially set “1” to indicate that all solutions are selected. If the value
in the Mask channel is set to dummy (*), then this solution will be ignored in
operations making use of the mask channel.

T O W INDOW R ESULTS :

1. Select (highlight) the database (Standard_Euler.gdb).

2. On the Euler3D menu, select Window results and then select Window solution
list. The Refine solution list dialog is displayed.

3. Using the dropdown lists, select the Solution List as (Solutions), the Channel with
results as (dZ) and Mask channel as (Mask).
4. Specify the Mask results less than as (0) and the and greater than as (7). This
means that only those solutions where the depth uncertainty is less than or equal
to 7% will be plotted.
5. From the Masking mode dropdown list, select (Append mask).
Note: In the Append mode, the mask channel values are set to dummy “*” for all
values outside the specified range. No values are reset from dummy to “1” if
they happen to be inside the range. In the Preset mode, the mask channel
value of every item where there is valid data in the selected channel is first set
 Chapter 3: Euler 3D Tutorials 45

to “1”, and then the windowing is performed. (Solutions where the data
channel value is a dummy will have the mask channel value dummy as well).
6. Click the [OK] button to have the system window the values specified.
7. Repeat the Simple Windowing procedure three more times (i.e. repeat steps 1 to 4
with the parameter settings below).

Channel Mask results less than… …and greater than

dXY 0 15

X_offset -2000 2000

Y_offset -2000 2000

8. The windowing procedure changes the values in the Mask channel of the
database. Mask = 1 means plot this point. Mask = * mean do not plot. For more
information on the windowing results, click the [Help] button.

Tutorial 10: Making a windowed channel

Use the Build solution channel menu option to copy one channel to another against a
Mask channel. If a new channel does not exist, it will be created with the same
definition as the original channel.

T O M AKE A W INDOWED C HANNEL :

1. On the Euler3D menu, click Window results|Build solution channel. The Copy a
channel against mask channel dialog is displayed.

2. Using the dropdown lists, select the Copy From channel as (Elevation)and in the
To box, specify the name of the new channel as (Elev_win) and then select the
Mask Chan as (Mask).
3. Click the [OK] button. The system applies the formula Elev_win = Elevation *
Mask and creates a new windowed elevation channel.

Displaying windowed channel

You can now display the windowed solutions from the new channel.
46 Chapter 3: Euler 3D Tutorials

1. After creating the windowed Elevation channel (Elev_win), create another new
map (as before).
2. Then on the Euler3D menu, select Plot solution symbols and then select Zone
Coloured. The Coloured symbols dialog is displayed. Use this dialog to plot the
windowed results leaving the symbol plotting options as before, but choose
Elev_win (instead of Elevation) for the channel to plot.
3. The system displays the windowed solutions.

Standard Euler Solutions “Windowed” Standard Euler Solutions

Note that many spurious solutions have been eliminated. The remaining solutions are
well focused to the anomalies of interest, particularly the two main dykes.

Tutorial 11: Printing maps for interpretation

Oasis montaj uses your installed Windows system drivers to create output maps.

T O P RINT A M AP :

1. On the File menu, click Print. The Print dialog is displayed.

2. Click [Print] to send the plot to your default printer or plotter. Click the [Help]
button for more information about printing maps.

Final Note

You will note that Euler Deconvolution is not perfectly selective. For instance, in our
example not every solution for Structural Index = 1.0 corresponds to a dyke even
after windowing. In the upper right area of the map, many solution points occur over
 Chapter 3: Euler 3D Tutorials 47

magnetic contacts. A magnetic contact has an SI of 0.5 — close enough to the dyke
that it may be difficult to distinguish.

Extracting a meaningful depth value for a given source body depends on an

appropriate choice of structural index. You can experiment with using different
structural indices and window sizes, to detect and solve for the depth of various types
of source bodies in the survey area.

If you have more detailed questions about the Euler 3D system, please refer to the
following chapters.
48 Chapter 4: Euler Deconvolution Theory and Methodology

Chapter 4: Euler Deconvolution Theory and

Methodology

Standard Euler 3D
The Standard Euler 3D method is based on Euler’s homogeneity equation — an
equation that relates the field (magnetic or gravity) and its gradient components to the
location of the source, with the degree of homogeneity N, which may be interpreted
as a structural index (Thompson, 1982). The structural index is a measure of the rate
of change with distance of a field. For example, in a magnetic field a narrow 2-D
dyke has a structural index of N=1, while a vertical pipe gives N=2. In a gravity field,
a pipe has a structural index of 1, while a sphere has a structural index of 2.

This system uses a least squares method to solve Euler's equation simultaneously for
each grid position within a sub-grid (window). A square window, say 10 by 10, is
moved along each grid row. At each grid point there will be 100 equations (for a 10
by 10 window), from which the four unknowns (location X, Y, Z, and a background
value B) and their uncertainties (standard deviations) are obtained for a specified
structural index.

A solution is recorded if the depth uncertainty of the calculated depth is less than a
specified tolerance and the solution is within a limiting distance of the centre of the
data window. When the process is finished, a database file containing the depth
solutions is obtained as output.

This chapter briefly describes the theory on which Euler Deconvolution is based and
provides additional details about Euler Deconvolution algorithm used in the system.

Understanding Euler deconvolution

Any three-dimensional function f (x , y ,z ) is said to be homogeneous of degree n if the

function obeys the expression.

f (tx ,ty ,tz ) = t n f (x , y , z )

From this, it can be shown that the following (known as Euler's equation) is also
satisfied:

∂f ∂f ∂f
x +y +z = nf
∂x ∂y ∂z

Considering potential field data, Euler's equation can be re-stated as follows:

( x − x 0 ) ∂T + ( y − y 0 ) ∂T + ( z − z 0 ) ∂T = N (B − T )
∂x ∂y ∂z
 Chapter 4: Euler Deconvolution Theory and Methodology 49

where (x 0 , y 0 , z 0 ) is the position of a magnetic source whose total field T is measured

at (x, y, z ) . The total field has a regional value of B. Note that N in this expression is
equivalent to -n in Euler's equation.

It can easily be shown that simple magnetic and gravity models conform to Euler's
equation (Thompson, 1982). The degree of homogeneity, N, can be interpreted as a
structural index (SI), which is a measure of the rate of change with distance of a
potential field.

A magnetic point dipole corresponds to N = 3, while a gravity point mass, a magnetic

pole (theoretical) and a line of magnetic dipoles corresponds to N = 2; whereas a
magnetic dyke and an anomalous pipe mass corresponds to N = 1. Reid et. al. (1990)
have shown that a magnetic contact will yield an index of 0.5 provided that an offset
A is introduced to incorporate anomaly amplitude, strike and dip factors:

∂T ∂T ∂T
A = (x − x0 ) + ( y − y0 ) + (z − z 0 )
∂x ∂y ∂z

Given a set of observed total field data, we can determine an optimum source location
(x0 , y 0 , z 0 ) by solving Euler's equations for a given index N by least-squares
inversion of the data. The inversion process will also yield an uncertainty (standard
deviation) for each of the fitted parameters, and this can be used as a criterion to
accept or reject a solution. This inversion process is often called Euler
Deconvolution.

Understanding the algorithm

In the Standard Euler 3D Deconvolution method, Euler Deconvolution is applied by

selecting a square window of data from grids of the total field and its orthogonal
derivatives, solving for (x 0 , y 0 , z 0 ) and its uncertainties, saving the solution if it
passes certain criteria, and moving the window to the next grid point. Solutions with a
depth error (standard deviation) that exceeds a defined tolerance (typically 15%), and
which are within a limiting distance of the observed window are accepted. The
accepted solutions are saved in a Geosoft database file that can be used to plot the
results.

Note that the standard deviation is a measure only of how well the measured field
conforms to Euler's equations for a given index. This assumes that the index chosen is
correct for the source of the field observed in the sample window, and that only one
source is producing the field observed in the window.

For magnetic data, a significant advantage of Euler's equation is that it is insensitive

to magnetic inclination, declination and remanence since these become a part of the
constant in the anomaly function of a given model.
50 Chapter 4: Euler Deconvolution Theory and Methodology

However, low magnetic latitude problems will still exist for magnetic North-South
geologic features because the physics of this situation dictates a low signal to noise
ratio in the data. Pole reduction and careful North-South leveling noise rejection
filters can be applied to low-latitude data to improve the situation somewhat.

For a more complete discussion of the Euler Deconvolution method described in this
manual, please refer to Reid et. al. (1990)

Located Euler 3D
When using the Located Euler 3D method, unlike the Standard Euler method, where
all grid locations are tested and only those locations with valid solutions passing the
input criteria are retained, the Located method begins by calculating the analytic
signal grid, finding peaks in the grid, then using these peak locations for Euler
deconvolution. This Located solution method produces far fewer solutions than the
Standard Euler and is the method used for UXO detection (for example, used in UX-
Detect.)

Calculate and display analytic signal grid

To perform the Located Euler Deconvolution you must first calculate and display an
analytic signal grid. The analytic signal grid is computed from the existing derivative
grids. The analytic signal is the square root of the sum of the squares of the
derivatives in the x, y, and z directions:

asig = sqrt ( dx*dx + dy*dy + dz*dz )

The analytic signal is useful in locating the edges of magnetic source bodies,
particularly where remanence and/or low magnetic latitude complicates
interpretation.

Locate peaks in analytic signal grid

The second step is to locate peaks in the analytic signal grid. This step includes an
analysis of peaks in the analytic signal grid, application of the Blakely grid peak-
picking algorithm to locate the peaks, then the peak locations are both written to a
solutions database and displayed as symbols on the current (Analytic signal) map.

The Blakely method is used to find peaks in a grid. For each grid cell, the GX
compares its value with the values of its eight (8) nearest grid cells in four directions
(along the row, along the column, and along both diagonals). There are four
sensitivity levels, which may be used to determine whether a grid cell will be selected
as a peak:
• Normal (4) - grid values in all of the nearest grid cells are lower
• More peaks (3) - grid values in any three directions are lower
 Chapter 4: Euler Deconvolution Theory and Methodology 51

• Even more peaks (2) - grid values in any two directions are lower
• All ridge peaks (1) - grid values in one direction are lower

Located Euler deconvolution

The third step in the Located Euler 3D method is to apply the Located Euler
Deconvolution. The Standard Euler 3D moves a window of a fixed size over a grid of
data and calculates Euler Deconvolution solutions for each window. There are
typically many solutions, virtually one for every window location, which approaches
the number of cells in the grid.

The Located Euler 3D modifies this procedure by first locating only those windows
which encompass peak-like structures in the data. A peak-finding routine is first run
which locates peaks and estimates a window size using the locations of adjacent
inflection points. These locations and window sizes are then used to define the
windows for the Located Euler Deconvolution, using much the same algorithm as
used in the Standard Euler method. The Located Euler method typically produces far
fewer solutions than the Standard Euler method because only a small subset of the
grid cells will be the centres of "peaks" in the data.

Theory

The apparent depth to the magnetic source is derived from Euler’s homogeneity
equation (Euler deconvolution). This process relates the magnetic field and its
gradient components to the location of the source of an anomaly, with the degree of
homogeneity expressed as a "structural index". The structural index (SI) is a measure
of the fall-off rate of the field with distance from the source.

Euler’s homogeneity relationship for magnetic data can be written in the form:

( x − x0 ) δ T + ( y − y 0 ) δ T ( z − z 0 ) δ T = N (B − T )
δx δy δz
where:

( x0 , y 0 , z 0 ) is the position of the magnetic source whose total field (T) is detected
at (x, y, z,).
B is the regional magnetic field.

N is the measure of the fall-off rate of the magnetic field and may be
interpreted as the structural index (SI).

The Euler deconvolution process is applied at each solution. The method involves
setting an appropriate SI value and using least-squares inversion to solve the equation
for an optimum xo,yo,zo and B. As well, a square window size must be specified
52 Chapter 4: Euler Deconvolution Theory and Methodology

which consists of the number of cells in the gridded dataset to use in the inversion at
each selected solution location. The window is centred on each of the solution
locations. All points in the window are used to solve Euler’s equation for solution
depth, inversely weighted by distance from the centre of the window. The window
should be large enough to include each solution anomaly of interest in the total field
magnetic grid, but ideally not large enough to include any adjacent anomalies.
 Chapter 5: Preparing and Processing Potential Field Data 53

Chapter 5: Preparing and Processing Potential

Field Data
The Euler Deconvolution process requires four Geosoft format grids as input data:
• A starting grid, which may be the total magnetic field, gravity field,
or the first vertical derivative of these fields.
• The first vertical derivative.
• The first horizontal derivative in the X direction.
• The first horizontal derivative in the Y direction.
You can supply all four grids if they are available, or you can use the Process grids
menu option to calculate the derivative grids from the total magnetic or gravity field
grid. The menu option computes the X and Y derivatives in the space domain using a
simple nine-point convolution filter. The Z derivative filter does not have such a
simple spatial representation, so it is implemented in the frequency domain using a
Fast Fourier Transform (FFT) procedure.

Implications of convolution filtering

Using FFT, methods to calculate derivatives can produce undesirable side effects
such as ringing and edge distortion. Although first vertical derivatives are best
calculated in the frequency domain, horizontal derivatives are better obtained with a
simple 3x3 convolution filter. This avoids side effects due to the grid filling in the
FFT filter process. Examples of the convolution filters are shown below:

X horizontal derivative

0 0 0
-1/2*cell 0 1/2*cell
0 0 0

Y horizontal derivative

0 1/2*cell 0
0 0 0
0 -1/2*cell 0

Implications of FFT filtering

FFT filtering is used to calculate the z-derivative. This involves expanding and filling
the input grid before calculating the Fourier transform. The default is to expand to a
54 Chapter 5: Preparing and Processing Potential Field Data

square grid, but if the input grid is long and narrow, it is more accurate and time-
efficient to expand to a rectangle.

Upward continuation may be necessary if the data contains significant short

wavenumber noise. Upward continuation effectively attenuates noise without
changing the physical significance of the data (as opposed to a low-pass filter, which
does change the significance of the data). Continuation distances up to one cell
dimension are not uncommon, although you should be aware of how much you are
losing in the smoothed data when choosing an appropriate distance.

If the data is upward continued, the original grid must also be upward continued so
that all grids represent data observed at the same elevation. You must specify a new
grid file name for the upward continued grid and this grid must be used during Euler
Deconvolution processing.

The procedure for preparing a grid is quite involved and can take a considerable
amount of time depending on the size of the input grid. You will need at least eight
times the size of input grid in available free disk space to complete this process.
 Chapter 6: Performing Euler Deconvolution Analysis 55

Chapter 6: Performing Euler Deconvolution

Analysis
The object of Euler 3D Deconvolution processing is to produce one or more maps
that display the locations and depths of the sources of potential field anomalies. The
source type and structural index are very important and you should decide what types
of sources are of interest before beginning. For example, in kimberlite exploration
you are looking for pipes, so an index of 2 for magnetic data is most appropriate.

In regional interpretation, you are interested in contacts and faults, so an index

between 0 and 1 should be used (an index of 0.5 is also common). You may require a
set of maps that show the results from different structural indices, all of which will be
used to guide your final interpretation.

Selecting a structural index (SI)

The Euler 3D Deconvolution System is based upon the Euler's homogeneity
relationship, which does not assume any particular geologic model. Therefore, Euler
Deconvolution can be applied in a wider variety of geologic situations than
conventional model-dependent techniques. The degree of homogeneity N (-n in the
Euler's equation) may be interpreted as a structural index — a measure of the rate of
change with distance of the field.

The following table summarizes the structural indices for simple models in a
magnetic field and for a gravity field:

SI magnetic field gravity field

0.0 contact sill/dyke/step

0.5 thick step ribbon
1.0 sill/dyke pipe
2.0 pipe sphere
3.0 sphere

Another straightforward way to determine an appropriate structural index is to

determine how many infinite or reasonably large dimensions are present in a given
model. The model SI is this number subtracted from the maximum SI for a given
field, which is 3 for magnetic data (the magnetic field from a point dipole falls off as
1/r3) and 2 for gravity data (the gravity field from a point source or sphere falls off as
1/r2):
56 Chapter 6: Performing Euler Deconvolution Analysis

Geologic model number of infinite dimensions Magnetic SI Gravity SI

sphere 0 3 2
pipe 1 (z) 2 1
horizontal cylinder 1 (x-y) 2 1
dyke 2 (z and x-y) 1 0
sill 2 (x and y) 1 0
contact 3 (x, y and z) 0 NA

Note that a 0 index implies that the field is a constant regardless of distance from the
source model. In the case of a gravity contact, the field would be infinite. These
situations are physically impossible for real data, and a zero index represents a
physical limit, which can only be approached as the so-called 'infinite' dimensions of
the real source increase. In practice, an index of 0.5 can often be used to obtain
reasonable results when an index of 0 would otherwise be indicated.

However, this presents a problem when we wish to obtain results from a 'contact'
model with gravity data. A solution is to work with the first vertical derivative of the
gravity field in place of the original gravity field. This increases the rate of fall-off by
one power (one SI) and makes the gravity SI equivalent to the magnetic SI for the
same model source. The four component grids provided to the Euler 3D System
∂ g ∂ 2g ∂ 2g ∂ 2g
would then be , 2 , and . These can be calculated by following
∂z ∂z ∂z ∂x ∂z ∂y
the same procedure described in section 4, except that the starting grid will be the
first vertical derivative instead of the original field grid. Calculate the first derivative
grid by going through the process using the original data, then calculate the vertical
and horizontal derivatives of this grid.

The correct SI for a given feature is that which gives the tightest clustering of
solutions. From this, you can think of SI as a focus control, in which the correct SI
produces the sharpest focus of results. This is illustrated in the following diagram
(magnetic field):
 Chapter 6: Performing Euler Deconvolution Analysis 57

Vertical Pipe

SI=3 SI=2
poor good

Dyke

SI=0 SI=1
poor good

The solutions on the right represent the correct SI for a magnetic pipe-like body (top)
and a dyke (bottom). It can be shown that, an index that is too low gives depths that
are too shallow; one that is too high gives estimates that are too deep. But even if the
index is correct, it is clear that depth estimates are more precise for high-index
sources than for low. (Reid et al. 1990).

Understanding the computational procedure

The Standard Euler decon. menu option applies Euler depth Deconvolution as a tool
for potential field interpretation. The system reads in the four data grids (the total
field magnetics or gravity grid and its three orthogonal derivatives) and applies Euler
Deconvolution by using the data extracted from a square window (typically 10 by 10
cells in size). The window is then shifted one grid cell and a new solution is
calculated. This procedure continues until the window has moved across the entire
grid, as illustrated in the following figure:

grid limits

Euler window

>>> window shifts one grid cell for each solution.

58 Chapter 6: Performing Euler Deconvolution Analysis

Solutions that are found to pass the following criteria are saved in the database:
1. The solution depth error is less than a defined tolerance (typically 15% of the
depth).
2. The distance from the centre of the window to the source is less than a limit
specified by the user.

Choosing a window size for computation

The window size should be chosen with the following criteria in mind:
1. It should be large enough to incorporate substantial variation of the field and field
gradient.
2. It should be small enough not to include significant effects from multiple sources.

If the anomalies arising from different sources are so close together that both occupy
any given window, poor fit statistics cause the solution to be rejected. There is,
therefore, reason to keep the window as small as possible.

On the other hand, broad anomalies arising from deep sources are poorly represented
in a small window, and unreliable estimates of depth and position of source are likely.
There is evidence that the maximum depths returned are about twice the window size.

The default windows size of 20 x 20 should be adequate for typical exploration

targets. A smaller size (5 x 5) may be used for sources at or near the ground surface.
For deep sources it is better to increase the grid cell size rather than increasing the
number of cells in the search window- otherwise computation will be slow and the
inversion may be confused by many small sources within the search window.

Understanding solution uncertainties

Euler Deconvolution is a statistical process and the resulting solution positions have
associated uncertainties. The (X,Y) position error is a combination of the separate X
and Y uncertainties. Both the depth uncertainty and location uncertainty are reported
as a percentage of the depth below the sensor, which has an appropriate normalizing
effect.

It is necessary to eliminate solutions with high uncertainties, because the system

produces a solution for all window positions whether any significant source is present
or not. We can only recognize spurious solutions by their high uncertainty.

Low SI values are associated with bodies which give rise to low gradients, so
solutions with low SI values have high uncertainties. The general level of uncertainty
depends on the data quality. Selection criteria must therefore be established by
examination of the database solutions. The choice of selection criteria should be
guided by the two requirements that:
 Chapter 6: Performing Euler Deconvolution Analysis 59

• Position uncertainties should be kept low

• Sufficient solutions should be retained to delineate the structures sought
We have generally found that acceptable results are obtained for real data if selection
criteria are set more or less as follows:

SI Depth uncertainty Horizontal uncertainty

0 20% 40%
0.5 15% 30%
1.0 10% 20%

It is most convenient to run Euler Deconvolution with a large depth error tolerance so
that the result file will contain as many solutions as are reasonably possible. You can
then use the Window results... menu option to further reduce the number of solutions
based on reductions in the depth and location uncertainty tolerance. This often
requires trial and error, with each selection plotted and viewed to determine if the
solutions are acceptable.

Guidelines for eliminating spurious solutions

After examining results, you may want to eliminate spurious solutions. The
Euler3D|Window Results|Window solution list menu enables you to select which
solutions to plot from the database created previously. There are four methods that
you can use (alone or in combination) to eliminate solutions:
• Windowing on depth (or elevation)
• Windowing on X and Y Offset
• Windowing on Depth Uncertainty (dZ)
• Windowing on Location Uncertainty (dXY)

Windowing on depth (or elevation)

If you are not interested in finding sources outside a certain depth range, you can
window based on depth. )

When windowing on depth, you should also keep in mind that the width of the search
window used with the Standard Euler Decon menu option determines the depth range
for which solutions are meaningful. The relationship is complex and depends on the
anomaly model used and the character of the data. Generally solutions at depths
much greater or much smaller than the width of the search window may be
considered as errors, resulting from either multiple sources, a source which does not
fit the model geometry, or else no significant source in the vicinity of the search
window.
60 Chapter 6: Performing Euler Deconvolution Analysis

Windowing on X and Y offset

Similarly to above, solutions lying too far from the centre of the search window are
most likely errors resulting from the failure to find a single, clear anomaly which
matches the model.

Windowing on solution offset is particularly useful as a means of "focusing" the

solutions onto the location of a real source. It may be used to help correct the
scattering of solutions caused when the real source geometry differs from the model
geometry chosen in the Standard Euler Decon menu option.

Windowing on depth uncertainty (dZ)

The Standard Euler Decon menu option enables you to reject solutions with a depth
uncertainty greater than a certain percentage (default 15%). However it is best not to
set this parameter too low, since it is easy to eliminate solutions afterward with the
Window results menu option — but the only way to add solutions is to use the
Standard Euler Decon menu option again, which is more time-consuming. )

In addition, the uncertainty is not necessarily the best criterion for rejecting solutions,
since a geological anomaly may be important even if it does not exactly conform to
the shape of the model anomaly you used.

Windowing on location uncertainty (dXY)

The Standard Euler Decon menu option does not restrict the location uncertainty of
output solutions, so it is generally a good idea to window on dXY with the Window
results menu option to eliminate at least the worst cases for location uncertainty.

However, the maximum dXY limit need not be the same as the maximum dZ limit.
For instance if Euler Deconvolution is being used to locate drill targets, then the
location error will likely be more critical than the depth error.

Windowing on new channels

The user may wish to add new channels to the solution database to help select more
meaningful or relevant solutions. For example, if geological information is available
in digital format, then the user may create a geology channel, and then use the
Standard Euler Decon menu option to select only solutions lying within a certain
formation.
 Chapter 7: Interpreting Your Results 61

Chapter 7: Interpreting Your Results

The main issues to think about when interpreting your results are:
1. Are the geological features of interest represented in the results?
2. For a given feature, is the structural index correct?

Getting acceptable solutions for features of interest can involve some trial and error
while changing SI, the window size, and the uncertainty tolerances. For large data
sets, it is usually best to select a smaller sub-set, perhaps where your knowledge of
the geology is best, and run tests to obtain the optimum deconvolution parameters.
The entire area can then be processed using these 'optimum' settings.

If results from a number of different SI are desired, it will be necessary to remove bad
results from different SI runs and combine the accepted results into a single map. This
is most easily done by repeating the Euler Deconvolution process, then creating new
databases for each structural index and combining them into a single map.