Orient 3.4.

2
Spherical Projection and
Orientation Data Analysis Software
User Manual
Copyright © 1986-2016
Frederick W. Vollmer
Contents
iii
Legal

v
Installation

1
1. Introduction
1.1 Overview 1
1.2 Tutorial 1 – Quick Start 4
1.3 Graphics Output 7

2. Data and Coordinate Systems 8

2.1 Introduction 8
2.2 Data Types 8
2.3 Coordinate Systems 9
2.4 Data Entry 10
2.5 Hidden Data Display 11
2.6 Spreadsheet Integration 11
2.7 Project Files 13
2.8 Tutorial 2 – Plotting Spreadsheet Data 14
2.9 Digitizing Data 16

3. Circular Plots 18
3.1 Introduction 18
3.2 Circular Scatter Plots 19
3.3 Circular Histograms 29
3.4 Circular Frequency Polygons 20
3.5 Circular Mean 21
3.6 Tutorial 3 – Circular Plots 22

4. Spherical Projections 23
4.1 Introduction 23
4.2 Geometry of Spherical Projections 23
4.3 Orthographic Projection 24
4.4 Stereographic Projection 25
4.5 Equal-Area Projection 26
4.6 Density Distortion 27
4.7 Maxima and Eigenvectors 28
4.8 Tutorial 4 – Scatter Plots 30
4.9 Contouring 30
4.10 Tutorial 5 – Contour Plots 34
4.11 History and Terminology 34
4.12 Schmidt Plots 36

i
5. Coordinates and Rotations 38
5.1 Introduction 38
5.2 Geographic Coordinates 38
5.3 Tutorial 6 – Geographic Coordinates 41
5.4 Projection Rotation 42
5.5 Tutorial 7 – Rotation to Maxima 43
5.6 Tutorial 8 – Data Visualization 44
5.7 Data Rotation 45

6. Confidence Cones and Bootstrapping 47

6.1 Introduction 47
6.2 Spherical Distributions 47
6.3 Bootstrapping 50

7. Conical Data and Small Circles 55

7.1 Introduction 55
7.2 Conical Data 55
7.3 Fitting Small Circles 56
7.4 Small Circle Confidence 57
7.5 Tutorial 9 – Data Point Weighting 60

8. Kinematic Analysis 62
8.1 Introduction 62
8.2 Entering Line-Plane Pairs 62
8.3 Orthonormalization 62
8.4 M-Planes 63
8.5 Moment Tensors 65
8.6 Tutorial 10 – Kinematic Analysis 67
8.7 Confidence, Synoptic Plots, and Weighting 68

9. Orientation Plots 71
9.1 Introduction 71
9.2 Triangular Orientation Plot 71

10. Cluster Analysis 74

10.1 Introduction 74
10.2 Axis Cluster Partitioning 74
10.3 Vector Cluster Partitioning 77
10.4 Girdle Cluster Partitioning 78

11. Orientation Maps 80

11.1 Introduction 80
11.2 Latitude, Longitude, UTM Conversion 80
11.3 Tutorial 11 – UTM Conversion 80
11.4 Google and Web Maps 81

ii
11.5 Google Earth 81
11.6 Tutorial 12 – Web and Google Earth 81
11.7 Orientation Fields and Domain Analysis 82
11.8 Tutorial 13 – Domain Analysis 86

Acknowledgements 90

References 90

History 94

iii
Legal
License
Orient software and accompanying documentation are Copyright © 1986-2016 Frederick W. Vollmer.
They come with no warrantees or guarantees of any kind. The software is free and may be downloaded
and used without cost, however the author retains all rights to the source, binary code and
accompanying files. It may not be redistributed or posted online. It is requested that acknowledgment
and citation be given for any usage that leads to publication.
This software and any related documentation are provided as is without warranty of any kind, either
express or implied, including, without limitation, the implied warranties or merchantability, fitness for
a particular purpose, or non-infringement. The entire risk arising out of use or performance of the
software remains with you.

Citation
Orient is the result of countless hours of work over three decades. It is released for free in the hope that
it will be useful for scientific and educational purposes. Commercial institutions should contact the
author with details of the intended use. In return for free use, any significant use of the software in
analyzing data or preparing diagrams must be cited in publications, presentations, reports, or other
works. One or more of the following should be cited as appropriate:
Citation for the modified Kamb contouring method (automatic Kamb contouring on a sphere):
Vollmer, F.W., 1995. C program for automatic contouring of spherical orientation data using a modified
Kamb method: Computers & Geosciences, v. 21, p. 31-49.

Citation for the triangular orientation plot or automated structural domain analysis:
Vollmer, F.W., 1990. An application of eigenvalue methods to structural domain analysis. Geological
Society of America Bulletin, v. 102, p. 786-791.

Citation for use of the Orient software:

Vollmer, F.W., 2015. Orient 3: a new integrated software program for orientation data analysis,
kinematic analysis, spherical projections, and Schmidt plots. Geological Society of America
Abstracts with Programs, v. 47, n. 7, p. 49.

Citation for Orient 3 software:

Vollmer, F.W., 2016. Orient 3: Spherical projection and orientation data analysis software.
www.frederickvollmer.com.

Citation for the Orient 3 User Manual (this document):

Vollmer, F.W., 2016. Orient 3: Spherical projection and orientation data analysis software user manual
www.frederickvollmer.com. 96 p.

iv
An acknowledgement such as, “I thank Frederick W. Vollmer for the use of his Orient software”,
“Frederick W. Vollmer's Orient software was used to prepare figures”, or even “Orient was used to
prepare figures”, is greatly appreciated.

Registration
Please consider registering the software, registration is free. This helps determine usage, and justify the
time spent in it's upkeep. To register, send an email to vollmerf@gmail.com with your user name,
affiliation, and usage. You will not be placed on any mailing list or contacted again, other than my
response with a thank you. For example, send me an email with something like:
User: Dr. Frederick Vollmer
Affiliation: SUNY New Paltz, Geology Department
Usage: Research on joint orientation analysis, Catskill Mountains, NY; fault kinematics in
the Hudson Valley fold and thrust belt. Teaching an undergraduate structural
geology course with approximately 35 students per year.
If you are specific about the type of project, this can help me in developing future releases. If you are
using Orient in a teaching environment, I am interested to know the course and approximate number of
students.

v
Installation
Orient is compiled, tested, and debugged on Macintosh OS X, Windows, and Linux Ubuntu. Macintosh
OS X 10.5 to 10.10, Windows XP, 7 to 10, and Linux distributions should all run without problem.
On Macintosh OS X, double click the disk image file (.dmg), and drag the Orient application on to the
Applications folder icon, or to any other desired location. If you get the App Can’t Be Opened message
when double clicking on the Orient icon, right-click on the icon and choose Open from the popup
menu.
Gatekeeper in OS X 10.7 to 10.10 must be set to allow applications other than from the Mac App Store
to be opened. To do so, open the System Preferences and the Security & Privacy option. Under
General select Allow apps downloaded from: Anywhere.
On Windows, unzip the zip file (.zip) using the Extract All option, and drag the Orient application
(Orient.exe) to any desired location. Do not try to run the Orient application folder from inside the zip
file, this is the most common installation problem.
On Linux unpack the gzip file (.tgz), and copy the Orient application (orient) to any desired location.
The Example Data folder should also be copied for use in the tutorials. After installing a new version,
you may wish to reset the preferences using the Restore Defaults command in the Help menu. This will
clear any options that may have changed and set them to default values. The preferences are stored in
the file Orient3.xml, which is located in the folder Orient in your operating system's application
preferences folder. To deinstall simply delete the Orient application folder, and optionally delete the
preference folder. No other files are installed on your computer. No administrative permissions are
required to install Orient, and it is possible to keep a copy on a thumb drive to run on any computer.

vi
1. Introduction
1.1 Overview
Orient is a fast, professional, easy to use spherical projection and directional data analysis program. In
1986 Orient implemented modified Kamb contouring (automatic contouring on the sphere, Figure 1.1),
triangular orientation plots, orientation fields, and automated structural domain analysis (Vollmer,
1988, 1989, 1990, 1993, 1995). Orient 3 brings a new level of accuracy and speed, with many new
tools, including interactive data analysis, coordinate conversions, digitizing, and file integration with
applications such as Microsoft Excel, LibreOffice, Adobe Illustrator, InkScape, CorelDRAW, and
Google Earth.
Orient is for plotting and analyzing directional data, data that can be described by the orientation of an
axis or vector in space or, equivalently, by a position on a sphere or circle. Examples of data that are
represented by unit vectors (directed) or axes (undirected) include geologic bedding planes, fault
planes, fault slip directions, fold axes, paleomagnetic vectors, glacial striations, current flow directions,
crystallographic axes, earthquake epicenters, cosmic ray arrival directions, comet orbital planes,
positions of galaxies, whale migration paths, and the locations of objects on the Earth. Orient has been
written to apply to a wide variety data types, however many examples come from structural geology,
which requires extensive manipulation and analysis of directional data.

Figure 1.1 Lower hemisphere equal-area Figure 1.2 Circular frequency polygon plot of the
modified Kamb contour plot of ice means of two joint sets from each of 24 counties in
crystallographic axes with contours at 10% central New York, plotted as undirected strikes
density, an example of undirected axial data (data (data from Parker, 1942).
from Kamb, 1959).

Spherical projections (Figure 1.1) are used to display three dimensional directional data by projecting
the surface of a sphere, or hemisphere, onto a plane. Lines and planes in space are considered to pass
through the center of a unit sphere, so lines are represented by two diametrically opposed piercing
points. Planes are represented by the great circle generated by their intersection with the sphere or,
more compactly, by their normal.

1
Spherical projections include equal-area (used for creating Schmidt nets), stereographic (used for
creating Wulff nets or stereonets), and orthographic projections, these can be plotted on either upper or
lower hemispheres. Point distributions are analyzed by contouring and by computing eigenvectors of
undirected data from orientation matrixes, or vector means of directed data. Figure 1.3 is an example of
directed data plotted on both upper and lower hemispheres. Data sets and projections can be rotated
about any axis in space, or to principal axes. For two-dimensional data, such as wind or current
directions, circular plots and circular histograms, including equal-area and frequency polygon
diagrams, can be prepared (Figure 1.2).

Figure 1.3 Schmidt plot of directed data, magnetic remanence directions from Precambrian
volcanics, with modified Kamb contours at 20% density (data from Schmidt and Embleton,
1985, in Fisher, et al., 1987). Projection on left is inverted to display upward directed data.

Data can be input as spherical coordinates, longitude and latitude, azimuth and altitude, declination and
inclination, trend and plunge, strike and dip, or other measurements. Orient does kinematic analysis of
fault data, which is represented by a plane and the direction of slip within that plane, by generating P
and T kinematic axes, tangent line diagrams (Figure 1.4), and beachball plots (Figure 1.5).
Spherical projections represent data directions, but not spacial locations. Orient therefore includes
orientation maps to analyze spacial distributions of orientation data, such as the location of domains of
cylindrical folding in polydeformed regions (Figures 1.6 and 1.7). Orient can plot the distribution of
data globally (Figure 1.8), and integrates with internet maps, like Google Maps, and with Google Earth.
Additional features include statistical confidence cones, bootstrap analysis, plotting of conical data,
small circle fitting with confidence regions, and projection and data rotations. Figure 1.9, for example,
is a projection of poles to bedding in graywacke rotated to display fold a cylindrical axis.

2
Figure 1.4 Lower hemisphere equal-area slip Figure 1.5 Lower hemisphere equal-area
tangent plot of 38 normal faults from Crete, beachball diagram of data in Figure 1.4 showing
Greece. Each data point is defined by both a quadrants of P shortening (red) and T extension
plane and a directed line (data from Angilier, (cyan) kinematic axes.
1979).

Figure 1.6 An axial orientation field of Figure 1.7 Triangular orientation diagram (Point Girdle
eigenfoliation dip lines derived from 625 foliation Random) diagram showing the variations in orientation
planes, Dovrefjell mountains, Norway (data from data symmetry and scatter among structural domains
Vollmer, 1990). defined from the data in Figure 1.6.

3
Figure 1.8 Upper hemisphere equal-area Figure 1.9 Lower hemisphere orthographic
projection of 14,229 earthquake epicenters (1980- projection of 56 poles to bedding in graywacke,
1990), an example of directed data plotted by Albany County, New York rotated to display fold
latitude and longitude (data from NOAA). axis (data from Vollmer, 1981).

1.2 Tutorial 1 – Quick Start

Open the Orient application to display the data entry spreadsheet, the main data display area (Figure
1.9). By default, columns for ID, Station, Strike, Dip, Trend, and Plunge are displayed. The ID is an
integer that should be unique for each measurement, Station is any alphanumeric string to identify the
measurement (to see all available columns, select the menu command View Data Columns, these
additional columns will be covered later). Measurements for planes should be entered in the strike and
dip columns, and lines in the trend and plunge columns (these can be changed, for example, if you
prefer dip direction over strike).
Use the mouse to examine the icons in the toolbar, from left to right these are Open, Save As,
Spherical Projection, Circular Histogram, PGR Plot, Orientation Map, and Preferences. Most of the
controls in Orient have tooltips, or help hints displayed when the mouse is over the control. Click on
the Spherical Projection icon to display the default Schmidt net (Figure 1.10). The toolbar icons in the
spherical projection window are Export Image As, Find, Zoom In, Zoom Out, Zoom Fit, and
Preferences. Export Image As will save the image to a graphics file in various raster (Adobe Photoshop
and GIMP compatible) and vector (Adobe Illustrator and Inkscape compatible).
Begin entering numbers into the strike and dip columns, and the spherical projection will automatically
update to display them. The ID number is automatically incremented, although a different number can
be entered. To identify individual measurements, click on the Find icon in the spherical projection
window. In this mode any selected data point will be highlighted in the other window, a colored bar in
the spreadsheet, and a selection icon in the spherical projection window (Figure 1.11).
To only display data items selected in the spreadsheet, turn on the Plot Selected option in the Graph
menu. In this case only the selected data will be displayed on the spherical projection. Turn this off
before continuing the tutorial.

4
Figure 1.10 The Orient Data window with the spreadsheet for data entry and display, and a Spherical Projection displaying
the default Schmidt net.

Figure 1.11 Display of data in spreadsheet and spherical projection. The data point can be selected in either window, and
will be highlighted in the other window.

Next, open the Preferences dialog by clicking its icon. Select the Spherical Projection option from the
pulldown menu, and the Data Symbols panel. Check the Great Circle checkbox (Figure 1.12), and
press OK. The projection will update to show great circle arcs for each of the data points (Figure 1.13).
A common calculation required for geologic data analysis is to determine the intersection between two
planes, such as bedding and cleavage, or of multiple bedding planes to find a fold axis. To do this in
Orient, select the planes, the result is displayed in the status bar and plotted on the spherical projection.
The result is displayed in the status bar as the calculated maximum and minimum eigenvectors, and
also displayed on the spherical projection (Figure 1.14). The maximum ( Max) is a mean value, and the
minimum (Min) is the intersection (see Section 4.7 for details).

5
Finally, click on the Circular Histogram icon to display a circular histogram, or rose diagram, of the
data (Figure 1.15). By default the data is displayed as undirected data, and planes are displayed by their
strikes.

Figure 1.12 The Orient Preferences dialog showing spherical Figure 1.13 The projection after checking
projection data symbol options. the Great circle option.

Figure 1.14 The projection after selecting all Figure 1.15 Circular histogram of the data
data points, the minimum (blue) and displayed in Figure 1.7, showing the strikes
maximum (red) eigenvectors are displayed. of the data.

1.3 Graphics Output

Graphics output of plots and diagrams is to bitmap (raster) and vector graphics files. Bitmap files are
compatible with photo editing programs such as Adobe Photoshop and GNU Image Manipulation

6
Program (GIMP, free), and can be imported into most word processing programs. Formats include
Portable Networks Graphics (png), Windows Bitmap (bmp), Tagged Image File Format (tiff or tif) and
Joint Photographic Experts Group (jpeg or jpg). The default format is png, is a widely supported
optimally compressed non-lossy format. bmp and tiff files are also non-lossy and widely supported.
The jpg file format was designed mainly for photographic images, and is less suitable for plots and
illustrations.
Vector graphics files allow editing in programs such as Adobe Illustrator, InkScape (free), and
Computer Aided Design (CAD) programs such as AutoDesk AutoCAD and QCAD (free). Orient
supports Scaled Vector Graphics (svg), Encapsulated PostScript (eps), and AutoCAD Drawing
Exchange (dxf) formats.
The default svg format is an open source vector graphics format that is widely supported, including by
Adobe Illustrator, InkScape, CorelDRAW, and most web browsers. Some versions of Adobe Illustrator,
however, contain bugs that prevent opening svg files. For example, on Macintosh, Illustrator 5 can
import svg files, but Illustrator 6 can not. On Windows, Illustrator 5 can not open svg files, but
Illustrator 6 can. As a work around, try opening the svg in another program, such as InkScape or
Macintosh Preview, and saving it in another format, such as pdf. Alternatively, save the plot as an eps
file which Illustrator should be able to open.
The vector graphics file formats vary in their support of drawing features. svg is the most complete,
and Orient saves plot elements as nested groups. svg does not support layers, but grouped elements can
be moved to layers once opened in a vector graphics editing program. eps format is a widely supported
PostScript vector format, but does not support layers, groups, or transparency. The dxf format does
support layers, but not transparency, only simple polygon fills are supported, and line styles, text, and
colors may vary from other formats.
Note that none of the vector file formats supports bitmaps, such as the gradient contour fills in Orient.
To use these in a vector graphics editing program, save the background only, turning off data, label, and
net display, and save as a bitmap file. The resulting background can be imported into a vector graphics
editing program with the vector graphics overlain.

Format Extension Format Comment

Portable Networks Graphics png Bitmap Best choice for bitmap output
Tagged Image File Format tiff Bitmap Good choice
Windows Bitmap bmp Bitmap Good choice
Joint Photographic Experts Group jpg Bitmap Poor choice for bitmap output, lossy
Scaled Vector Graphics svg Vector Best choice for vector output
Encapsulated PostScript eps Vector Good choice, Adobe format
AutoCAD Drawing Exchange dxf Vector Standard for CAD software
Table 1.1 Summary of graphics output formats supported by Orient.

7
2. Data and Coordinate Systems
2.1 Introduction
This chapter defines the data types that Orient can analyze, including specification of the angles and
coordinates used to describe them. There are numerous ways of describing a direction, including
latitude, longitude, trend, inclination, strike, and others. There are also many possible coordinate
reference frames, the basics are described here with more detail for those needing it in Section 4.8.
The main goal of this chapter, however, is to get the user quickly started by explaining how to enter
data. Often it is simplest to enter the data directly into Orient's spreadsheet, however Orient can read
and write files compatible with spreadsheet software such as Microsoft Excel and LibreOffice, and, for
many workers, this is a better option. Finally, Section 2.4 explains how data can be digitized from
scanned images of spherical projections or maps.

2.2 Data Types

The orientations of lines, axes, and planes in two or three dimensional space are a common data type in
the earth sciences, and are referred to as orientation data (Watson, 1966; Howarth, 1999; Borradaile,
2003). However, directional data and directional statistics (Mardia, 1972; Mardia and Jupp, 2000),
more specifically refer to lines, axes, and planes, while orientation statistics deals with the full spacial
orientations of rigid objects in space, including orthonormal frames such as foliation-lineation pairs
(Downs, 1972; Mardia and Jupp, 2000; Davis et al., 2015).
Directional data are either unit vectors, directed data, or unit axes, axial or undirected data. Current
flow directions, for example, are directed, while fold axes are undirected. Plotting, contouring, and
statistical analysis of these data types is different. Geometrically, data represents either lines or planes.
On spherical projections planes and lines are considered to pass though the center of a unit sphere.
Planes are represented by either their great circle, the intersection of the plane with the unit sphere, or
by their normal (often referred to as the plane's pole).
Unit vectors or axes in three dimensions can be specified by their coordinates on the surface of the unit
sphere, or direction cosines. However, it is more common to specify two independent angles, a
horizontal angle (such as strike, trend, or azimuth) and a vertical angle (such as dip, plunge or
inclination). Two dimensional data require a horizontal angle only. Available angular measures and
their definitions are listed in Table 1, and discussed in Section 2.2. Orient has separate columns for
lines and for planes, this allows data that contains both, such as kinematic orientation data (Chapter 5),
to be entered as well. Unused columns can be hidden if desired by opening the Data Columns Dialog
from the View Menu.
Angle units can be specified as either degrees (0 to 360 degrees), gradians (0 to 400 gradians), or
radians (0 to 2π radians). Orient stores all numbers internally as radians, which are transparently
converted to user units. This is a global preference that sets how angles are read from files and user
input, and stored in files. There is no mechanism to determine this from file input, so this should only
be done if all files share the same format, or to convert files from one format to another. Degrees is the
default value, and will not normally need to be changed.

8
2.3 Coordinate Systems
Orient is designed to be used with all types of directional measurements and coordinate systems, and
converts to and from user coordinates for data entry and output. The user does not normally need to be
concerned about the underlying coordinate system, except in the case of rotations or selecting different
geographic coordinates (Section 4.8). Orient uses a standard right-handed spherical coordinate system
defined with default axes X, Y, Z = Right, Top, Up. In geologic usage these normally correspond to
East, North, Up, which conforms to map coordinate systems using Easting, Northing, Elevation.
Section 4.8 explains how the default coordinates can be changed to any other orientation, such as
global geographic (Orient has 13 standard coordinate systems, which can be modified by rotations
about any axis).
Spherical coordinates are specified by two angles, θ (theta) and ϕ (phi). The standard mathematical
definitions are that θ is the longitude, or the counterclockwise angle from X in the XY plane, and that ϕ
is the colatitude, the angle from Z. However, there are numerous ways to specify the same information
with two angles, typically specified by the scientific discipline, such as geology, geography, or
astronomy (e.g., Fisher et al, 1987; Mardia and Jupp, 2000). Planes are represented by their normal, or
pole. Alternatively, coordinates can be specified by three direction cosines in this coordinate system,
which are the coordinates of points on a unit sphere. In Orient the user can specify θ and ϕ for lines and
planes according to their data or discipline (Table 2.1).
Geographic data are generally given using longitude as the horizontal angle, the vertical angle is
commonly latitude or colatitude, azimuth and altitude are used in astronomy. Geologic data, however,
are typically specified using azimuths for horizontal angles, measured clockwise from North (Y), and
dips or plunges for vertical angles, measured down from horizontal (the XY plane). Geologic angles are
typically strike and dip for planes, or trend and plunge for lines. Orient supports all common
conventions (Table 2.1), and converts among them.
To convert between units or to set the default units for data entry, open the Data Orientation Units
Dialog from the Data Menu. This will set the data units entered in the spreadsheet, and will convert any
preexisting values to the new format. For example, a common conversion is from strike and dip, to dip
and dip direction. The data will be saved in the new format. Note that data files contain the format,
opening a saved file will always have the format as saved.
Due to the standard use of geographic locations, typically from GPS (Global Positioning System)
positioning, of data points in geology, Orient requires a distinction between two usages of latitude and
longitude for data input. A geologist typically records data locations expressed as Cartesian
coordinates, such as UTM (Universal Transverse Mercator), or as latitude and longitude pairs, as well
as collecting directional data at that location. Therefore it is necessary to determine whether a latitude,
longitude pair is meant as the location of a data point, or if it is the data point itself. The convention
adopted here is that latitude, longitude (or lat, long) refer to a data point location, and that latitude
sphere, longitude sphere (or lats, longs) refer to the data point to be plotted on a spherical projection.

9
 Column Header Abbreviation Notes
Plane θ Horizontal angle of plane
Strike Clockwise from North (Y), dip is to right along strike
Strike left strikel Clockwise from North (Y), dip is to left along strike
Dip direction dipdir Azimuth of dip line
Plane ϕ Vertical angle of plane
Dip Angle from XY plane down toward -Z
Line θ Horizontal angle of line
Azimuth az Clockwise angle from North (Y)
Declination dec Equal to azimuth
Longitude Sphere longs Counterclockwise angle from X
Trend Equal to azimuth
Line ϕ Vertical angle of line
Altitude alt Equal to latitude
Colatitude colat Angle from Z down toward XY plane
Inclination inc Angle from XY plane down toward -Z
Latitude Sphere lats Angle from XY plane up toward Z
Nadir Angle from -Z up toward XY plane
Plunge Equal to inclination
Zenith Equal to colatitude
Table 2.1 Column headers used to specify data formats in Orient. The header or its abbreviation is used in
data files to identify a column of data. A plane requires two header, such as strike and dip. A line also
requires two header, such as trend and plunge.

There are, unfortunately, more than one convention for strike and dip, including two contradictory ones
both called the right-hand rule (e.g., Ragan, 2009). A distinction must be made, so by default, Orient
uses the convention that the dip is to the right looking along the strike (e.g., Pollard and Fletcher, 2005;
Twiss and Moores, 2007). A second convention, where the dip is to the left, the thumb of the right hand
points down the dip (e.g., Barnes, 1995), is referred to in Orient as strike left (or strikel). This
convention can be selected using the Data Orientation Units command. A third convention, using a dip
octant (N, NE, E, SE, etc.) is automatically converted to one of the above as described in Section 2.4.
Finally, dip and dip direction are another common way of giving the orientation of a plane, which is
also supported.

2.4 Data Entry

Each data point must include a pair of angles specifying its orientation in space. The first angle is
measured in a horizontal plane (θ), and the second in a vertical plane (ϕ). For typical geologic data,
these are strike and dip for planes, and trend and plunge for lines. However, all common units are
supported (Table 2.1). Two dimensional data require horizontal angles only.
Before entering data, select the desired data format using the Data Orientation Units command. You
may also wish to hide or show appropriate columns using the View Data Columns command. Separate
columns are used for planes and for lines, so make sure the required columns are visible. The Type
column can contain any alphanumeric identifier, for example S0 and S1 are often used in geology to
designate bedding and cleavage. This is optional, but is required if multiple data types are entered in a

10
single file. All settings, such as symbol sizes and color, are saved for each type. Additional data
attributes include station identifiers, location coordinates, domains, and comments (Table 2.2).
Orient does several automatic data conversions. If data is entered as a bearing with compass quadrants
it is automatically converted to a numerical azimuth. Bearings are given as degrees east or west of
north or south, for example, N30W will be converted to 330. A conversion is also done for planes in
strike and dip, or strike left and dip, formats if a dip octant (N, NE, E, SE, S, SW, W, or NW) is given.
The strike (or strike left) will be corrected if necessary. For example a strike, dip pair entered as 10,
30W will be converted to 190, 30.
Normally only one plane or line is entered on a row, however kinematic analysis requires the entry of
plane-line pairs, which introduces some complexity. Data entry for this type of data is covered in
Section 5.2.
Each data point may be individually weighted by entering a value into the Weight column. The default
weight is 1, enter 0 to discount the point, or a positive value to increase its weight, This can be used, for
example, to weight count joint or fault data by length or surface area. See Section 7.4 for an example of
weighting applied to fitting a small circle. Weighting applies to all calculations except bootstrap
confidence cones, which require the resampling of equally weighted points.

2.5 Hidden Data Display

Directed data (vectors) intersect the sphere at only one point, which may occur in either hemisphere.
Consequently, some of the data point symbols may be hidden from view. One option is to plot both
hemispheres separately (e.g., Figure 1.3), however it may also be useful to plot the data points on the
hidden hemisphere on the same plot using a different symbol. In the Preferences dialog Spherical
Projection pane, the Directed option can be selected to plot hidden contours, confidence cones, and
symbols at 10% to 90% opacity, so they appear semitransparent (e.g., Figure 1.9). Alternatively,
symbols on the hidden hemisphere may be plotted as Unfilled.
Undirected data (axes) intersect the sphere at two diametrically opposed points, by convention usually
only one is plotted. For example, the one in the lower hemisphere is normally plotted in structural
geology, and the one in the upper hemisphere in mineralogy. However, it is sometimes useful to plot
both symbols on the same plot, which can be dome using the Undirected option, which similarly allows
plotting of hidden contours, confidence cones, and symbols.
When learning the properties of spherical projections, it is instructive to plot one or more data points
with both the Symbol and Ray options on, and Directed and Undirected set to 50% opacity. With a
spherical projection displayed, use the Rotate Projection command discussed in Section 5.4 to
sequentially rotate the projection until the points change hemispheres. Try this for both undirected and
directed data, by using the Preferences dialog Symbols Pane Directed checkbox. See Section 5.6 for a
Tutorial on this method.

2.6 Spreadsheet Integration

Data entered into the Orient spreadsheet can be saved in several spreadsheet formats: tab separated
values (tsv), comma separated values (csv), OpenDocument spreadsheet (ods), Microsoft Excel (xls),
and Excel Open Office XLM (xlsx). These are all compatible with most spreadsheet programs,
including Microsoft Excel and LibreOffice (Table 2.2).

11
Orient can also read all of these formats, allowing users to enter data files into Microsoft Excel,
LibreOffice, or other spreadsheet software. The only requirements are that the data have a header row,
consisting of headers listed in Tables 2.1 and 2.3, and that any initial comment lines start with two
slashes (//). Table 2.4 gives a simple example, and Table 2.5 gives a more complete example. The
simplest possible file would be a list of horizontal angles, such as trends or azimuths, for two
dimensional analysis. Three dimensional analysis also requires vertical angles, such as dips or plunges.
The included folder Example Data has numerous examples of compatible files. Open them in Excel or
LibreOffice to examine them, they are used in the following tutorials.
Format Extension Comment
Excel XML Spreadsheet xlsx Microsoft Excel 2007 and later
Excel Spreadsheet xls Microsoft Excel pre-2007
OpenDocument XML Spreadsheet ods LibreOffice/OpenOffice, free open source software
Tab Separated Value tsv Simple text format with values separated by tabs
Comma Separated Value csv Simple text format with values separated by commas
Table 2.2 Summary of spreadsheet formats supported by Orient for opening and saving.

12
Column Header Abbreviation Notes
ID N Integer identification number for data point
Station Alphanumeric station identifier
Zone Alphanumeric zone, such as UTM grid zone
Easting East, X X or easting coordinate of data location
Northing North, Y Y or northing coordinate of data location
Elevation Z Elevation of data location
Latitude lat Latitude of data location (D°, D°M', or D°M'S”)
Longitude long Longitude of data location (D°, D°M', or D°M'S”)
Plane θ Horizontal angle of plane from Table 2.1
Plane ϕ Vertical angle of plane from Table 2.1
Line θ Horizontal angle of line from Table 2.1
Line ϕ Vertical angle of line from Table 2.1
Rake Rake of line in plane
Sense Kinematic movement sense, Normal or Reverse
Error Angular error of line in plane
Alpha Half apex angle of cone
Weight Data point weight, default is 1
Type Alphanumeric data type, as S0 or S1
Label Alphanumeric label
Domain Integer identification number for domain
Comment Alphanumeric comment
Table 2.3 Column headers used to specify data attributes in Orient. The header or its abbreviation is used
in data files to identify a column of data. See Table 2.2 for data angles.

Strike Dip
230 24
018 54
141 15
Table 2.4 Example of a simple data file as displayed in a spreadsheet, such as Microsoft Excel or
LibreOffice, with strikes and dips of planes. See Table 2.1 and 2.2 for other possible column headers.

13
 // Example of an Orient 3 data file in a spreadsheet, this is a file comment, 2015-05-04
Easting Northing Strike Dip Trend Plunge Type Comment
8452.05 12885.05 230 24 S0 approximate
8456.03 12825.03 018 54 S0 overturned
8432.12 12922.30 141 15 S1 spaced
8466.08 12865.40 138 02 L S0/S1
8492.01 12872.00 140 09 L mineral
Table 2.5 Example of a more complex data file as displayed in a spreadsheet, such as Microsoft Excel or
LibreOffice, with an initial file comment, data locations, multiple data types, and data comments. See
Tables 2.1 and 2.3 for other possible column headers.

While a number of data files formats are supported (Table 2.2), tab separated value (tsv) files are
recommended for long term storage and archiving data. The format is a simple text file that can be read
by virtually any text editor, word processor, or spreadsheet program. While the comma separated value
(csv) format is also simple, the fact that text often contains commas requires additional processing.

2.7 Project Files

Orient stores all system settings in an XML project file (extension oprx) that is by default stored in the
Application Support folder for your system (e.g., on a Macintosh, it is in Users/yourname/Library
Support/Orient). Most of the settings, for example the symbol styles and colors, are stored by Type, and
will remain in effect for any files that use that data type.
It is also possible, however, to save the project file in another directory. The project file contains links
to all the files in the project, as well as all of the settings. The data is not stored in the project file, and
remains untouched in your spreadsheet files. When reopening the project file, if the data files are
moved or renamed, Orient will report them missing. If so, simply reopen them. When a project is open,
the project name is prefaced to the file name in the caption of the spreadsheet data window.
The oprx project file is an XML (Extensible Markup Language) Unicode UTF-8 encoded file that
conforms to the W3C XML 1.0 specification, a free open standard format widely used for data files
(e.g., Microsoft Office, LibreOffice). It consists of ordered key-value pairs stored in recursively nested
dictionaries. XML files are designed to be both human-readable and machine-readable, and can be
viewed in text editors and internet browsers.

2.8 Tutorial 2 – Plotting Spreadsheet Data

Open a spreadsheet program, such as LibreOffice or Microsoft Excel, and enter the data shown in Table
2.5. Then save the file using a standard format (see Table 2.2).

14
 // My data file
Strike Dip Trend Plunge Type
060 65 S0
10 45 S0
40 60 S0
310 60 S0
280 70 S0
90 40 FH
330 55 S1
Table 2.5 Example of a data file. Enter this into a spreadsheet such as LibreOffice or Microsoft Excel.

Open the file in Orient using the File Open command. Click on the Spherical Projection icon, to get a
plot as in Figure 2.1. If you have previously modified settings, however, they will still be in effect. The
Help Restore Defaults command can be used to reset all settings to default.

Figure 2.1 Lower hemisphere equal-area projection Figure 2.2 Equal-area scatter plot as in Figure 2.1
of data from Table 2.5. after removing the Schmidt net, and assigning colors
to the data symbols.

To prepare the plot for presentation, it is good to simplify it and focus on the data, so change some of
the settings as follows. From the spherical projection window, press the Preferences icon. Under
Spherical Projection in the Net pane, turn off both Net and Axes. Click on Preview to see the result. In
the Labels pane, set the Increment to -90, Offset to 14, and Size to 12.
Next, assign different colors to the data types. In the Symbols pane select data type S0 (bedding), click
on the data symbol icon, and assign the Fill Color blue. Select data type S1 (cleavage) and assign its fill
color to green, and FH (a fold hinge) yellow. Press Preview, and the plot should be as shown in Figure
2.2.

15
Now add great circles to represent the planes. In the Symbols pane select data type S0 again and check
the Great Circle checkbox. Then select data type S1, check the Great Circle checkbox, and set the
Stroke Color to green (Figure 2.3). A plot of planes represented by their great circles is a β (beta)
diagram. With a lot of data this type of diagram becomes crowded, and loses any statistical
significance, but for a small number of points it can help with visualization. A plot with planes
represented by their poles is a π (pi) diagram, or S-pole diagram, and is preferred over a β diagram as a
better statistical representation.
Finally, add the S0 maxima. In the Maxima pane select the S0 data type, and check Visible. For this plot
just add the Minimum Eigenvector, which gives us a best-fit fold axis. Uncheck the Symbol for both the
Maximum and Intermediate Eigenvectors. For the Minimum Eigenvector check the Symbol and Great
Circle checkboxes, and set their colors to red (Figure 2.4).

Figure 2.3 Equal-area plot as in Figure 2.2 with Figure 2.4 Equal-area plot as in Figure 2.3 with
planes represented as great circles. best-fit great circle and pole given by the minimum
eigenvector of S0 (bedding).

2.9 Digitizing Data

When digital data files or text listings of directional data are not available, data can be digitized from
scanned images of spherical projections or maps. To digitize a spherical projection from a suitable
source image, insure that the image is undistorted, that is circular, and that the north point is unrotated.
An image editing program such as Adobe Photoshop or GIMP can remove distortion and rotate the
image if necessary.
Open the image in Orient using the File Open Image command, which will display the Digitizing
window (Figure 2.5). Click on the Digitize icon to display the Digitize dialog (Figure 2.6). Select the
correct options for the spherical projection, the projection options are stereographic, equal-area, or
orthographic, in either upper- or lower-hemisphere (one hopes that the original author correctly
specified these). Select the data type, line or plane, and press OK.
A prompt will ask for three points on the perimeter of the projection, after entering these, a circle will
be drawn over the projection. If the circle is a reasonable fit, proceed to digitize the points, which will
appear in the Data window spreadsheet. If not, the circle can be reinitialized, or distortion of the image

16
may need to be removed. Spurious data points can be deleted in the spreadsheet as necessary. If
digitizing needs to be interrupted, save the file and reopen the file and image later to resume. To change
data type, lines or planes, reopen the Digitize dialog, select the desired one, press enter, and continue
digitizing.
Digitizing directional data from a map is similar. Make sure the image is undistorted, and that the X
and Y directions are orthogonal with north to the top. The X and Y coordinate scales do not need to be
equal, but they must each be scaled linearly within the map area. If necessary, remove any distortion
using an image editor. Select Map as the source, the correct element, line or plane, and press OK. There
will be a prompt for two points with known coordinates to define the area and coordinates. If the result
is satisfactory, begin digitizing the elements by clicking on a start point and then an end point. The
coordinates entered into the spreadsheet will be the center point between the two points.

Figure 2.6 The Digitize Dialog showing options for digitizing

orientation data from images.

Figure 2.5 An image of a spherical

projection opened in the Digitize Window.

17
3. Circular Plots
3.1 Introduction
Circular plots for two dimensional directional data include scatter plots, circular histograms, and
circular frequency polygons (Davis 1985; Fisher et al. 1987; Cheeney, 1983; Howarth, 1999; Mardia
and Jupp, 2000). Circular plots can also be used to display the horizontal angles of lines and planes,
such as lineation trends. For planes it is possible to plot the strike direction, dip direction, or the
azimuth of the plane normal. The data may be directed or undirected. Undirected data plots two points
at 180°, or can be plotted on a double angle, modulo 180°, plot. The settings for these plots are in the
Preferences dialog using the Circular Histogram selection.

3.2 Circular Scatter Plots

A simple circular scatter plot shows the data distribution on a circular plot, normally the perimeter of a
unit circle (Cheeney, 1983; Mardia and Jupp, 2000). Directional rays may be drawn from the circle
center, or symbols plotted on the perimeter. The vector mean, for directed or undirected data (Section
3.4), can also be displayed. There are many variations on scatter plots, Orient implements ray plots
with lines drawn from the center, and with symbols on the perimeter (Figures 3.1 and 3.2).
Figure 3.1 is a circular scatter plot of the travel directions of 76 turtles after laying eggs (data from
Gould cited in Mardia and Jupp, 2000), an example of directed circular directional data. In order to
better visualize overlapping data points, the opacity of the data point symbols is set to 50%. Figure 3.2
is a plot of the means of two joint sets from each of 24 counties in central New York (data from Parker,
1942). This data is plotted as strike azimuths, an example of undirected data.

Figure 3.1 Circular scatter plot of the orientations of Figure 3.2 Circular scatter plot of the means of two
76 turtles after laying eggs (Gould's data from joint sets from each of 24 counties in central New
Mardia and Jupp, 2000). York (data from Parker, 1942).

18
3.3 Circular Histograms
Two dimensional directional data is commonly displayed as a circular frequency histogram, where the
data count is tallied for bins or sectors of a set angular width. A commonly used graph is a rose
diagram constructed with sector radii proportional to class frequency, an equidistance rose diagram.
Figure 3.3 shows an example for directed data. Unfortunately, such a diagram is biased, and not a true
histogram, because the area displayed for a single count increases with the radius. An unbiased plot is
an equal-area circular histogram where each count has an equal area, and the sector area is proportional
to class frequency (Cheeney, 1983; Mardia and Jupp, 2000; Figure 3.4). Figures 3.5 and 3.6 are
examples for undirected data.

Figure 3.4 Equal-area circular histogram, or rose

Figure 3.3 Equidistance rose diagram of the diagram, of the directed turtle data shown in Figure
directed turtle data shown in Figure 3.1. The 3.1. Each count has an equal area, removing area
increasing area for larger bin counts results in an bias.
area bias, so this is not a true histogram.

Each of these plots is drawn with 24 bins, or 15° sectors. The selection of bin size will change the
appearance of the diagram, an example is shown for circular frequency polygons in Section 3.3.

19
Figure 3.5 Equidistance rose diagram of the Figure 3.6 Equal-area circular histogram of the
undirected joint data shown in Figure 3.2. The undirected joint data shown in Figure 3.2. Each
increasing area for larger bin counts results in an count has an equal area, removing area bias.
area bias, so this is not a true histogram.

3.4 Circular Frequency Polygons

A circular frequency polygon (Haughton, 1864, in Howarth, 1999), or kite diagram (Davis, 1986; Swan
and Sandilands, 1995), is an alternative graph for displaying the circular directional data. In a circular
frequency polygon diagram the bin sector centers are connected by straight lines to form a polygon.
Figure 3.5 is a circular frequency polygon diagram of the directed turtle data (Figure 3.1) using 24 15°
bin sectors, the same size as in Figures 3.3 to 3.6.

Figure 3. 7 Circular frequency polygon plot, or

kite diagram, of the directed turtle data shown in
Figure 3.1, using 24 15°

20
Figure 3.8 is an example of an undirected circular frequency polygon using the joint data (Figure 3.2),
also using 24 15° bins. To illustrate the effect of bin size on circular histograms, Figure 3.8 is the same
data plotted using 12 30° bins.

Figure 3.8 Circular frequency polygon plot, or kite Figure 3.9 Circular frequency polygon diagram of
diagram, of the undirected joint data shown in the joint data as in Figure 3.8, but using 12 30°
Figure 3.2, using 24 15° sector bins. sector bins.

3.5 Circular Mean

A simple measure of location, or best-fit, to circular directional data is the mean direction, which is
calculated as a vector sum. For directed data the two sums are calculated:

Then the mean resultant length and direction, or center of mass of the coordinates, is calculated as:

Since the mean resultant length approaches 1 as directions converge, it is common to cite the sample
circular variance:

21
which is 0 when all the directions are identical. For undirected data the same calculation is done,
however θ is doubled prior to the summation, and the result is halved. For details of statistical measures
and tests for circular directional data see Mardia (1972), Cheeney (1983), Davis (1985), Fisher et al.
(1987), and Mardia and Jupp (2000).

3.6 Tutorial 3 – Circular Plots

Open the file Gould from the Example Data folder (any of the csv, tsv. ods, or xlsx versions) in Orient,
and click on the Circular Histogram icon. In the Preferences dialog check Directed in the Circular
Histogram Symbols pane. If no settings have been previously modified the projection will look as in
Figure 3.4. The Help Restore Defaults command can be used to reset the preferences if desired.
This is an example of directed circular data. The default display is an equal-area circular histogram, or
rose diagram. In the Symbols pane this data type is displayed as Default, as none was specified in the
file. Next, in the Histogram pane, select the Equidistance plot (Figure 3.3), and then the Frequency
Polygon (Figure 3.7) to compare plot types. Note how the equal distance plot unequally scales the
areas, giving more significance to each additional data point in a bin. This is why it is generally
preferable to use an equal area or frequency polygon (kite) plot.
Next, open the file Parker 1942. This is an example of undirected circular data. They are treated as
circular (instead of spherical) data because, although joints are planar, only the strikes were reported.
Note that, because these are planes recorded by strike, the direction plotted should be specified as
Strike, otherwise the plane's normal would be plotted. Reset the Histogram to Equal-Area.
There are two data types in this file, J1 and J2. In the Symbols pane, select J1 and set the Histogram Fill
Color to red, and uncheck Symbol for both data types, J1 and J2, and uncheck Directed. The resulting
plot should be as in Figure 3.6. Next, in the Histogram pane, select the Equidistant plot (Figure 3.5),
and then the Frequency Polygon (Figure 3.8) to compare plot types.

22
4. Spherical Projections
4.1 Introduction
A primary function of Orient is the creation and manipulation of spherical projections of directional
data, in particular azimuthal spherical projections that project the surface of a sphere onto a plane. This
chapter discusses mathematical concepts related to spherical projections, in particular the geometry of
several common projections, and the spherical nets which are commonly used to display and work with
these projections. A final section on nomenclature discuses terminology and common errors that occur
in the literature.

4.2 Geometry of Spherical Projections

A spherical projection is a mathematical transformation that maps points on the surface of a sphere to
points on another surface, commonly a plane. Astronomers, cartographers, geologists, and others have
devised numerous such projections over thousands of years, however two, the stereographic projection
and the equal-area projection, are particularly useful for displaying the angular relationships among
lines and planes in three-dimensional space. A third projection, the orthographic projection, is less
commonly used, but is important for some applications, and is easily visualized. These are azimuthal
spherical projections, projections of a sphere onto a plane that preserve the directions (azimuths) of
lines passing through the center of the projection. This is an important characteristic as azimuths, or
horizontal angles from north (strike, trend, etc.), are standard measurements in structural geology,
geophysics, and other scientific disciplines.
The directions of lines and planes in space are fundamental measurements in structural geology. Since
planes can be uniquely defined by the orientation of the plane's pole, or normal, it is sufficient to
describe the direction of a line. If only the direction of a line, and not it's position, is being considered,
it can be described in reference to a unit sphere, of radius, R = 1. A right-handed cartesian coordinate
system is defined with zero at the center of the sphere. A standard convention, used here, is to select X,
Y, Z = East, North, Up (see Section 4.8 for alternative conventions). A line, L, passing through the
center of the sphere, the origin, will pierce the sphere at two diametrically opposed points (Figure 4.1).
If the line represents undirected axial data (as opposed to directed data), such as a fold axis or the pole
to a joint plane, it is allowable to choose either point. In structural geology the convention is to choose
the point on the lower hemisphere, P (the opposite convention is used in mineralogy). The three
coordinates of point P are known as direction cosines, and uniquely define the direction of the line.
More commonly, the trend (azimuth or declination) and plunge (inclination) of the line are given. In
Figure 4.1, the trend of the line is 090°, and it's plunge is δ. It is a helpful reminder to designate
horizontal angles using three digits, where 000° = north, 090° = east, 180° = south, etc., and to specify
vertical angles using two digits, from horizontal, 00°, to vertical, 90°. Note that directed data, such as
fault slip directions, may have negative, upward directed, inclinations.
An important tool for plotting line and plane data by hand, and for geometric problem solving, is a
spherical net. A spherical net is a grid formed by the projection of great and small circles, equivalent to
lines of longitude and latitude. Nets are commonly either meridianal or polar, that is, projected onto a
meridian (often the equator) or a pole. The terms equator and pole (or axis) will be used to refer to the
equivalent geometric features on the net, it is essential to remember that they do not have an absolute
reference frame, that is, the net axis is not equivalent to geographic north. When used to plot data by

23
hand, an overlay with an absolute geographic reference frame (North, East, South, etc.) is used (Ragan
2009).
The projections described here are spherical projections, so equal-area projection is assumed to mean
equal-area spherical projection. Other projections are possible, such as hyperboloidal projections,
which include equal-area and stereographic hyperboloidal projections (Yamaji, 2008; Vollmer, 2011).
In these projections the surface of a hyperboloid is projected onto a plane. These are used in the context
of strain analysis, and are unlikely to be confused with the more common spherical projections.

Figure 4.1 Definition of the

point, P, on the unit sphere that
defines the orientation of the
undirected line L. The line is
trending toward X (east) and it's
plunge is δ. The Y coordinate axis
(north) is into the page.

4.3 Orthographic Projection

The orthographic projection is an important projection in which points are projected along parallel rays,
as if illuminated by an infinitely distant light source. Figure 4.2 gives the geometric definition of the
orthographic spherical projection. A corresponding orthographic polar net is shown in Figure 4.3, and
an orthographic meridianal net in Figure 4.4. The projection of point P in the sphere to point P' on the
plane is parallel to the Cartesian axis Z, effectively giving a projection following a ray from Z equals
positive infinity. This type of projection gives a realistic view of a distant sphere, such as the moon
viewed from Earth. It is azimuthal, but angles and area are not generally preserved. When plotting
geologic data it is important that area, and therefore data densities, are preserved, so the orthographic
projection unsuitable for such purposes. The net does, however, have other uses, such as the
construction of block diagrams (e.g., Ragan, 2009).

24
Figure 4.2 Geometric definition of the Figure 4.3 Polar orthographic Figure 4.4 Meridianal
orthographic spherical projection. Point net. orthographic net.
P on the sphere is projected to point P'
on the plane.

4.4 Stereographic Projection

The stereographic or equal-angle spherical projection is widely used in mineralogy and structural
geology. It was known to the Greeks Hipparchus and Ptolemy, and given its present name by François
d'Aguilon in 1613 (Snyder, 1987). It is defined geometrically by a ray passing from a point on the
sphere (here Z = 1) through a point P on the sphere to the projected point P' on the plane (Figure 4.5).
The corresponding stereographic nets are shown in Figures 4.6 and 4.7. Both hemispheres can be
represented, however the convention in structural geology is to use the lower hemisphere. The
meridinal stereographic net is known as a stereonet (Bucher, 1944; Billings 1954; Donn and Shimer,
1958; Badgley, 1959), or Wulff net, named after the crystallographer G.V. Wulff who published a
stereographic net in 1902 (Whitten, 1966), although a much older example was published by François
de Aguilón in 1613 (Section 4.15). The stereonet is commonly used in mineralogy, however, the
convention is to use the upper hemisphere. It is therefore good practice to clearly label all projections,
for example lower-hemisphere stereographic projection.

25
Figure 4.5 Geometric definition of the Figure 4.6 Polar stereographic Figure 4.7 Meridianal
stereographic projection. Point P on the net. stereographic net, stereonet or
sphere is projected to point P' on the Wulff net.
plane.

The projection is azimuthal, so lines passing through the center of the projection have true direction,
these represent great circles. Note that area in Figures 4.6 and 4.7 is distorted, the projection preserves
angles (is conformal), but it does not preserve area. An important consequence is that great circles
(such as meridians) and small circles project as circular arcs. These properties make it useful for
numerous geometric constructions in structural geology (Bucher, 1944; Phillips, 1954; Donn and
Shimer, 1958; Badgley, 1959; Lisle and Leyshorn, 2004; Ragan, 2009).

4.5 Equal-Area Projection

The Lambert azimuthal equal-area spherical projection was presented by Johann Heinrich Lambert in
1722 as a map projection that preserved area (Lambert, 1722; Snyder, 1987; Section 4.11). It is not
conformal, however it preserves area, so densities are not distorted. This equal-area projection is used
widely in the geologic literature for the presentation of directional data, and is the most likely of these
projections to be encountered in literature related to structural geology. Figures 4.8, 4.9, and 4.10
illustrate the geometric definition, polar net, and meridianal net respectively.
In 1925 Walter Schmidt (see Section 4.12) recognized that the equal-area projection should be used for
plotting samples of rock fabrics, such as crystallographic axes, this was later extended to bedding
planes, joints, and other directional data (Schmidt, 1925; Sander, 1948, 1950, 1970; Phillips, 1954;
Badgley, 1959; Turner and Weiss, 1963; Whitten, 1966; Ramsay, 1967; Hobbs et al., 1976; Fisher et al.,
1987; Mardia and Jupp, 2000; Van der Pluijm and Marshak, 2004; Pollard and Fletcher, 2005; Twiss
and Moores, 2007; Ragan, 2009; Fossen, 2016).
The term azimuthal indicates that, like stereographic and orthographic projections, lines passing
through the center have true direction, and that it is projected onto a plane. This distinguishes it from
other equal-area projections, which include the projection of a sphere onto conical and other surfaces,
however, in structural geology, it can usually be referred to simply as an equal-area projection without
ambiguity. The projection is also known as the Schmidt projection, after Walter Schmidt who first used
it in structural geology (Schmidt, 1925; Turner and Weiss, 1963), and the meridianal equal-area net

26
(Figure 4.10), is known as a Schmidt net (Knopf and Ingerson, 1938; Billings, 1942; Sander, 1948,
1950, 1970; Mardia and Jupp, 2000). The polar equal-area net, or Billings net (Figure 4.9), is useful for
rapidly plotting measurements by hand.

Figure 4.8 Geometric definition of Figure 4.9 Polar equal-area net, Figure 4.10 Meridianal equal-
the equal-area projection. or Billings net. area net, or Schmidt net.

4.6 Density Distortion

A common requirement is to plot a sample of directional data to evaluate their density distribution, a
descriptive statistical procedure intended to identify significant clusters, girdles, or other patterns. In
1925 Walter Schmidt recognized that the distortion of area makes the stereographic projection
unsuitable for studying rock fabric data, and proposed the use of the Lambert equal-area projection for
such directional data (Figure 4.10; Section 4.11).
To illustrate the effect of density distortion, 2048 directed data points were calculated on a spherical
Fibonacci grid (Swinbank and Purser, 2006; command Data Fibonacci Sphere), and plotted using
orthographic (Figure 4.11), stereographic (Figure 4.12), and Lambert equal-area (Figure 4.13)
projections. The points have equal densities on the sphere (except at the very center), but the densities
are highly distorted by the orthographic and stereographic projections.

Figure 4.11 Lower-hemisphere orthographic Figure 4.12 Lower-hemisphere stereographic

projection of 2048 directed data points on a projection of data as in Figure 4.11 to illustrate
spherical Fibonacci grid to illustrate density density distortion.
distortion. The Data Fibonacci Sphere command
will generate this distribution. 27
 Figure 4.13 Lower hemisphere Lambert equal-
area projection (Schmidt plot) of data as in Figure
4.11 illustrating the lack of density distortion.

The lack of distortion in Figure 4.13 illustrates the advantage of the Lambert equal-area projection for
plotting fabric data, and why it is normally used (Schmidt, 1925; Sander, 1948, 1950, 1970; Phillips,
1954; Badgley, 1959; Turner and Weiss, 1963; Whitten, 1966; Ramsay, 1967; Hobbs et al., 1976;
Fisher et al., 1987; Mardia and Jupp, 2000; Van der Pluijm and Marshak, 2004; Pollard and Fletcher,
2005; Twiss and Moores, 2007; Ragan, 2009; Fossen, 2016). The term Schmidt plot has been proposed
for such a plot (Vollmer, 2015; Section 4.12), and is used in subsequent sections.

4.7 Maxima and Eigenvectors

Spherical directional data is characterized by either directed unit vectors or undirected axes in space.
The concept of a mean value is familiar when dealing with scalar values like temperature. Determining
such a value for directional data is more complex, averaging trends and plunges separately does not
work. As with circular data, a useful measure is the vector mean, the center of mass for directed data.
The procedure for finding the mean is similar to that for directed circular data. A data point on a unit
sphere can be represented by three coordinates, which are its direction cosines. If we designate this
point as the vector x = (x1, x2, x3), then the mean vector is the averaged sum:

28
with mean resultant length:

and mean direction, the normalized mean vector:

Axial data requires the computation of eigenvectors, an important concept that gives best-fit values, or
moments of inertia, for tensors, such as the principal stresses of a stress tensor. In the context of
directional data, imagine that each line passing through the center of the unit sphere is represented by a
small mass at each of the two points where it pierces the sphere. If the sphere were spinning, it would
have a tendency to spin about the axis of minimum density, this is the minimum eigenvector. If it were
rolling, it would have a tendency to stop with the maximum density at the bottom, this is the maximum
eigenvector, the intermediate eigenvector is exactly 90° from the other two. As an example, Figure 4.14
is a Schmidt plot of bedding plane normals represented by points on the lower hemisphere.
The orientation tensor, or scatter matrix (Fisher, Lewis and Emblton, 1987; Mardia and Jupp, 2000), is
given by:

the averaged sum of the unit directional vectors times their transpose. The eigenvectors of this matrix
are determined and plotted. In Figure 4.15 the maximum eigenvector is plotted as a red circle, and the
minimum a blue circle. In cylindrical folds the minimum eigenvector of poles to bedding gives an
estimate of a cylindrical fold axis, and a great circle normal to the minimum eigenvector gives the best-
fit plane through the plane poles (Figure 4.15).

Figure 4.14 Schmidt scatter plot of 56 poles to Figure 4.15 Schmidt plot as in Figure 4.14, with
folded bedding in graywacke, Albany County, New maximum (red), minimum (blue) eigenvectors, and
York (data from Vollmer, 1981). 29 the great circle normal to the minimum eigenvector.
The minimum eigenvector is an estimate of the fold
axis.
See Figure 1.9 and Section 5.6 for a visualization of this data. For details of statistical measures and
tests for spherical directional data see Mardia (1972), Cheeney (1983), Davis (1985), Fisher et al.
(1987), and Mardia and Jupp (2000).

4.8 Tutorial 4 – Scatter Plots

Open the file Vollmer 1981a from the Example Data folder (any of the csv, tsv, ods, or xlsx versions) in
Orient, and click on the Spherical Projection icon.
If not already done, remove the Schmidt net from the background. Click on the Preferences icon, and
locate the Net pane under the Spherical Projection settings. Uncheck both Axes and Net. Next, in the
Labels pane, change the Increment to -90, Offset to 14, and the Size to 12.
In the Symbols pane, Data Type S0 should be selected, and the Symbol checkbox checked. Set the Fill
Symbol Color to yellow. At this point the plot should look as in Figure 4.14.
To add eigenvector maxima, select the Maxima pane, and check Visible. Under Maximum Eigenvector
check Symbol, and under Minimum Eigenvector check Symbol and Great Circle. Set the great circle
Stroke Width to 2. The plot should now look as in Figure 4.15. This type of scatter plot gives a simple
summary of the data distribution. Section 4.9 shows how to add contouring, and Chapter 6 discusses
how to add confidence cones about the maxima.

4.9 Contouring
A simple Schmidt scatter plot of data using an equal-area projection (Figure 4.14), or with maxima
displayed (Figure 4.15) may suffice for some data sets. However, since the first use of lower
hemisphere equal-area projections for displaying directional data (Schmidt, 1925), they have
commonly been contoured to bring out underlying patterns such as girdles or point clusters (Figure
4.16).

Figure 4.16 Early examples of contoured lower hemisphere equal-area projections. (A) Crystallographic
fabric in a sample of granite (Schmidt, 1925). (B) Crystallographic axes in a sample of ice (Kamb, 1959).

30
Schmidt (1925) devised a method to hand contour directional data using percentages. This method, or
variants of it, is still used (Ragan, 2009), although it is generally superseded by computerized methods
(Vollmer, 1995; Ragan, 2009). Schmidt contouring, however, suffers in that it does not correctly take
into account the sample size. Kamb (1959) therefore introduced a method using binomial statistics,
giving a greater statistical validity to the contours (Kamb, 1959; Vollmer, 1995).
A problem with both the Schmidt and Kamb methods, in addition to being hand-contouring methods, is
that the density calculations are done on the projection plane after projecting the data points from the
sphere, causing distortion. Therefore a modified Kamb method, which calculates density directly on the
sphere, was introduced in Orient in 1986 (Vollmer, 1988, 1990, 1995). An alternative method, using
probability density estimation on the sphere, is given by Diggle and Fisher (1985). Orient implements
the modified Kamb, modified Schmidt (Vollmer, 1995), and probability density methods.
The settings related to contouring are in the Preferences dialog on the Spherical Projection Contours
and Gridding panels. Gridding is the first step in contouring, in which the density calculations are done.
In this panel select the Method, Modified Kamb or Probability Density (Modified Schmidt is not
recommended). Normally Weighting should be left at the default setting Exponential, Sigma at the
default 3, and Calculate Kappa checked. Details for these settings can be found in Vollmer (1995) and
Diggle and Fisher (1985). The number of nodes, Nodes, sets the number of calculated grid points, the
default of 100 gives 10000 calculated nodes. A value of 30 is the approximate minimum for acceptable
contours, the default value of 100 generally gives very good resolution.
Once the gridding method is selected, the Contour Settings pane is used to set the number and type of
contour lines. Figure 4.17 is a Schmidt scatter plot of crystallographic axes of ice (from Kamb, 1956),
which is shown contoured using the 3σ modified Kamb method with a 2σ contour interval in Figure
4.18. This is the data and contour levels that Kamb used to hand-contour Figure 4.16 B. The Contour
levels option selected, and an Interval of 2 entered for this plot.

Figure 4.17 Schmidt scatter plot of crystallographic Figure 4.18 Schmidt plot with 3σ modified
axes of ice (data digitized from Kamb, 1956). Kamb contours of the data from Figure 4.17 with
contours at 2σ. Compare with Figure 4.16 B. This
plot uses a WBGYR Gradient option.

31
It is also possible to scale the contour levels to multiples of uniform density, by checking Uniform
density from the Gridding Settings pane. This option is only available for Exponential weighting of the
Modified Kamb method, and when Contour interval is selected. This rescales the grid so the contour
interval is in multiples of the expected uniform density. A scale bar can also be optionally displayed for
the color gradient scale (Figure 4.19).
An alternate, and perhaps preferable, choice for selecting contours is to contour the density distribution
at equal levels (Fisher et al., 1987). Figure 4.20 is a 3σ modified Kamb contour plot of this data
contoured at 20% density, 5 equally spaced levels over the density distribution. To do this, select
Contour levels In the Contour Settings panel, and enter the number of Contour levels, in this case 5
which gives contours at 20% density.

Figure 4.19 Schmidt plot of data in Figure 4.17 Figure 4.20 Schmidt plot of data in Figure 4.17
with 3σ modified Kamb contours at multiples of with probability density contours at 20% density, 5
uniform density, with scale bar. equally spaced levels over the density distribution.

Figure 4.21 is an additional example of this data, but contoured at 10% density, using the BCYR
gradient. Figure 4.21 is the same plot, but with the Fill contours option selected in the Contour Settings
pane.
An additional method that can be used is the probability density method of Diggle and Fisher (1985).
An example plot is shown in Figure 4.23. In this case the resulting contours are similar to those given
by the 3σ modified Kamb method (Figure 4.21), however this is not the case for all data sets.
Finally, Figure 4.24 is a 3σ modified Kamb contoured Schmidt plot of the folded graywake data from
Figure 4.14 for an example of data with a strong girdle disstribution.

32
Figure 4.21 Schmidt plot of data in Figure 4.17 Figure 4.22 Schmidt plot of data in Figure 4.17
with 3σ modified Kamb contours at 10% density. with 3σ modified Kamb contours at 10% density,
This plot uses the BCYR Gradient option. and the Gradient and Fill Contours options on.

Figure 4.24 Schmidt plot of poles to bedding

Figure 4.23 Schmidt plot of data in Figure 4.17 with from Figure 4.14 with modified Kamb contours at
probability density contours at 10% density. In this 20% density.
case, the method gives contours similar to the 3σ
modified Kamb method (Figure 4.21).

33
4.10 Tutorial 5 – Contour Plots
Open the file Vollmer 1981a from the Example Data folder (any of the csv, tsv. ods, or xlsx versions) in
Orient, and click on the Spherical Projection icon. Click on the Preferences icon, and locate the Net
pane under the Spherical Projection settings. Uncheck both Axes and Net. Next, in the Labels pane,
change the Increment to -90, Offset to 14, and Size to 12. If still on from the previous tutorial, turn off
eigenvector and confidence cone display in the Maxima panel. The Help Restore Defaults command
can be used to reset all preferences if desired.
In the Symbols pane the S0 (bedding) data set should be selected. Uncheck the Symbol option, and
check both Contour and Gradient. Click on the gradient paint picker and select the WBGYR (White
Blue Green Yellow Red) preset. Finally, in the Contours pane select Contour levels, and set the Levels
to 5. This gives modified Kamb contours at 20% density levels, and the plot should appear as in Figure
4.24.

4.11 History and Terminology

Spherical projections have long been used for mapping the terrestrial and celestial spheres. The
stereographic projection was known to the Greeks Hipparchus and Ptolemy, and was given its present
name by François d'Aguilon in 1613 (de Aguilón, 1613; Snyder, 1985; Figures 4.25 and 4.26). The
more recent Lambert azimuthal equal-area projection was invented by Lambert in 1772 (Lambert,
1772; Snyder, 1985; Figure 4.27).

Figure 4.25 Illustration by Peter Paul Rubens showing a globe held by Atlas illuminated by winged putti to
illuminate the construction of a stereographic net (de Aguilón, 1613).

34
 Figure 4.26 Stereographic net, or stereonet, published by François de
Aguilón, 1613.

Figure 4.27 Illustration of the Lambert azimuthal equal-area

projection by Johann Heinrich Lambert (Lambert, 1722).

In 1925 Walter Schmidt recognized that the stereographic projection was unsuitable for directional data
analysis due to its distortion of area (Section 4.6), and introduced the use of the equal-area projection
for fabric analysis. Rejecting the stereographic, or Wulff, net used by mineralogists, Schmidt
introduced the equal-area net, or Schmidt net, as well as data contouring (Figures 4.28 and 4.29).
In 1944 Walter H. Bucher introduced the stereonet for use in structural geology in North America
(Bucher, 1944; Billings 1954; Donn and Shimer, 1958; Badgley, 1959). Bucher defined stereonet as a
shorthand term for stereographic net (Bucher, 1944, p. 193), and stereogram as a diagram produced
using the stereographic net (Bucher, 1944, p. 194). The phrases equal-area stereonet, and equal-area
stereogram are contradictions.

35
 Figure 4.28 Lambert equal-area net, or Figure 4.29 Lower hemisphere equal-area
Schmidt net, as published by Walter projection, or Schmidt plot, of
Schmidt in 1925. crystallographic fabric (Schmidt, 1925).

The equal-area projection is correctly referred to as the Lambert azimuthal equal-area projection
(Snyder, 1985), although in the context of directional data analysis, it is usually referred to simply it as
the equal-area projection. The projected hemisphere should be given, as lower hemisphere equal-area
projection. Note that other spherical equal-area projections exist, such as the cylindrical equal-area
projection (Snyder, 1985). Additionally, hyperboloidal equal-area and stereographic projections exist,
and are used for some geologic data (Yamaji, 2008; Vollmer, 2011). The term Schmidt projection has
been used for the projection (Mardia and Jupp 2000), and the term Schmidt plot is suggested for plots
produced using the projection (Vollmer, 2015; Section 4.12).
Early references (Schmidt, 1925; Billings 1942, 1954; Bucher, 1944; Sander, 1948, 1950, 1970;
Phillips, 1954; De Sitter, 1956; Donn and Shimer, 1958; Badgley, 1959; Turner and Weiss, 1963; Hills,
1963, Whitten, 1966; Ramsay, 1967, Hobbs et al., 1976) are careful to use correct terminology, as are
most current structural geology texts (Marshak and Mitra, 1988; Van der Pluijm and Marshak, 2004;
Pollard and Fletcher, 2005; Twiss and Moores, 2007; Ragan, 2009; Fossen, 2016). Note that:
• The equal-area projection is not a type of stereographic projection
• A stereonet, or Wulff net, is a stereographic net
• A Schmidt net, or equal-area net, is not a stereonet
• Scatter plots and contour plots are not stereonets
• The phrase equal-area stereonet is a contradiction
• The phrase equal-area stereographic projection is a contradiction

4.12 Schmidt Plots

The Schmidt net (Schmidt, 1925) was widely used in structural geology (Knopf and Ingerson, 1938;
Billings 1942; Turner and Weiss, 1963) prior to the introduction of the stereonet (Bucher, 1944;
Billings, 1952; Phillips, 1954), however it is common to see spherical projection scatter plots and
contour plots mislabeled as stereonets. As they are not (cf. Figures 4.7 and 4.14), what they represent,
and the projection used to make them, is unclear. The equal-area projection and the Schmidt net have a
long and important history in structural geology, mislabeling them as stereonets does not give due

36
credit to their inventor. In 1925 Schmidt recognized that the stereographic projection was not suitable
for directional data analysis, and invented the Schmidt net 90 years ago.
Although diagrams produced using Schmidt's equal-area method (Schmidt, 1925) are ubiquitous in
structural geology and tectonics, no succinct term exists for them. The term Schmidt plot therefore has
been suggested for a lower-hemisphere Lambert azimuthal equal-area spherical projection of three-
dimensional directional data, such as foliation planes, joints, slickensides, magnetic vectors,
crystallographic axes, fold axes, and lineations (Vollmer, 2015). These plots, which are often
contoured, have been in common use in structural geology, tectonics, and related disciplines, since their
introduction by Walter Schmidt in 1925.

37
5. Coordinates and Rotations
5.1 Introduction
The rotations used for projection and data display are normally transparent to the user, and the default
settings are sufficient in most cases. The default coordinates are local coordinates commonly used in
geology (North at the top, and East at the right), however, it is easily configured for other common
geographic or spherical coordinate systems, which is the first topic covered in in this section.
The second topic covered here is coordinate system rotation, or projection rotation. In some cases it is
useful or necessary to view data in a more specialized coordinate system, for example centered over a
specific location, or chosen parallel to one of the data maxima. The third topic is data rotation, which
may be required for a number of reasons, such as paleomagnetic fold tests, or rotation of paleocurrent
measurements back to horizontal.
Note that for directed data, symbols for data not on the current hemisphere will not appear. Rotations
involving directed data will therefore often hide those data points. The Directed and Undirected drop-
down menus contain options to display hidden directed data, as well as the hidden duals of undirected
data. See Section 5.6 for a discussion of these options.

Figure 5.1 The Preferences dialog Spherical Projection

Settings pane displaying options for projections. The Axes
display, here X Right, Y Top, Z Out, gives the current
positions of the coordinate axes.

5.2 Geographic Coordinates

The primary coordinate system is selected in the Preferences dialog in the Spherical Projection
Projection Settings panel (Figure 5.1). In addition to the Projection and Hemisphere drop-down menus,
there is an Orientation drop-down menu from which six geographic coordinate systems ( Polar,
Antipolar, Equatorial 0, Equatorial 90, Equatorial 180, and Equatorial 270) and six spherical systems
(XY, ZX, YZ, YX, XZ, and ZY planes) can be selected. These options allow views along any of the three
axes in either direction. Additionally, there are checkboxes for inverting the projection about the Y and
X axes. These are used, for example, to produce plots like Figure 1.3, a plot of directed magnetic data

38
displayed on both hemispheres, as they essentially flip the hemisphere over. The Axes display changes
when any of these are modified, so the current coordinate system is clear, for example X Right, Y Top, Z
Out is displayed for local coordinates.
Figures 5.2 to 5.7 are upper hemisphere equal-area projections of 14,229 earthquake epicenters with
magnitudes greater than 4.5 from 1980 to 1990 (data from NOAA) plotted on the six standard
geographic coordinate systems.

Figure 5.2 Upper hemisphere equal-area Figure 5.3 Projection as in Figure 5.2, but
projection of 14,229 earthquake epicenters with centered at 0°, 90°, the Equatorial 90 coordinate
magnitudes greater than 4.5 from 1980 to 1990, system. Compare with Figure 4.26 from Lambert,
centered at latitude, longitude 0°, 0°, the 1722.
Equatorial 0 coordinate system (data from
NOAA).

39
Figure 5.4 Projection as in Figure 5.2, but Figure 5.5 Projection as in Figure 5.2, but
centered at 0°, 180°, the Equatorial 180 centered at 0°, -90° , the Equatorial 270
coordinate system. coordinate system. Compare with Figure 4.26
from Lambert, 1722.

Figure 5.6 Projection as in Figure 5.2, but Figure 5.7 Projection as in Figure 5.2, but
centered at 90°, 0°, the Polar coordinate system. centered at -90°, 0°, the Antipolar coordinate
system.

The standard coordinate systems provide views of data along all coordinate axes, however it is also
possible to select a view along an arbitrary axis using the Rotate Projection command. The Rotate
Projection dialog is shown in Figure 5.8.

40
 Figure 5.8 The Rotate Projection dialog
showing the settings for a rotation of 74.156°
about the Z axis, which is currently at the top
of the projection.

For example, to produce a projection centered at latitude, longitude 41.764, -74.156, first set the
coordinate system to Equatorial 0. The projection, as in Figure 5.2, is now centered at 0°, 0°, with X
Out, Y Right, Z Top. Set Z as the Axis, enter 74.157 into the Angle edit box, and press Apply. The
projection is now centered at 0°, -74.157°. Then set Y as the Axis, enter 41.764 into the Angle edit box,
and press Apply. The projection is now centered at 41.764°, -74.157° (Figure 5.9).

Figure 5.9 Projection as in Figure 5.2, but

centered at latitude, longitude 41.764°, -74.157°.

5.3 Tutorial 6 – Geographic Coordinates

Open the file Earthquakes With Continents from the Example Data folder, using any of the provided
formats, and click on the Spherical Projection icon. The projection will now appear extremely
cluttered. To remedy this, open the Preferences dialog and select the Spherical Projections Symbols
pane. For the data type CO, uncheck Symbol, check Polyline, and check Directed. For the data type EQ,

41
set the Symbol Size to 1, the Stroke Color and Fill Color to red, and the Stroke and Fill Opacities to 50%.
Check Directed.
Next select the Net pane, uncheck Axes, and check Net. For the Y axis, uncheck both Major and Minor
Great Circles and Small Circles. For the Z axis, check Major Great Circles and Major Small Circles, and
set both Stroke Colors to a very light gray.
Select the Projection panel, set the Hemisphere to Upper Hemisphere, and the Orientation to Equatorial
0. The projection is now centered at latitude, longitude 0°, 0°, X is out, Y is right, Z is at the top, and
the plot should appear as in Figure 5.2.
To center the projection at latitude, longitude 41.764°, -74.156°, open the Rotate Projection dialog
(Figure 5.8). Set Z as the Axis, enter 74.157 into the Angle edit box, and press Apply. The projection is
now centered at 0°, -74.157°. Then set Y as the Axis, enter 41.764 into the Angle edit box, and press
Apply. The projection is now centered at 41.764°, -74.157° (Figure 5.9).

5.4 Projection Rotation

As was shown in the Section 5.2, it is possible to rotate the projection to any orientation. The Rotate
Projection dialog (Figure 5.8) allows the sequential application of rotations about any coordinate axis,
or about arbitrary axes, by pressing the Apply buttton. The rotations can be undone in the order they
were applied, using the Undo button, or can be completely removed using the Revert button.
In examining some data sets, such as crystallographic axes, it may be desirable to produce a plot whose
axes are related to the distribution, such as the eigenvectors (Fisher et al., 1987). Figure 5.10 shows the
Rotate Projection dialog with settings to rotate a data set to the maximum eigenvector. Any of the three
eigenvectors, the mean vector, or the pole to the best fit small circle (Section 7.3), can be selected.

Figure 5.10 The Rotate Projection dialog

showing the settings for a rotation to center
the projection on the maximum eigenvector
of the S0 data type.

Figure 5.11 is a Schmidt plot with modified Kamb contours at 20% of the poles to bedding data shown
in Figure 4.14, which is then shown rotated to the maximum (Figure 5.12), intermediate (Figure 5.13),
and minimum (Figure 5.14) eigenvectors.

42
Figure 5.11 Schmidt plot with modified Kamb Figure 5.12 Contour plot as in Figure 5.11,
contours of poles to bedding data from Figure rotated to the maximum eigenvector.
4.14 with contours at 20% density.

Figure 5.13 Contour plot as in Figure 5.11, Figure 5.14 Contour plot as in Figure 5.11,
rotated to the intermediate eigenvector. rotated to the minimum eigenvector.

5.5 Tutorial 7 – Rotation to Maxima

Open the file Vollmer 1981 from the Example Data folder (any of the formats is fine) in Orient, and
click on the Spherical Projection icon.
If not already done, remove the Schmidt net from the background. Click on the Preferences icon, and
locate the Net pane under the Spherical Projection settings. Uncheck Great circles, Small circles, and
Minor Tick marks. Set the Major Tick marks increment (Inc) to 10, and the length (Len) to 4 for the X, Y,
and Z axes. Check Above. Next, in the Labels pane, change the Increment to -90, the Offset to 14, and
the Size to 12.

43
In the Spherical Projection settings, restore the projection to Lower Hemisphere and Local coordinates
if necessary. In the Symbols pane, uncheck Symbol for the S0 data type, and check both Contour and
Gradient. Click on the gradient paint picker and select the W BGYR (White Blue Green Yellow Red)
preset. Finally, in the Contours pane set Levels to 5. This gives modified Kamb contours at 20%
density levels, and the plot should appear as in Figure 5.11.
Use the Rotate Projection command to open the Rotate Projection dialog (Figure 5.10). Set Axis to
Maxima, Data Type to S0, and Eigenvector to Maximum to produce Figure 5.12, Intermediate for
Figure 5.13, and Minimum for Figure 5.14.

5.6 Tutorial 8 – Data Visualization

It can be difficult to visualize the three dimensional geometry of data plotted on spherical projections.
In this tutorial we will use the folded bedding data shown in Figure 4.14 with projection rotation to get
a visualization of the geometric significance of poles to folded beds.
Open the example data file Vollmer 1981a, and display a spherical projection using the Graph
Spherical Projection command. For this visualization we will use the orthographic projection. While it
distorts area and density (Section 4.6), it gives an easily visualized image of a sphere as viewed from a
distance. You may also want to try using an upper hemisphere projection which you may find easier to
visualize, but start with the lower hemisphere, which is the standard one for such plots.
From the Spherical Projection pane of the Preferences dialog, select Orthographic. In the Symbols
pane, set the symbol Fill Color to yellow, and check the Ray option. Next, in the Maxima pane check
Visible. For the Maximum Eigenvector check Symbol and Ray. For the Minimum Eigenvector, check
Symbol, Ray, and Great Circle. Set the Stroke Width of the Great Circle to 2. The resulting Schmidt
plot is shown in Figure 5.15. Finally, in the Spherical Projection panel, select Undirected 50% Opaque,
to give Figure 5.16.
At this point the data elements on the hidden (upper) hemisphere are translucent. Imagine that you can
see through them down to the lower hemisphere. Now open the Rotate Projection dialog using the
Graph Rotate Projection command, set Axis to X and Angle to 10. Apply increments of rotation using
the Apply button, example results are shown in figures 5.17 and 5.18.

44
Figure 5.15 Schmidt plot of data from Figure Figure 5.16 Plot as in Figure 5.15 with
4.14 with symbol rays, and maximum (red) and Undirected 50% opaque selected in the Spherical
minimum eigenvectors (blue) displayed with rays Projection pane, so data on the hidden (upper)
and a great circle girdle. hemisphere is transparent.

Figure 5.17 Plot at in Figure 5.15 after a rotation Figure 5.18 Plot as in Figure 5.16 after a rotation
of 30° about the X axis. of 160° about the X axis.

5.7 Data Rotation

The final topic in this section is data rotation. The principal is the same as projection rotation however
the rotations are applied to the data, not the projection. The data can be sequentially rotated about any
of the coordinate axes (Figure 5.19). Before using this command be sure you have a backup file of your
data, as this will change the values in the data window spreadsheet.

45
 Figure 5.19 The Rotate Data dialog allows
rotation of the data about any of the
coordinate axes, or about an arbitrary axis of
any orientation.

To try this command, open the file Vollmer 1981 as in Section 5.5, select all or part of the data, and
apply several rotations using the Rotate Data command. Pressing the Apply button sequentially applies
rotation by premultiplication of a rotation matrix, and updates any open projections. Press OK to accept
the rotated data, or Cancel to revert.

46
6. Confidence Cones and Bootstrapping
6.1 Introduction
Section 4.7 introduced the concept of calculating maxima for samples of directional data. In particular,
the maximum, intermediate, and minimum eigenvectors for undirected, or axial, data, and the vector
mean for directed, or vector data. The maximum eigenvector and vector mean are measures of the mean
value, or direction about which the samples cluster. The minimum eigenvector is used to locate the
center of a great circle girdle, such as a cylindrical fold axis. The relative magnitudes of the
eigenvectors give additional information about the type of distribution, this is discussed further in
Chapter 9. The best fit small circle is discussed in Section 7.3.
The concept of a mean value is of great significance, but it does not say anything about the certainty of
the estimate. This chapter discusses ways to provide uncertainty estimates, in particular deriving
confidence cones about the mean. Only a basic overview is provided here, the statistical analysis of
directional data is discussed in detail by Fisher et al. (1987), and Mardia and Jupp (2000). These texts
should be consulted for additional details and statistical tests.

6.1 Spherical Distributions

There are a number of models for distributions of spherical directional data (Mardia, 1972; Fisher et al.,
1987; Mardia and Jupp, 2000). By analogy, the Gaussian, or normal, distribution is a familiar model for
scalar data. Four important distributions on the sphere are the Fisher, Kent, Watson, and Bingham
distributions.
The Fisher (or von Mises-Fisher) distribution is the basic model for unimodal rotationally symmetric
directed data, so, for example, is applicable to paleomagnetic vectors. This is a general and useful
statistic for directed data. A confidence region is calculated for a symmetrical unimodal distribution,
which gives a circular confidence cone.
The Kent distribution (Kent, 1982) is also a model for unimodal directed data, but does not assume the
distribution is rotationally symmetric, so the confidence cone is elliptical, rather than circular. The Kent
statistic is calculated using moment estimation, similar to the Watson and Bingham statistics, but is
centered on the vector mean.
The Watson distribution serves as the basic model for rotationally symmetric undirected lines (axes).
The distribution has rotational symmetry as a bipolar or girdle distribution. The confidence region is
bipolar, giving two confidence cones centered 180° apart in opposing hemispheres. Confidence cones
can be calculated for the three axes of asymmetrical polar, girdle, or multimodal distributions, giving
elliptical confidence cones. This distribution can be used for features such as normals to joint planes, or
quartz c-axes. It is common to determine the axes of folded bedding surfaces using the pole to a girdle,
however note that assigning a younging direction to bedding allows it to be treated as vector data.
The Bingham distribution is also for undirected lines (axes), but does not assume rotational symmetry.
It can be regarded as an extension of the Watson distribution, which is the Bingham distribution with
rotational symmetry (Mardia and Jupp, 2000). Confidence cones are similarly bipolar. Fisher et al.
(1987) recommend the simpler Watson model in most cases. Confidence cones can be calculated for
each of the three axes, giving elliptical cones.

47
Using these models, confidence cones can be drawn around either the mean direction (Fisher or Kent),
or the eigenvectors (Watson or Bingham), at confidence levels of 90%, 95%, or 99%. Note that a
minimum sample size of 25 is required for these, bootstrapping (Section 6.2) should be applied for
sample sizes less than 25.
Figure 6.1 shows the directed paleomagnetic data from Figure 1.3 with a 95% confidence cone based
on the Fisher distribution. Figure 6.2 shows the same data with a 95% confidence cone based on the
Kent distribution.

Figure 6.1 Schmidt plot of paleomagnetic data as Figure 6.2 Paleomagnetic data as in Figure 6.1
in Figure 1.3 with 95% confidence cone based on with 95% confidence cone based on a Kent
a Fisher distribution. Upper hemisphere data distribution.
shown as 30% opaque.

Figure 6.3 shows 95% confidence cones using the Watson distribution for folded bedding data. Note
that, because of the strong girdle pattern in this data, the semiapical angles of the confidence cone
about the minimum eigenvector are small, while the maximum is strongly elliptical within the girdle.
This shows that the minimum is well constrained, while the position of the maximum within the girdle
is less well defined. Figure 6.4 shows similar confidence cones for ice crystallographic data. Figures
6.5 and 6.6 show confidence cones based on the Bingham distribution for the same data sets.

48
Figure 6.3 Schmidt plot of poles to folded Figure 6.4 Ice crystallographic axes as in Figure
greywacke beds, as in Figure 4.14 with 95% 4.17 with 95% confidence cones based on the
confidence cones based on a Watson distribution Watson distribution. Colors are as in Figure 6.3.
about the maximum (red), intermediate (green),
and minimum (blue) eigenvectors.

Figure 6.5 Plot as in Figure 6.3 with 95% Figure 6.6 Plot as in Figure 6.4 with 95%
confidence cones based on the Bingham confidence cones based on the Bingham
distribution. distribution.

Data point weighting is implemented for Fisher, Kent, Watson, and Bingham confidence cones, but is
not for bootstrap confidence cones (Section 6.3). Bootstrapping creates replicates by random selection,
and assumes each data point is equally weighted.

49
6.3 Bootstrapping
The models discussed in section 6.2 assume an underlying distribution model and require a minimum
sample size of 25. An alternate method is to use bootstrapping to infer confidence regions about the
mean. The data set is resampled numerous times and the mean calculated each time to give a large
number of replicate means, which can be used to determine elliptical confidence regions (Fisher et al.,
1987; Mardia and Jupp, 2000; Davis et al., 2015; Figures 6.3 and 6.4). This approach is particularly
useful for small samples (and is required for sample sizes less than 25), or where the underlying
distribution model is not known.
Bootstrapping requires creating a large number of replicates, the default is 1000, with an allowed range
from 100 to 100,000. The means of the replicates can be displayed (Figures 6.8, 6.10) if desired. The
process involves pseudorandom (mersenne twister) number generation, and the resulting confidence
regions will vary slightly each time. Confidence results are presented in the Log window by giving the
axis of the elliptical cone (a direction, which may differ from the mean), the semiapical angle of the
rotationally symmetric (circular) cone, and the semiapical angles and directions of the major and minor
semiaxes of the elliptical cone.
Note that, while data point weighting is implemented for Fisher, Kent, Watson, and Bingham
confidence cones (Section 6.2), it is not implemented for bootstrap confidence cones because
bootstrapping creates replicates by random selection, and assumes each data point is equally weighted.
Figure 6.7 illustrates bootstrap confidence cones for folded graywacke bedding planes treated as
undirected axes, Figure 6.8 shows an example of the bootstrap replicates used to calculate the cones.
Figure 6.9 shows similar bootstrap cones for ice crystallographic data, with an example of replicates
shown in Figure 6.6. While no underlying model is assumed, the confidence cones generated are
comparable to those generated using the Watson distribution, as shown in Figures 6.3 and 6.4.

Figure 6.7 Bedding data as in Figure 6.3 with Figure 6.8 Bedding data as in Figure 6.3 showing
95% confidence cones from 5000 bootstrapped the means of 5000 bootstrap replicates used to
replicate means. This gives confidence cones calculate the ellipses as in Figure 6.7.
similar to the Watson model in Figure 6.3.

50
Figure 6.5 Schmidt plot of ice c-axis data, as in Figure 6.6 Plot of data as in Figure 6.5, with the
Figure 1.1, with bootstrapped elliptical means of 5000 bootstrap replicants on each of the
confidence regions. This gives confidence cones eigenvectors.
similar to the Watson model in Figure 6.2.

The Log file generates a large number of results, however in most cases it is probably sufficient to
report one or two mean values, either the vector mean for directed data, the maximum eigenvector for
undirected data, or the minimum eigenvector for a girdle pole, along with a measure of uncertainty,
such as the 95% symmetric cone semiapical angle. This confidence cone is plotted if the Circular
option is selected for bootstrapping, the confidence cone is then calculated using the angular deviation
from the mean, giving a circular confidence cone which encloses 95% of the replicate means.
Examples are shown in Figure 6.7 of bedding data, in Figure 6.8 of ice crystallographic axes.

51
Figure 6.7 Plot as Figure 6.7 showing symmetric Figure 6.8 Plot as in Figure 6.2 showing
95% confidence cones about the maximum and symmetric 95% confidence cones about the
minimum eigenvectors. The mean bedding as maximum and minimum eigenvectors. The mean
strike, dip is 172.29°, 14.52° with 95% axis as trend, plunge is 256.52°, 8.74° with 95%
confidence of 26.72°. The mean fold axis as confidence of 13.74°. The mean girdle pole is
trend, plunge is 182.76°, 2.69° with 95% 164.84°, 10.90° with 95% confidence 25.66°.
confidence 4.61°.

Figure 6.9 illustrates bootstrap analysis applied to directed data, magnetic remnant vectors. In this
diagram the upper hemisphere is inverted on the left using the Invert about Y option. Figure 6.10 is a
projection of the same data using a 30% transparent upper hemisphere, and Figure 6.11 shows the
projection rotated to the vector mean to more clearly show the distribution.

Figure 6.9 Schmidt plots (data inverted on left) of magnetic remanence data, an example of
directed or vector data, from Figure 1.3 with the means of 5000 bootstrap replicants on the
mean vector.

52
 Figure 6.11 Magnetic data as in Figure 6.10
Figure 6.10 Magnetic data as in Figure 6.9 with rotated to the vector mean. Compare this to the
95% bootstrap confidence cone about the mean, confidence cone generated using the Kent
and 30% transparency of upper hemisphere. distribution in Figure 6.2.

Figure 6.12 shows the replicate means used to calculate the confidence cones. Figure 6.13 shows the
symmetric confidence cone generated with the Circular option. For comparison, the corresponding
circular Fisher confidence cone is shown in Figure 6.14.

Figure 6.12 Magnetic data as in Figure 6.11 with Figure 6.13 Magnetic data as in Figure 6.11 with
5000 bootstrap replicate means displayed. A set of circular bootstrap confidence cone. The vector
such means is used to calculate the confidence mean as trend, plunge is 315.78°, -8.65° with
cone. 95% confidence 13.84°. This gives similar
confidence cones as the Fisher model shown in
Figure 6.3.

53
One final option is to display the confidence region using the Hull option, this displays the convex hull
of all points that fall within the confidence region calculated using the angular deviations. Figure 6.14
shows an example using undirected crystallographic data and centered on the maximum eigenvector.
Figure 6.15 is an example using directed magnetic data centered on the vector mean.

Figure 6.14 Ice crystallographic data as in Figure Figure 6.15 Magnetic data rotated to the vector
6.5 centered on the maximum eigenvector with mean as in Figure 6.11 with 95% bootstrap
bootstrap 95% confidence region. The polygonal confidence region. The polygonal hull encloses
hull encloses 95% of the replicate means sorted 95% of the replicate means sorted by distance
by distance from the maximum. from the mean.

54
7. Conical Data and Small Circles
7.1 Introduction
This chapter concerns two distinctly different topics, data that is entered as a direction and an angle,
and data that can be fitted to a cone.. An example of the first are drill core data from cores that
penetrate layers at a measured angle. Each data point is characterized by three angles (θ, φ, α), where θ,
φ give the direction (e.g., trend and plunge) of the core, and α is the half apex, or semiapical, angle
defining the circular cone.
The second topic, fitting to a cone, is most commonly applied to fitting bedding or foliation data to a
conical fold. Folds are treated as approximately cylindrical when possible (Badgley, 1959; Turner and
Weiss, 1963; Whitten, 1966; Ramsay, 1967; Hobbs et al., 1976), however non-cylindrical folds are
common in nature, and conical folds are an alternate model that may be applicable in some locations
(Wilson, 1967; Kelker and Langenberg, 1982).
A tutorial on data point weighting is included in this chapter as well, using the example of fitting a
small circle. The weighting of individual points applies to other plots and calculations, such as circular
plots and kinematic analysis, but the fitting of small circles provides a useful example for visualization.

7.2 Conical Data

Data that are characterized by both a direction and an angle define a cone, which projects as a small
circle on a spherical projection. In exploration geology it is common to drill a core to determine rock
composition and structure. If the core intersects a plane of interest, such as a bedding plane, the angle
between the core with orientation (θ, φ), and the plane gives a core-plane angle, α, defining the cone (θ,
φ, α). When the core is extracted the orientation of the plane is lost, however it must lie tangent to the
cone (θ, φ, α).
This is one of a class of geometric problems that are commonly solved using a stereographic net, or
stereonet (Bucher, 1944; Phillips, 1954; Donn and Shimer, 1958; Badgley, 1959; Lisle and Leyshorn,
2004; Ragan, 2009). An advantage of the stereographic projection is that, because of its equal-angle
property, small circles project as circular arcs (Section 4.4), and can be constructed using a compass.
Unless this equal-angle property is required, such problems can also be solved using a Schmidt net.
Orient allows such conical data to be entered and will plot small circles, avoiding manual construction,
and aiding in the visualization and solution of drill hole problems. The direction of the drill core (θ, φ)
is typically given by its trend and plunge, enter these, and then the half apex (semiapical) angle α in the
spreadsheet column Alpha. Use the View Data Columns command to display the column if not visible.
As an example, Table 7.1 gives the orientations of three drill cores with core-plane angles of bedding
planes (from Ragan, 2009), the problem is to determine the bedding plane orientation.

55
ID Trend Plunge Alpha
1 0 60 51
2 270 50 67
3 45 55 38

Table 7.1 Example drill core data (from Ragan, 2009). Given the trend and plunge of three drill cores, and the core-plane
angle α (Alpha) of bedding planes, determine the orientation (strike and dip) of the bedding.

To solve the problem, open a new data sheet in Orient, and display a projection using the Graph
Spherical Projection command. In the Symbols pane of the Preferences dialog, check Small Circle.
Use the View Data Columns command to display the Alpha column if not visible, and enter the data
from Table 7.1. The resulting Schmidt plot is shown in Figure 7.1, and lower-hemisphere stereographic
projection (stereogram) in Figure 7.2. The three cones intersect at a single point, which represents the
bedding plane orientation. Use the mouse to determine the solution on either projection. The bedding
normal trend, plunge is TP = 120°, 60°, and the plane strike dip is SD = 210°, 30°. While an analytical
solution can be calculated (Ragan, 2009), a visualization of the solution is desirable, and precision to
less than a degree unlikely to be required.

Figure 7.1 Schmidt plot of drill core data from Figure 7.2 Lower hemisphere stereographic
Table 7.1 showing cones defined by the core- projection, or stereogram, of data as in Figure 7.1.
plane angles. The solution is the intersection of Either projection can be used to find the solution.
the three cones.

7.3 Fitting Small Circles

In some cases it may be required to fit directional data to a cone. The most common case in geology is
to determine the best fit axis, λ (a vector, specified, for example, by trend and plunge), and half apex
(semiapical) angle, ψ, of conical folds. The most common model for folds is cylindrical, however the
alternate model of conical folds may be useful in some cases (Wilson, 1967; Gray et al., 1980; Kelker
and Langenberg, 1982, 1988; Mulchrone et al., 2013).

56
This may also be viewed as a general model for fitting vector data to great or small circles, since great
circles are given by ψ = 90°. Note that the data (θ i, φi) are treated as directed (vector) data, if these are
bedding measurements, any overturned beds should be entered using negative dips (θ + 180°, -φ).
The problem, given a data sample (θ i, φi), find the cone (λ, ψ), which can be formulated by minimizing
the sum of either the distance or angular residuals, requires an iterative solution (Mardia and Gadsden,
1977; Gray et al., 1980; Mancktelow, 1981; Fisher et al., 1987; Mulchrone et al., 2013). The fitting
method used here follows Fisher et al. (1987) after Mardia and Gadsden (1977) with modifications
(Vollmer, in preparation).
Figure 7.3 is a Schmidt plot of data from an area of cylindrical folding in Australia (Cohen, 1983, from
Fisher, 1987; example data file Cohen 1983). The fitted cone is (λ, ψ) = (3.23°, 18.38°, 76.19°) given
as azimuth, declination, semiapical angle. Figure 7.4 shows the folded graywacke bedding data with the
best fit small circle (λ, ψ) = (183.63°, 8.09°, 85.57°), very close to the great circle (ψ = 90°) shown in
Figure 4.15.

Figure 7.3 Schmidt plot of 155 facing directions Figure 7.4 Schmidt plot of folded graywacke
of conically folded planes to a small circle. Data bedding data from Figure 4.14 with best fit small
from Cohen (1983), in Fisher et al. (1987). circle. Compare with Figure 4.15, which shows
the best fit great circle.

7.4 Small Circle Confidence

The method for fitting a small circle can be extended to estimate confidence regions by using
bootstrapping methods similar to that discussed in Section 6.3 (Fisher et al., 1987). A large number of
estimates of the mean are calculated from resampled replicates, and used to estimate confidence
regions. Figure 7.5 shows the 95% confidence for the cone axis of the fold data in Figure 7.3 calculated
using 1000 replicates, and Figure 7.6 shows the data rotated to the cone axis, and the means of 1000
replicates.

57
Figure 7.5 Plot of conical fold data as in Figure Figure 7.6 Plot of conical fold data as in Figure
7.3 with 95% bootstrap confidence cone. 7.4, rotated to the best fit cone axis, with means
of 1000 bootstrap replicates used to calculate the
confidence ellipse as in Figure 7.5.

If the Circular option is selected, the confidence cone is calculated using only the angular distance,
producing a circular cone. In this case it is possible to draw confidence rings about the best fit small
circle by checking the Ring option. Figure 7.7 shows the conical fold data with a circular confidence
cone and small circle confidence rings, Figure 7.8 is a plot rotated to the small circle axis.

Figure 7.7 Conical fold data as in Figure 7.3 Figure 7.8 Plot of conical fold data as in Figure
displayed with 95% circular confidence cone and 7.7 rotated to the small circle axis.
small circle confidence rings. The best fit small
circle is (λ, ψ) = (3.23°, 18.38°, 76.19°) with
95% confidence about (λ, ψ) as (3.10°, 2.39°).

58
Figure 7.9 is a Schmidt plot of the folded graywacke beds from Figure 4.14 with 95% circular
confidence cones and small circle confidence rings. The best fit small circle is (λ, ψ) = (183.63°, 8.09°,
85.95°) with 95% confidence about (λ, ψ) estimated as (7.04°, 11.91°). Figure 7.9 shows the same data
rotated to the small circle axis. Note that the axis is calculated with the bedding data as vectors, even if
they are displayed as axes with the Symbols panel Directed option unchecked. From Figure 7.10 it is
clear that one of the 95% confidence includes the plane normal to the axis, that is, ψ (85.95°) plus the
confidence estimate on ψ (11.91°) is greater than 90°. Figure 7.11 shows the same data with the
symmetrical confidence cone for the minimum eigenvector, the standard method for determining the
axis of a cylindrical fold.

Figure 7.9 Schmidt plot of folded graywacke Figure 7.10 Plot as in Figure 7.8 rotated to the
beds from Figure 4.14 with 95% circular fitted small circle axis. Note that confidence rings
confidence cones and small circle confidence on ψ include the plane normal to the axis.
rings. The best fit small circle is (λ, ψ) =
(183.63°, 8.09°, 85.95°) with 95% confidence on
(λ, ψ) estimated as (7.04°, 11.91°).

Due to the long processing time required for the small circle fitting algorithm, bootstrap statistics are
not calculated in the Log window unless either the bootstrap Cone, Hull, or Replicates options are on
(e.g, Figures 7.4 to 7.11). Increasing the number of replicates can cause the calculation to take a very
long time, depending on your processor speed, 1000 resamples should only take a few seconds
however.

59
Figure 7.11 Plot as in Figure 7.9 with the Figure 7.12 Plot as in Figure 11, but using lower
minimum eigenvector, its pole, and 95% circular hemisphere orthographic projection and
confidence region displayed (blue), the standard visualization techniques discussed in Section 5.6.
method for determining a cylindrical fold axis.

7.5 Tutorial 9 - Data Point Weighting

As noted in Section 2.4, individual data points can be weighted. While this applies to all calculations
(except bootstrap confidence cones), including circular plots and kinematic analysis, the fitting of a
small circle gives a good visualization of the effect of weighting a data point.
Make the Weight column visible by using the View Data Columns command if necessary. Select Small
Circle in the Preferences dialog Maxima pane, and enter the following data. The Schmidt plot should
appear as Figure 7.13.

ID Trend Plunge Weight

1 30 60
2 60 20
3 120 60
4 180 60
5 270 60
Table 7.2 Example data to illustrate data point weighting.

60
Figure 7.13 Schmidt plot of data from Table 7.2 Figure 7.14 Data as in Figure 7.6 with a weight
with fitted small circle to illustrate data point of 5 entered for data point 2. Data point 2 now
weighting. counts the same as 5 points, shifting the small
circle towards it.

Enter 5 in the Weight column for data point 2, this data point now counts the same as 5 points, shifting
the small circle towards it (Figure 7.14). Finally, enter 0 for the weight of data point 2, it is now not
included in the calculation and has no effect as shown in Figure 7.15.

Figure 7.15 Example data as in Figure 7.6 with a

weight of 0 entered for data point 2. Data point 2
is now discounted, with no effect on the small
circle calculation.

61
8. Kinematic Analysis
8.1 Introduction
Some data comprise planes that contain lines. These data may indicate movement along faults or shear
directions, such as striae on slickenside surfaces. Other data, such as hinge lines in fold axial planes, or
current directions in bedding planes, also have that characteristic, however this section deals only with
the kinematic analysis of faults and shear zones. Numerous methods exist for the determination of
stresses, strains, or displacements from populations of faults, Orient implements a kinematic analysis
based on M-plane, or movement plane, geometry (Angelier, 1979; Marshak and Mitra, 1988; Twiss
and Unruh, 1998; Marrett and Allmendinger, 1990; Twiss and Moores, 2007).

8.2 Entering Line-Plane Pairs

Kinematic analysis requires entering plane-line pairs, as each data point includes both a plane, such as a
fault plane, and a directed line contained in that plane, such as a fault striation. Data is commonly
entered as strike, dip, trend, plunge, however it is also possible to enter strike, dip and just one of the
three values trend, plunge, or rake. and have Orient calculate the other values. Given the strike and dip
of a plane, only a single additional angle is required to completely specify the orientation of the
contained line.
Three additional columns are provided for kinematic analysis, Sense, Rake, and Error, and an
additional command, Calculate Lines, is provided to calculate missing values, or to recalculate existing
values exactly. Use the View Data Columns command to set these columns visible. The Sense column
is used to indicate if the fault has a normal or reverse component (it is also possible to enter negative
values for the plunge of reverse faults, and add 180° to the trend). The Rake column is for the positive
(clockwise) angle between the strike of the plane and the striation direction, this will be from 0° to
180°, and is positive for faults with a normal component, and negative for faults with a reverse
component.
The Error column is read only, and contains the calculated angular error between the plane and its
contained line. A common problem with field data is that it is over constrained, only three angles are
required for plane-line pairs, while four are often measured. The Calculate Line command is used to
calculate, or recalculate line-plane pairs.
Given the orientation of a plane and one of the angles trend (or other horizontal angle), plunge (or other
vertical angle), or rake (or pitch), the other two can be calculated using the Calculate Line command.
The resulting line-plane pair will have a 0° angular mismatch error.

8.3 Orthonomalization
If both the plane and line have been measured directly, then the line-plane pair is over-constrained with
a mismatch error (Davis et al., 2015). In this case the Error column will contain a non-zero angular
value. The Calculate Line command has two options to recalculate the line-plane pair and remove the
angular error. The Projection option recalculates the line as its projection on the plane. The
Orthonormalize option recalculates both the line and the plane to create a normalized orthogonal frame.
Given over-constrained data, the selection of the Calculate Line option requires judgement of the
source of measurement error. The angular errors should be checked first, any data point with a large

62
error should be discarded or corrected. It is common that the plane is easier to measure accurately than
the contained line, if so, the Projection option is the best. The Orthonormalize option will modify both,
and is better if no assumption is made about the source of error. Alternatively, the line-plane data can be
corrected using the trend, plunge, or rake, possibly taking into account the dip. Note that potential
measurement errors depend on the orientation of the line or plane, for example, the likely error in the
strike of a plane increases with decreasing dip (Ragan, 2009).

8.4 M-Planes
Each data point for kinematic analysis must include the orientation of a kinematic plane, such as the
strike and dip of a fault plane, and the orientation of a movement direction in that plane, such as the
trend and plunge of a fault striation, signifying the motion of the hanging wall (upper) block. Upward
directed vectors, with a reverse component, can be entered using negative plunges, or by entering
Reverse (r) in the Sense column. Note that this is equivalent to adding 180° to the trend, and
multiplying the plunge by -1. Normal (n) may also be entered into the Sense column, but has no effect.
Although the term fault is used here, the kinematic analysis can be applied to shear zones if the
displacement directions can be determined. As the slip lineation lies within the fault plane, the
displacement direction must lie in the plane containing the lineation and the pole to the fault, this plane
is the M-plane. Figure 8.1 shows the geometric relationships with the M-plane shown in gray.

Figure 8.1 Schmidt plot of a normal fault, with a

slip lineation (red), pole to the fault (blue) and the
M-axis, or movement axis, (green), the M-plane is
shown in gray. Directions P and T bisect the angle
between the pole and slip line. Arrows are tangent
lines that show the displacement sense of the
hanging wall with respect to the footwall.

Two additional kinematic axes can be defined within the M-plane, at ±45° from the lineation (Figure
8.1). These are commonly referred to to as the P-axis and T-axis, terms derived from first motion
studies in seismology, however they should not be confused with stress axes. Rather P is an incremental
shortening axis, and T is an extension axis.

63
Figure 8.2 A lower hemisphere equal-area slip- Figure 8.3 A lower hemisphere equal-area slip
linear, or pole tangent line, diagram of a population tangent line diagram of the data shown in Figure
of 38 Neogene normal faults from central Crete, 8.2. The arrows through the slip lineation
Greece. The arrows through the fault pole projections show the displacement sense of the
projections show the displacement sense of the hanging wall with respect to the footwall (data
hanging wall with respect to the footwall (data from from Angelier, 1979).
Angelier, 1979).

The arrows drawn though the lineation and the fault normal are tangent lines, directed lines tangent to a
projected point on a sphere. In Figure 8.1 tangent lines are drawn through the lineation and the fault
pole projections to show the placement sense of the hanging wall with respect to the footwall. They can
be viewed as instantaneous rotation vectors about the M-axis. The term slip-linear has been used for a
tangent line drawn through the fault pole using this reference frame (Marshak and Mitra, 1988). Figure
8.2 shows a slip-linear diagram of a population of 38 Neogene normal faults from central Crete, Greece
(data from Angelier, 1979).
The term tangent-lineation has been used for the tangent line drawn through the fault pole showing the
placement sense of the footwall with respect to the hanging wall (Twiss and Unruh, 1998; Twiss and
Moores, 2007). However, the reference frame used in Figures 8.1 and 8.2 for slip-linears is compatible
with the definition of fault slip as a displacement vector of the hanging wall with respect to the
footwall, and has priority, so is the default setting. Since Orient offers both conventions, the term pole
tangent line (tangent line through the projected fault pole) is used to include both slip-linear and
tangent-lineation reference frames.
A second type of tangent line can be drawn through the slip lineation within in the M-plane. This is
referred to here as a slip tangent line (tangent line through the projected slip line), drawn though the
projected lineation towards, or away from, the pole to the fault (Figure 8.1). This contains the same
information as a pole tangent line. Note that many spherical projections of fault data in the literature
show ticks or arrows parallel to the trend of the slip lineation, which is not the same.

64
8.5 Moment Tensors
The P and T axes (Figure 8.1) indicate shortening and extension axes respectively, and fault kinematic
analysis can therefore done by examining populations of P and T axes independently using moment
tensors (Marrett and Allmendinger, 1990). These tensors are orientation matrixes discussed in Sections
4.7. When fault data is entered, normally a strike and dip for a fault plane, and a trend and plunge for a
slip lineation, Orient automatically generates M-planes, P-axes, and T-axes. An extension is added to
the data type indicate each of the five data elements (Table 8.1).

Data Type Extension Data Element

.S Plane, fault, slickenside, or shear plane
.L Line, slip direction, slickenline, or shear direction
.M M movement plane
.P P shortening axis
.T T extension axis
Table 8.1 Data type extensions automatically generated for kinematic data.

Each of these extended data types can be plotted and analyzed independently. For kinematic analysis
the P and T axes are contoured (Figures 8.5 and 8.6; Section 4.9), and their maximum values and
scatter determined by eigenvector analysis (Section 4.7), and for the construction of beachball plots
(Figure 8.7).

Figure 8.4 Lower hemisphere equal-area projection

of poles to M-planes (green), P shortening axes
(red), and T extension axes (blue) for data in Figure
5.2.

65
Figure 8.5 Schmidt plot with modified Kamb Figure 8.6 Projection as in Figure 8.4 with
contours of the fault data from Figure 5.3 with contours on the T extension axes.
10% density contours on the P shortening axes.

Figure 8.7 Lower hemisphere equal-area

beachball plot showing the P shortening and T
extension quadrants using the data shown in
Figure 8.3.

66
8.6 Tutorial 10 – Kinematic Analysis
Open the file Angelier 1979 from the Example Data folder (any of the csv, tsv. ods, or xlsx versions) in
Orient, and click on the Spherical Projection icon. If no settings have been previously modified the
projection will look as in Figure 8.8. The Help Restore Defaults command can be used to reset the
preferences if desired.
To remove the Schmidt net from the background, click on the Preferences icon, and locate the Net
pane under the Spherical Projection Settings. Uncheck both Axes and Net. Next, in the Labels pane,
change the Increment to -90, the Offset to 14, and the Size to 12.
Next, go to the Symbols pane. Note that there are 5 data types displayed in the Data Type pulldown list.
Select SL.P, check Visible, then select SL.T and check Visible. At this point all data type elements
should be displayed with a symbol as in Figure 8.9.
The first step is to prepare a pole tangent line, or slip-linear, plot. In the Symbols pane uncheck Visible
for all the elements except SL.S so only the poles to the fault planes are displayed. Now go to the
Kinematics pane and check Tangent Lines. The resulting plot is shown in Figure 8.2.

Figure 8.8 Lower hemisphere equal-area scatter Figure 8.9 Projection as in Figure 5.7 after turning
plot of fault data in Figure 5.2 showing fault off display of the Schmidt net, and showing the P
normals (white), slip lines (red), and M-planes shortening (magenta) and T extension (cyan) axes.
(green), with Schmidt net.

Next prepare modified Kamb contour plots of the P shortening and T extension axes. Uncheck Tangent
Lines in the Kinematics pane, and open the Symbols pane. Uncheck Visible for SL.S. For SL.P, check
the following: Visible, Symbol, Contour, and Gradient. Modify the defaults as follows. Set the Symbol
Fill Color to white, and the Symbol Size to 6. Click on the gradient color picker, and select YR (yellow
red) from Preset. Next go to the Contours pane and set Levels to 5 to give modified Kamb contours at
20% density. The resulting plot is shown in Figures 8.10.

67
 Figure 8.11 Projection as in Figure 5.10 with
Figure 8.10 Lower hemisphere equal-area contours of the T extension axes.
modified Kamb contour plot of the fault data
from Figure 5.2 with 20% density contours of the
P shortening axes.

To prepare a contour plot for the T extension data, uncheck Visible for SL.P. For SL.T, check: Visible,
Symbol, Contour, and Gradient. Set the Symbol Fill Color to white, and the Symbol Size to 6. Click on
the gradient color picker, and select CB (cyan blue) from Preset. The resulting plot is shown in Figure
8.11.
Finally, to prepare a beachball plot, uncheck the SL.T options for Contour and Gradient, and set the
Symbol Fill Color to cyan. For SL.P, check Visible, uncheck Contour and Gradient, and set the Symbol
Fill Color to red. In the Kinematics pane check Beachball. The resulting plot is shown in Figure 8.7.

8.7 Confidence Levels, Synoptic Plots, and Weighting

This section explains the use of additional options, including determining confidence cones, using the
same example fault plane data set as in previous sections. Following Section 8.3, note that the line-
plane data should be adjusted to insure that each line-plane pair form an orthogonal frame. This is
easily done by selecting all the data with Edit Select All, and then Data Calculate Line with the
Orthonormalize option. Other options are discussed fully in Section 8.3.
It may be desirable to produce a contoured summary, or synoptic, plot of both the P and T axes. Figure
8.12 is a such a diagram with combined contours as displayed in Figures 8.10 and 8.11. Both the P and
T axes have contours and gradients (Symbols Contour and Symbols Gradient selected), with the
Contours Fill Contours option selected, and Gradient set to Gray. Contours Levels is set to 5 to give
contours at 20%.
Note that the P gradient obscures the T gradient because both are opaque, and the P gradient bitmap is
on top of the T gradient bitmap. To fix this, click on the P Gradient Picker, and set the opacity to 50%.
The result is an overlay as shown in Figure 8.13. The use of overlays can be similarly applied to any
plot with multiple data types, such as bedding and cleavage.

68
Figure 8.12 Schmidt plot of P and T axes as in Figure 8.13 Schmidt plot as in Figure 8.12 with
Figure 8.9 with modified Kamb contours for each the P contour gradient transparency set to 50%, to
at 20%. The Gradient option is selected for each, create a transparent overlay.
but only one is displayed.

The bootstrapping technique discussed in Section 6.3 can be applied to kinematic analysis to derive
confidence intervals about the P and T axis. Figure 8.14 shows the confidence intervals as elliptical
cones about both P and T axes, and Figure 8.15 shows 5000 means of the bootstrap replicants used to
calculate the cones. The confidence values are displayed in the Log window after running the Data
Statistics command. Locate the maximum eigenvectors for the P and T data elements in the log, these
give the 95% confidence radii for the P axis as 7.59°, 6.32°, and for the T axis as 9.87°, 7.99° (Figure
8.14; these will vary randomly a small amount). Points defining endpoints of the long and short ellipse
axes are also given.

Figure 8.14 Schmidt plot as in Figure 8.12 with Figure 8.15 Schmidt plot as in Figure 8.14
95% confidence cones about the P and T axes. displaying the means of 5000 bootstrap replicates
used to calculate the confidence cones.

69
Finally, as discussed in Sections 2.4 and 7.4, individual data points can be weighted by entering a
positive value in the Weight column. The default weight is 1, a weight of 0 will discount the data point.
In kinematic analysis it may be desirable to weight data points, for example by using a weight related
to the estimated area of a fault.
To see the effect of data weighting on kinematic analysis, make the Weight column visible using the
View Data Columns command if necessary, then use the Angelier 1979 data set to create a plot such as
shown in Figure 8.13. Enter weight values, such as 0, 2, or 10 for some of the line-plane pairs to see the
effect on the analysis.

70
9. Orientation Plots
9.1 Introduction
The orientation matrix, orientation tensor, or scatter matrix (Fisher, Lewis, and Embleton, 1987;
Mardia and Jupp, 2000), can be interpreted as a moment tensor associated with the spherical
distribution of axial direction data, by regarding the intersection points of each axis with the sphere as
unit masses. Such a weighted sphere would have three perpendicular moment axes. Section 4.3 showed
how the eigenvectors of this tensor can be used to determine the maximum, intermediate, and minimum
directions. The eigenvectors, however, have relative magnitudes, the eigenvalues, which are treated in
this chapter. Note, for example, that if are three eigenvalues are all equal, there is no preferred direction
and the measurements are randomly, or uniformly, distributed.

9.2 Triangular Orientation Plot

A triangular eigenvalue plot (Point Girdle Random) is particularly useful when a summary diagram is
required of numerous data sets. On the triangular plot a single point summarizes the type of directional
data distribution among point (or cluster), girdle, and random (or uniform) distributions. The plot is
also used in map domain analysis, where domains are defined in the Orientation Map portion of the
program, by maximizing the total cylindricity index.
Given the orientation matrix eigenvectors ε1, ε2, and ε3 for n data points, where the magnitudes ε1 ≥ ε2
≥ ε3, the following are defined (Vollmer 1989):

Point P = (ε1 - ε2)/n

Girdle G = 2(ε2 - ε3)/n
Random R = 3ε3/n
Cylindricity C=P+G

these have the property that:

P+G+R=1
and form the basis of the triangular plot. Cylindrical data sets plot near the top of the graph, along the
P-G join, point distributions plot near the upper left (P), girdle distributions plot near the upper right
(G), and random or uniformly distributed data will plot near the bottom of the graph (R). Figure 9.1 is a
plot of bedding plane poles from a cylindrical fold in Ordovician graywackes (Vollmer, 1981), and
Figure 9.2 is the corresponding triangular plot graph. These indicate a well defined girdle with a
distinct maximum.

71
Figure 9.1 Schmidt plot with modified Kamb Figure 9.2 Triangular orientation plot of data
contours of poles to bedding from a fold in shown in Figure 9.1.
graywackes (data from Vollmer, 1981).

A contour plot of ice fabric c-axes (Kamb, 1959) is shown in Figure 9.3, which shows a much more
scattered distribution, and plots nearer to the bottom of the plot (Figure 9.4). Finally a plot of fold axes
associated with the bedding data in Figure 4.14 is shown in Figure 9.5, and the associated triangular
plot in Figure 9.6 showing a strong point cluster.

Figure 9.3 Schmidt plot with modified Kamb Figure 9.4 Triangular orientation plot of data
contours of ice c-axes (data from Kamb, 1959). shown in Figure 9.3.

72
Figure 9.5 Schmidt plot with modified Kamb Figure 9.6 Triangular orientation plot of fold
contours of minor fold axes associated with the axis data in Figure 9.5 showing strong point
folded bedding data in Figure 4.12. cluster.

73
10. Cluster Analysis
10.1 Introduction
Many directional data samples are multimodal, and, while contouring (Section 4.9) may be sufficient to
locate likely modes (e.g., Schmidt, 1925; Section 4.9), it may be desirable to use numerical techniques
to partition the sample into discrete clusters (e.g., Press et al., 2007). One application is from geological
fracture analysis, in which the orientations of fracture or joint planes are measured, with no basis other
than orientation to use for subdividing the sample. This chapter discusses methods to attempt to
partition such multimodal samples into clusters or cluster domains.
Visual cluster analysis of directional data began with the work of Schmidt (1925), who used contouring
to identify modes, or deviations from uniformity. Shanley and Mahtab (1976) used a Schmidt
projection based counting method to identify clusters. Klose et al. (2005) give a summary of other
work, and present a numerical minimization method. The approach used here differs significantly from
that of Klose et al. (2005), however similarly attempts to locate a solution that minimizes residuals of a
measure of distance from the cluster centers (centroids). The present method allows weighting of data
points, and is implemented for vector, axis, and girdle distributions (Vollmer, in preparation; see Yamaji
and Sati, 2011, for a related method).
The method requires selecting the number of clusters, the cluster count, from k = 2 to 9. This will be
based on user expertise, either through theoretical expectation, or by examination of a contour plot.
Minimization may not converge to a unique solution for distributions that do not have distinct modes,
or where the number of modes is less than the input cluster count. The method selected, Axis, Vector, or
Girdle controls the minimization procedure. Data display for symbols and maxima should be selected
to correspond to the method, and set as either directed or undirected.

10.2 Axis Cluster Partitioning

Figure 10.1 is a Schmidt plot of ice crystallographic data as in Figure 4.17, after cluster partitioning
with two clusters. The two modes initially identified by contouring are clearly selected by the
partitioning method, although due to asymmetry, the axes are not precisely aligned with the contoured
maxima.
The clusters can be visualized more clearly by rotating the projections to the maximum eigenvectors of
the two cluster domains. Figure 10.2 shows the data rotated to the maximum of cluster domain 1
(yellow symbols), and Figure 10.3 shows the data rotated to the maximum of cluster domain 2 (green
symbols). Two distinct maxima are indicated by contouring the cluster domains, and by bootstrap
confidence cones. Figure 10.4 is a triangular orientation plot summarizing the change in fabric type
following partitioning, with the two domains plotting closer to the point, or cluster, vertex than the total
data set (red symbol).
The total data set maximum eigenvector (red symbols in Figures 10.4 and 6.5) as trend, plunge, 95%
confidence is (256.52°, 8.74°, 13.63°), cluster domain 1 (yellow in Figure 10.1) is (58.95°, 36.91°,
10.19°), and cluster domain 2 (green in Figure 10.1) is (261.17°, 28.23°, 8.98°).

74
Figure 10.1 Schmidt plot of ice crystallographic Figure 10.2 Plot as in Figure 10.1, rotated to the
data as in Figure 4.17, after cluster partitioning maximum of cluster domain 1, with contours on
with two clusters. 20% modified Kamb contours cluster domain 1.
on full data set, and 95% symmetric confidence
intervals on the two cluster axes.

Figure 10.3 Plot as in Figure 10.1 rotated to the Figure 10.4 Triangular orientation plot, showing
maximum of cluster domain 2, with contours on the full data set (red symbol), cluster domain 1
cluster domain 2. (yellow), and domain 2 (green). Note that the axis
domain clusters plot closer to the point, or cluster,
vertex.

Figure 10.5 illustrates a similar analysis on the folded graywacke bedding data displayed in Figure
4.14. In this case there is a reasonable expectation to find two modes, corresponding to the two the fold
limbs. Figure 10.6 shows the data rotated to the maximum of cluster domain 1, and Figure 10.7 shows
the data rotated to the maximum of cluster domain 2. As expected, two distinct maxima are indicated

75
by contouring the cluster domains, and by the confidence cones. Figure 10.8 is a triangular orientation
plot summarizing the change in fabric type following partitioning. Note again, that the cluster domains
plot closer to the point, or cluster, vertex than the full data set.
The axis cluster partitioning gives the two clusters (estimated fold limbs) as strike, dip, 95% confidence
equal to (4.10°, 58.79°, 11.84°) for cluster domain 1 (yellow symbols in Figure 10.5), and (180.05°,
23.54°, 6.06°) for cluster domain 2 (green symbols in Figure 10.5).

Figure 10. 5 Schmidt plot of folded graywacke Figure 10.6 Plot as in Figure 10.5 rotated to
bedding data as in Figure 4.14, with 20% cluster domain 1 maximum, with contours on
modified Kamb contours on the full data set, and cluster domain 1.
95% symmetrical confidence intervals on the two
cluster axes.

Figure 10.7 Plot as in Figure 10.5 rotated to Figure 10. 8 Triangular orientation plot, with full
domain 2 cluster axis, with contours on cluster data set (red symbol), cluster domain 1 (yellow),
domain 2. and domain 2 (green). Note that the domains
(estimated fold axes) plot closer to the point, or
cluster, vertex than the full data set.

76
10.3 Vector Cluster Partitioning
An analysis similar to axis cluster partitioning can be done on vector data. Figure 10.9 is a Schmidt plot
of the magnetic remanence data from Figure 1.3, after vector cluster partitioning. Data ploted on the
upper hemisphere are semi-transparent. Figure 10.10 shows the data rotated to the vector mean of
cluster domain 1, and Figure 10.11 shows the data rotated to the vector mean of cluster domain 2.
The two clusters given as [trend, plunge, 95% confidence] are (224.39°, 42.41°, 15.63°) for cluster
domain 1 (yellow symbols in Figure 10.9), and (333.87, -23.13, 7.37) for cluster domain 2 (green
symbols in Figure 10.9).

Figure 10.9 Schmidt plot of magnetic remanence

vectors as in Figure 1.3, after vector cluster
partitioning. 20% modified Kamb contours on the
full data set, and 95% symmetric confidence
intervals on the cluster domain vector means.
Symbols on upper hemisphere are 30%
transparent.

77
Figure 10.10 Plot as in Figure 10.9 rotated to the Figure 10.11 Plot as in Figure 10.9 rotated to the
vector mean of cluster domain 1 (yellow vector mean of cluster domain 2 (green symbols),
symbols), with contours on cluster domain 1. with contours on cluster domain 1

10.4 Girdle Cluster Partitioning

This section concerns girdle cluster partitioning, where an attempt is made to partition more than one
girdle distribution from a sample. A synthetic test example is shown in Figure 10.12, produced by
combining the data from Figure 10.5 with the same data rotated 30° about the Z (vertical) axis . The
best-fit great circles with confidence cones illustrate the effect of the partitioning. To more clearly
visualize the partitioning, Figure 10.10 shows the data rotated to the maximum eigenvector of cluster
domain 1, with contours on domain 1, and Figure 10.11 shows the data rotated to the maximum
eigenvector of cluster domain 2, with contours on domain 2.
The minimum eigenvector (estimated cylindrical fold axis) of the original data as trend, plunge, 95%
confidence was (182.76°, 2.69°, 4.70°) (Figure 6.7). The combined synthetic data after girdle cluster
partitioning gives the domain 1 axis as (187.38°, 3.38°, 1.88°), and domain 2 axis as (153.83°, 2.45°,
4.29°). Rotating domain 2 by -30° about Z gives the (restored) axis (183.8° 2.5°, 4.29°). Note that the
girdle cluster analysis has largely partitioned the combined data into the two original components, a
complete restoration can not be expected. Finally, Figure 10.15 is a visualization of the cluster analysis
results using methods discussed in Section 5.6.

78
Figure 10.12 Girdle partitioning of synthetic data Figure 10.13 Plot as in Figure 10.12 rotated to
produced by combining the data from Figure 10.5 the maximum eigenvector of girdle domain 1,
with the same data rotated 30° about Z. 20% with contours on domain 1.
modified Kamb contours on the full data set.

Figure 10.14 Plot as in Figure 10.12 rotated to Figure 10.15 Plot as in Figure 10.12 using
the maximum eigenvector of girdle domain 2, orthographic projection, rotation, and
with contours on domain 2. transparency as discussed in Section 5.6.

79
11. Orientation Maps
11.1 Introduction
Spherical projections aid in the analysis of the directional data, such as rock foliations, but they do not
display their spacial distribution. Understanding the spacial relationships of directional data is often one
of the primary goals of a geologist, particularly a structural geologist. A number of tools are provided in
Orient to assist with spacial analysis. These include coordinate conversions, integration with mapping
websites, and integration with Google Earth. Orient is designed to plot the spacial distributions of
directional data, and has the capability of calculating orientation data fields that can be used to study
regional tends and to do structural domain analysis. For example, a common problem in mapping areas
of complex geological structure is to identify domains of cylindrical folding. Orient provides unique
capabilities to automatically search for such domains (Vollmer, 1990).

11.2 UTM, Latitude, Longitude, Conversion

Orient includes UTM to latitude, longitude and latitude, longitude to UTM conversions (Snyder, 1987;
Dutch, 2015). Conversion among 14 datums, including WGS 1984 an, NAD 1983, is available. The
Data UTM Conversion dialog (Figure 10.1) can be accessed when there are either easting, northing
coordinates, or latitude, longitude coordinates. Select the conversion, and the desired datum. The
hemisphere can be specified using the UTM grid zone, or by hemisphere.

Figure 11.1 UTM, latitude, longitude

conversion dialog.

11.3 Tutorial 11 – UTM Conversion

Open the file World Earthquakes 1980-1990 Map from the included Example Data folder (any of the
provided file formats, csv, tsv, ods, or xlsx, is fine). Note that the file World Earthquakes 1980-1990
Sphere is formatted to plot the data on a spherical projection, and is not the one for this tutorial. This
data is a set of 14,229 earthquake epicenters between 1980 and 1990 with magnitudes greater than 4.5
(data from NOAA).
Next, open the UTM Conversion dialog (Figure 11.1), select Latitude, Longitude to UTM, WGS 1984,
and UTM latitude zone. Press OK, and the conversion is done. Finally, select View Data Columns and
check Zone, Easting, and Northing if not already checked. UTM grid zones and coordinates should all
be displayed.
Tutorial 12 covers web map integration, but you may wish to select one of the earthquake data rows,
and select Data Show Location Google Maps Satellite. As this is a global data set, you may need to

80
zoom out before determining where you are. Data point ID 49 is an interesting example of a non-plate
boundary seismic event.

11.4 Google and Web Maps

A number of internet web sites offer access to maps, including street maps, topographic maps, and
terrain maps, by entering search terms or geographic coordinates. Google Maps is probably the most
well known of these. Orient has commands to open many of these sites, including Google Maps,
ACME Mapper (which has many topographic maps), Bing Maps, HERE Maps, OpenStreetMaps, and
others in your default browser. If a data file contains latitude and longitude coordinates, these can be
opened directly from Orient. If the data contains UTM coordinates, they must first be converted to
latitude, longitude as covered in Section 11.1. To see the location displayed in your default browser,
select a data row containing latitude, longitude coordinates, and choose a website from the Data Show
Location menu.

11.5 Google Earth

The program Google Earth is an invaluable tool in numerous areas, and has become a widely used tool
in geologic mapping. Google provides an interface using KLM (Keyhole Markup Language) files that
can contain geographic coordinates, viewing instructions, and many other details. Orient uses this
interface to display outcrop locations, and symbols for directional data measurements. Symbols are
selected from the Preference dialog Orientation Map Settings Symbols pane where various symbols
can be selected, as well as color, size, and line width. The symbol size is set in meters in the KML
Length edit box, and the width in the KML Width edit box. To save the file, use the File Export to
Google Earth KML command. Double click the resulting file to open in Google Earth.

11.6 Tutorial 12 – Web and Google Earth

This tutorial covers internet web maps, such as Google Maps, as well as Google Earth integration.
Open the file Sky Top from the included Example Data folder (any of the provided file formats is fine,
csv, tsv, ods, or xlsx). This is a demonstration file with simulated geologic data for use in this tutorial.
You may wish to use the View Data Columns command to close some of the unused columns.
The file contains the latitude, longitude coordinates of one marker location, and eight simulated data
points of four different types of geologic data, bedding, cleavage, joints, and slickenside lineations. For
web map access, select one of the locations and choose a websites from the Data Show Location menu.
The selected web map will display the location of the data point, you can then zoom in, or change
display options as needed.
In order to display directional data symbols, as well as location, Orient writes KML files for Google
Earth, so directional symbols and complete data sets can be viewed. Open the Preferences dialog by
clicking on its icon, select the Orientation Map Settings option, and the Symbols pane. Select the data
types from the pulldown menu, and check just one symbol for each data type. For J, check Strike,
select a strike line with two ticks, and set Stroke Color to light green. For L, check Line, select an arrow
symbol, and set Stroke Color to magenta. For S0 check Strike, select a strike line symbol, and set
Stroke Color to red. For S1, check Strike, select a strike line with a tooth, and set Stroke Color to red.
Press OK when done.

81
Save the file as Sky Top.klm using the File Export to Google Earth KML command. Double click on it to
open in Google Earth. The location (that I can see from the window in my university office) should
come into view with the data symbols (Figure 11.2). For a quicker demonstration, there is a copy of the
resulting file in the Example Data folder.

Figure 11.2 Example data from the file Sky Top viewed in Google Earth. Example data types are S0 (bedding, red), S1
(cleavage, blue), J (joints, green), and a lineation (L, magenta).

11.7 Orientation Fields and Domain Analysis

Spherical projections aid in the analysis of directional data, but they do not show their spacial
distribution. Therefore Orient provides tools for analyzing the spacial variation of directional data,
including orientation fields and structural domain analysis. Orientation fields are a way to look for
regional patterns in data by defining eigenfoliations and eigenlineations on a grid over the area
(Vollmer, 1990). One method is to define discrete subdomains that include all data within them, for
example a one kilometer square area. A second method is to apply a weighting function to produce
area-smoothed orientation values. The fields are defined by calculating an orientation matrix at grid
locations, and using its eigenvectors and eigenvalues to characterize the fabric.
For example, a common problem in mapping areas of complex geologic structure is to identify
cylindrical domains within the map area (Ramsay, 1967). Figure 11.3 is an equal-area lower

82
hemisphere projection of poles to foliations from the Doverfjell mountains, Norway, and Figure 7.4 is a
modified Kamb contour plot of the data. Their spacial distribution is shown in Figure 7.5, plotted using
the Orientation Map command.

Figure 11.3 Lower hemisphere equal-area Figure 11.4 Modified Kamb contour plot of the
projection of poles to 625 foliation planes from data shown in Figure 7.3 with contours at 20%
the Doverfjell Mountains, Norway (data from density.
Vollmer, 1985).

Figure 11.6 Subdomain orientation field of

Figure 11.5 Orientation map of foliation
eigenfoliation strikes generated from the data
strikes of the data shown in Figure 7.3.
shown in Figure 7.3. The strike tick marks
represent the horizontal projection of the dip
line.

83
 Figure 11.7 Weighted orientation field of
eigenfoliation dip lines generated from the data
shown in Figure 7.3. The arrows represent the
horizontal projections of the dip line.

The geometry of the structure defined by the foliations is not obvious. Figure 11.6 is a subdomain
orientation field defined on a one kilometer grid. Note that the number of data points within each
subdomain varies, and that each has a discrete boundary. Figure 11.7 is a weighted orientation field
where each grid node is generated as a weighted sum of all other data points. This map shows the
horizontal projection of the dip lines, with steeper dips displayed as shorter lines.
Structural domain analysis (Vollmer, 1990) is done by attempting to maximize a quantity, or index,
related to the given problem. Orient provides several indexes that may be maximized, including point,
girdle, and cylindricity indexes. To locate areas of cylindrical folding the cylindricity index is
maximized:
C = (ε1 + ε2 – 2ε3) / n
Where ε is the orientation matrix eigenvalue for n data points, and ε1 ≥ ε2 ≥ 2ε3.
For a set of domains the sum of the products of the domain indexes (C1, C2, C3, ...) and the number of
data points within each domain (n 1, n2, n3, ...):
Z = C1n1 + C2n2 + C3n3 + ...
is maximized. Because:
n = n1 + n2 + n3 + ...
the maximum possible value for Z is equal to n. The normalized sum is:
C' = Z/n

84
The settings for the Orientation Map command are in the Preferences dialog Orientation Map panels.
For each data point, or field value, it is possible to plot a symbol in four directions, Line, Dip Line,
Strike, and Strike 180 (Figure 11.8).

Figure 11.9 The

Domain Search dialog
box.

Figure 11.8 Orient Preference dialog showing the

Orientation Map Symbols pane. The options shown are to
project strike lines as in Figure 10.6.

Line is the direction of a line, or of the plane normal (θ). Dip Line is the opposing direction (θ + 180°),
and used to display a plane's dip direction. Strike is the plane strike (θ + 90°), and Strike 180 is the
opposing direction (θ – 90°). Only Line and Dip line can be projected. To display a projected strike
symbol, select Dip Line and the strike symbol shaped like a sideways T (Figures 11.6 and 11.8). The
Length Scale scales the length of the symbol. The KLM settings are for the display of data in Google
Earth (Section 11.3).
To conduct a domain search, define the extents of the map area in the Extents panel, select the
Subdomain Method in the Field panel, along with the number of subdomains. Then use the Domain
Search command, which displays the dialog shown in Figure 11.9. This is used to maximize the
orientation index using an automated search process. The Value is the current value of the index, and N
is the number of points used to calculate it. Domains are numbered from 1 to 9, zero represents an
unassigned domain. Clear resets all subdomains to zero. Initialize sets all domains to the search domain.
Search will attempt to grow the search domain. The search process iteratively checks to see if
subtracting a subdomain from one domain and adding it to the search domain will increase the index. It
can only do so if both domains remain connected.
The user can manually edit domains, moving the mouse over the map will display subdomain
information in the status bar, and clicking on a subdomain will add it to the search domain if possible.
The constraint that the domains remain connected can cause edge effects, and vacant subdomains may
also effect connectivity. The general procedure is to do a search on a desired number of domains, then
manually edit them, and search again to see if a better solution is found.
Once the search is completed, pressing OK will assign domains to each of the data points. The domains
are added as attributes of the data, use the View Data Columns to view the Domain column. Extended
data types are formed by appending domain extensions, D1 to D9, to the data type so each domain can
be plotted independently, for example on a spherical projection. To update the results during a search,

85
click on the Apply button. This will apply the domains to the data and will update other graphs that are
open. If a spherical projection of the data is open it is possible to color code the data by domain to get a
visualization of the distribution.

10.8 Tutorial 13 – Domain Analysis

Open the file Vollmer 1985 included Example Data folder (any of the provided file formats). This is the
data displayed in Figures 11.3 to 11.7, 625 foliation planes from the Doverfjell mountains, Norway.
While folds are clearly present, the spherical projections (Figure 11.3 and 11.4) do not display a girdle
pattern that would indicate cylindrical folding (there are spectacular sheath folds in the Doverfjell and
adjacent Trollheimen ranges). The map of foliations (Figure 11.5) shows some areas of consistent
orientation, but the location of cylindrical domains is not obvious.
Click on the Orientation Map icon to display a map of the foliation data, then in the Orientation Map
window, click on the Preferences icon. In the Orientation Map Extents pane, uncheck Auto Scale, and
enter the 890 for the minimum X (Easting) coordinate, 150, for the minimum Y (Northing) coordinate,
1010 for the maximum X coordinate, and 300 for the maximum Y coordinate. To view the data,
uncheck all symbols in the Maxima pane, and check only Strike in the Symbols pane. The result should
appear as in Figure 11.5.
To set up the domain search, uncheck Strike in the Symbols pane, and check Strike in the Maxima pane.
In the Field pane the Method should be Subdomain, and Directed should be unchecked. Enter 12 and
15 for the number of X and Y subdomains respectively. The coordinates are UTM based in meters, so
this gives one square kilometer subdomains. The resulting map should be as in similar to Figure 11.6,
except that the strikes are not projected. Press OK when done.
To begin the domain search, select the Graph Domain Search command from the menu (Figure 11.9).
Press Initialize to set all subdomains to 1, then press Search to grow domain 2. Change the search
domain to 3, and press Search again. The result is should in Figure 11.10, which has a cylindricity
index, c = 0.738. There are some edge effects, particularly where domain 1 wraps around the other two.
Move the mouse over the map to see information about each subdomain. Clicking on one will change it
to the current search domain, if all domains remain connected.
Edit the domains by selecting a search domain and clicking on the map. When done, go back to the
dialog and search on each domain again. If the domains are stable, they will not change. Figure 11.11
shows a stable solution with C = 0.851, an improvement over the previous value. Press OK when done.

86
Figure 11.10 Initial automatic domain Figure 11.11 Final domain configuration
search formed by grouping the subdomains after iterative manual editing and automatic
of Figure 7.6 into three domains searching to find a stable configuration. This
maximizing cylindricity. This configuration configuration has a cylindricity index, C' =
has a cylindricity index, C' = 0.738. 0.851.

The map will now be cluttered with symbols, as the three new domains are plotted in addition to the
field. Before cleaning up the map, open a spherical projection by clicking on the Spherical Projection
icon in the data spreadsheet window, and then the Preferences dialog. In the Spherical Projection Net
pane, uncheck Axes and Net, and in the Labels pane set Increment to -90, Offset to 14 and Size to 12,
as in previous tutorials.
In the Symbols pane, uncheck Symbol the for data type S. Then for data types S.D1, S.D2, and S.D3,
the new data types, set the Symbol Fill Color to red, green, and blue respectively. In the Maxima pane,
select only the Minimum Eigenvector for S.D1, S.D2, and S.D3, and set the fill and stroke colors the
same. The result should be as in Figure 11.12. Note that the three domains show well defined girdles.
Turning off the data symbols give Figure 11.13, which suggests refolding of earlier folds about a
northwest plunging axis.
Next, open a triangular orientation (Point Girdle Random) plot using the PGR Plot icon in the data
spreadsheet window. In the Preferences dialog PGR Plot Symbols pane set the Symbol Fill Color for S
to white, and then for S.D1, S.D2, and S.D3, to red, green, and blue respectively, and increase the
symbol sizes to 16. The resulting triangular plot (Figure 7.14) shows the relative changes in cylindricity
from the whole area to the three domains. Note that domain 2 (green) has the strongest point
distribution, and domain 3 (blue) has the strongest girdle distribution.

87
Figure 11.12 Lower hemisphere equal-area Figure 11.13 Synoptic plot of best fit girdles and
projection of data from Figure 7.3, with foliation axes of data shown in Figure 7.12.
poles color-coded by domain, and the great circle
normal to the minimum eigenvector drawn for each
domain.

Figure 11.14 Triangular orientation plot of the three

domains compared to the total data set (white).
Figure 11.15 Data from Figure 7.5 color-
coded by domain.

To clean up the map, turn off the display of the maxima in the Orientation Map Maxima pane, turn off
display of data type S symbols, and set Strike on with Stroke Color for S.D1, S.D2, and S.D3, to red,
green, and blue respectively, giving the map in Figure 11.15.

88
As a final step, prepare lower hemisphere equal-area modified Kamb contour plots (Section 4.9) of the
three domains (Figures 11.16, 11.17, and 11.18).

Figure 11.16 Lower hemisphere equal-area Figure 11.17 Projection as in Figure 11.16 for
midified Kamb contour plot of poles to domain 2 (green domain in Figures 11.10 to
foliation for domain 1 (red domain in Figures 11.15)
11.10 to 11.15) contoured at 20% .

Figure 11.18 Projection as in Figure 11.16 for

domain 3 (blue domain in Figures 11.10 to
11.15) .

89
Acknowledgements
An early draft of this document was prepared for Teaching Structural Geology, Geophysics, and
Tectonics in the 21st Century, On the Cutting Edge, held July 15-19, 2012 at the University of
Tennessee, Knoxville, which rekindled my enthusiasm for completing this new version of Orient. I
thank the conveners Barbara Tewksbury, Gregory Baker, William Dunne, Kip Hodges, Paul Karabinos,
and Michael Wysession. I thank Eric Erslev, Steven Wojtal, Haakon Fossen, Christian D. Klose, Josh
Davis, Sarah Titus, Yvette Kueiper, Peter Hudleston, Robert Bauer, Mark Brandon, and Dazhi Jiang for
discussions at various times.
Thanks to Emily Lubicich (City College of New York), Chistopher Gahn (Vassar College), Ben
Michael Frieman (Colorado School of Mines), Haakon Fossen (University of Bergen), Josh Davis
(Carlton College), Tammy Xinran He (Colorado School of Mines), Steven Coombes (Rio Tinto
Exploration, Canada), and Andrea Bistacchi (University of Milano-Bicocca) for suggestions and bug
reports.

References
Angelier, J., 1979. Determination of the mean principal directions of stresses for a given fault
population. Tectonophysics, v. 56, p. T17-T26.
Badgley, P.C., 1959. Structural methods for the exploration geologist. Harper and Brothers, New York,
280 pp.
Barnes, J.W., 1995. Basic geologic mapping, 3rd Edition. Open University Press, Milton Keynes, UK.
Billings, M.P., 1942. Structural geology. Prentice-Hall, New York, 473 pp.
Billings, M.P., 1954. Structural geology, 2nd edition. Prentice-Hall, New York, 514 pp.
Bingham, C., 1974. An antipodally symmetric distribution on the sphere. Ann. Stats., v. 2, p. 1201-
1225.
Borradaile, G.J., 2003. Statistics of earth science data: their distribution in time, space, and orientation.
Springer, Berlin.
Bucher, W.H., 1944. The stereographic projection, a handy tool for the practical geologist. Journal of
Geology, v. 52, n.3, p. 191-212.
Cheeney, R.F., 1983. Statistical methods in geology. George Allen & Unwin, London. 169 p.
Davis, J.C., 1986, Statistics and data analysis in geology. Wiley, 646 pp.
Davis, J.R., Titus, S.J., and Tikoff, B., 2105. Statistical treatment of structural geology data. Geological
Society of America Short Course Notes.
De Aguilon, François, 1613. Opticorum libri sex philosophis juxta ac mathematicis utiles (Six books of
optics, useful for philosophers and mathematicians alike). Anvers.
De Sitter, L.U, 1956. Structural Geology. McGraw-Hill Book Company, Inc., London, New York,
Toronto. 552 p.
Downs, T.D., 1972. Orientation statistics. Biometrika, v. 59, n. 3, p. 665-676.
Dutch, S., 2015. Converting UTM to latitude and longitude (or vice versa).
www.uwgb.edu/dutchs/usefuldata/utmformulas.htm (accessed 27 April 2015).
Diggle, P.J., and Fisher, N.I., 1985, Sphere: a contouring program for spherical data. Computers &
Geoscience, v. 11, p. 725-766.
Donn, W.L., and Shimer, J.A., 1958. Graphic methods in structural geology. Appleton-Century-Crofts,
Inc., New York, 180 p.

90
Fisher, N.I., Lewis, T., and Embleton, B.J., 1987. Statistical analysis of spherical data (errata, 2000).
Cambridge University Press, Cambridge, 329 pp.
Fosson, H., 2016. Structural geology, 2nd Edition. Cambridge University Press, Cambridge, 463 pp.
Gray, N., Geiser, P., Geiser, J., 1980. On the least-squares fit of small and great circles to spherically
projected orientation data. Mathematical Geology 12, 173–184.
Haughton, S., 1856. On slaty cleavage, and the distortion of fossils. Philosophical Magazine, 12, p.
409-421.
Hills, E.S., 1963. Elements of structural geology. Chapman & Hall, London, 483 p.
Hobbs, B.E., Means, W.D., and Williams, P.F., 1976. An outline of structural geology. Wiley, New
York, 571 pp.
Howarth, R.J., 1999. Measurement, portrayal and analysis of orientation data and the origins of early
modern structural geology (1670-1967). Proceedings of the Geologists' Association, v. 110, p. 273-
309.
Kamb, W.B., 1959. Ice petrofabric observations from Blue Glacier, Washington, in relation to theory
and experiment. Journal Geophysical Research, v. 64, p. 1891-1909.
Kelker, D., and Langenberg, C. W., 1982. A mathematical model for orientation data from macroscopic
conical folds. Mathematical Geology, v. 14, n. 4, p. 289-307.
Kelker, D., Langenberg, C.W., 1988. Statistical classification of macroscopic folds as cylindrical,
circular conical, or elliptical conical. Mathematical Geology 20, 717–730.
Kent, J.T., 1982. The Fisher-Bingham distribution on the sphere. Journal of the Royal Statistical
Society B, v. 44, n. 1, p. 71-80.
Klose, C.D., Seo, S., and Obermayer, K, 2005. A new clustering approach for partitioning directional
data. International Journal of Rock Mechanics & Mining Sciences v. 42, p. 315-321.
Knopf, E.B., and Ingerson, E., 1938. Structural petrology. Geological Socity of America Memoir 6, 270
p.
Lambert, F.H., 1772. Anmerkungen und zusätze zur entwerfung der land und himmelscharten. Verlag
von Wilhelm Engelmann, 1894, Leipzig.
Lisle, R.J. and Leyshorn, P.R., 2004. Stereographic projection techniques for geologists and civil
engineers, 2nd edition. Cambridge University Press, Cambridge, 112 pp.
Mancktelow, N.S., 1981. A least-squares method for determining the best-fit point maxima, great circle,
and small circle to nondirectional orientation data. Mathematical Geology, v. 13, p. 507-521.
Mardia, K.V., 1972. Statistics of directional data. Academic Press.
Mardia, K.V., Gadsden, R.J., 1977. A circle of best fit for spherical data and areas of volcanism.
Applied Statistics 26, 238–245.
Mardia, K.V., and Jupp, P.E., 2000. Directional statistics. John Wiley & Sons, Ltd., Chichester, UK.
430 p.
Mardia, K.V., and Zemroch, P.J., 1977. Table of maximum likelihood estimates for the Bingham
distribution. Journal of Statistical Computation and Simulation, v. 6, p. 29-34.
Marrett, R., and Allmendinger, R.W., 1990. Kinematic analysis of fault-slip data. Journal of Structural
Geology, v. 12, p. 973-986.
Marshak, S., and Mitra, G., 1988. Basic methods of structural geology. Prentice Hall, 446 p.
Mulchrone, K.F., Pastor-Galán, D., Gutiérrez-Alonso, 2013. Mathematica code for least-squares cone
fitting and equal-area stereonet representation. Computers & Geoscience, v. 54, p. 203-210.
Parker, J.M., 1942. Regional systematic jointing in slightly deformed sedimentary rocks. Geological
Society of America Bulletin, v. 53, p. 381-408.
Phillips, F.C., 1954. The use of stereographic projection in structural geology. Edward Arnold, London,
86 pp.

91
Pollard, D.D. and Fletcher, R.C., 2005. Fundamentals of structural geology. Cambridge University
Press, Cambridge, 463 500 pp.
Press, W.H, Teukolsky, S.A., Vetterling, W.T., and Flannery, B.P., 2007. Numerical Recipes: The Art of
Scientific Computing, 3rd edition. Cambridge University Press, Cambridge, 1235 p.
Ragan, 2009. Structural geology: an introduction to geometrical techniques, 4th edition. Cambridge
University Press, Cambridge. 602 pp.
Ramsay, J.G., and Huber, M.I., 1987. The techniques of modern structural geology, volume 2: folds and
fractures. Academic Press, 700 pp.
Sander, B., 1970. An introduction to the study of fabrics of geological bodies, 1st English edition.
Pergamon Press, Oxford. 641 p. [Translated from Sander, 1948, 1950. German edition. Springer-
Verlag.
Schmidt W., 1925. Gefügestatistik. Tschermaks Mineralogische und Petrographische Mitteilungen 38,
p. 392–423.
Shanley R.J., and Mahtab, M.A., 1976. Delineation and analysis of clusters in orientation data.
Mathematical Geology, v. 8, n. 1, p. 9–23.
Snyder, J. P., 1987. Map projections: a working manual. United States Geological Survey Professional
Paper 1395, 383 p.
Swan, A.R.N., and Sandilands, M., 1995. Introduction to geological data analysis. Blackwell Science,
Oxford, 446 p.
Swinbank, Richard, and Purser, R. James, 2006. Fibonacci grids: a novel approach to global modeling.
Quarterly Journal of the Royal Meteorological Society, v. 132, n. 619, p. 1769-1793.
Turner, F.J. and Weiss, L.E., 1963. Structural analysis of metamorphic tectonites. McGraw-Hill Book
Company, New York, 545 pp.
Twiss, R.J., 1990. Curved slickenfibers; a new brittle shear sense indicator with application to a sheared
serpentinite. Journal of Structural Geology, v. 11, p. 471-481.
Twiss, R.J., and Unruh, J.R., 1998. Analysis of fault slip inversions: do they constrain stress or strain
rate? Journal of Geophysical Research,. v. 103, p. 12205-12222.
Twiss, R.J. and Moores, 2007. Structural geology, 2nd edition. W.H. Freeman, New York, 736 pp.
Van der Pluijm, B.A. and Marshak, S., 2004. Earth structure, 2nd edition. W.W. Norton, New York, 656
p.
Vollmer, F.W., 1981. Structural studies of the Ordovician flysch and melange in Albany County, New
York: M.S. Thesis, State University of New York at Albany, Advisor W.D. Means, 151 p.
Vollmer, F.W., 1985. A structural study of the Grovudal fold-nappe, western Norway. Ph.D. Thesis,
University of Minnesota, Minneapolis, Advisor P.J. Hudleston, 233 p.
Vollmer, F.W., 1988. A computer model of sheath-nappes formed during crustal shear in the Western
Gneiss Region, central Norwegian Caledonides. Journal of Structural Geology, v. 10, p. 735-743.
Vollmer, F.W., 1989. A triangular fabric plot with applications for structural analysis (abstract). Eos, v.
70, p. 463.
Vollmer, F.W., 1990. An application of eigenvalue methods to structural domain analysis. Geological
Society of America Bulletin, v. 102, p. 786-791.
Vollmer, F.W., 1993. A modified Kamb method for contouring spherical orientation data. Geological
Society of America Abstracts with Programs, v. 25, p. 170.
Vollmer, F.W., 1995. C program for automatic contouring of spherical orientation data using a modified
Kamb method: Computers & Geosciences, v. 21, p. 31-49.
Vollmer, F.W., 2011. Automatic contouring of two-dimensional finite strain data on the unit
hyperboloid and the use of hyperboloidal stereographic, equal-area and other projections for strain
analysis. Geological Society of America Abstracts with Programs, v. 43, n. 5, p. 605.

92
Vollmer, F.W., 2015. Orient 3: a new integrated software program for orientation data analysis,
kinematic analysis, spherical projections, and Schmidt plots. Geological Society of America
Abstracts with Programs, v. 47, n. 7, p. 0.
Vollmer, F.W., 2015. Orient 3 User Manual www.frederickvollmer.com. [this document]
Vollmer, F.W., 2015. Orient 3: Spherical projection and orientation data analysis program.
www.frederickvollmer.com.
Watson, G.S., 1966. The statistics of orientation data. The Journal of Geology, v. 74, n. 5, p. 786-797.
Whitten, E.H.T., 1966. Structural geology of folded rocks. Rand McNally, Chicago, 663 pp.
Wilson, G., 1967. The geometry of cylindrical and conical folds. Proceedings of the Geologists'
Association, 1967, v. 78, part 1, p. 179-209.
Yamaji, A., 2008. Theories of strain analysis from shape fabrics: A perspective using hyperbolic
geometry. Journal of Structural Geology, v. 30, p. 1451-1465.
Yamaji, A., and Sato, 2011. Clustering of fracture orientations using a mixed Bingham distribution and
its application to paleostress analysis from dike or vein orientations. Journal of Structural Geology,
v. 33, p, 1148-1157.

93
History
3.4.2.3 (2016-06-11)
• Now allows a minimum spherical projection contour level of 0.
• Added automatic range setting for spherical projection contour grid, now the gradient automatically scales to
full grid range by default.
• Added rescaling of spherical projection contour grid to expected uniform density to contour in multiples of
uniform density.
• Added gradient scale bar option to spherical projection.
• Added WBGYR gradient preset.
• Added an option to place spherical net above the gradient or beachball bitmaps.
• Fixed rotation of spherical projection tick marks.

3.4.1.3 (2016-03-25)
• Fixed orientation map subdomain maxima symbol drawing bug.
• Fixed orientation map subdomain maxima symbol count bug.
• Centered orientation map on page.
• Changed orientation map status display from one to two decimal places.
• Moved orientation map maxima symbols above data symbols.
• Fixed reading and display of explicit NaN values and, therefore, display of polylines (such as the Modern
Continental Outlines example data).
• Fixed domain search bug failure to assign domains when multiple files were open.
• Added Apply button to Domain Search dialog.
• The visibility of symbols on the orientation map is no longer linked to their visibility on the spherical
projection.
• Domain search subdomains are now saved and optionally displayed after Domain Search dialog is closed.
• Added an option in the Preferences dialog Orientation Map pane to show or hide the subdomains when
Domain Search dialog is closed.

3.4.0.16 (2016-03-20)
• Bound Windows and Linux resources into executable.
• Optimized with FPC 3.0.0 and LCL 1.6.0.
• Improved memory management by freeing unused forms.
• Added View Image Window command, moved location of File Open Image command.
• Replaced Timers with IdleTimers for better performance.
• Settings for Rotate Data, Rotate Projection, and UTM Conversion are now stored between sessions.
• Consolidated raster and vector image saving under Export Image As command.
• Improvements to digitize commands.
• Added Kent distribution confidence cones.
• Added Kent statistics to Log output.
• Added Bingham statistics to Log output.
• Data input behavior now more spreadsheet compatible. Pressing Enter to accept input moves down, additional
rows are always available for input, fixed bug causing a scroll up when on last data line.

3.3.3.11 (2016-02-03)
• Fixed Watson confidence cone weighting error, was dividing by count instead of weighted count.
• Added process cancelation code for lengthy operations in bootstrap routines.
• Added progress display status for lengthy operations in bootstrap routines.
• The preferences dialog now displays only the data types available in the current file, instead of all files.
• The spherical projection settings are now saved by file name, instead of globally.

94
• Only graphs of the current data are now redrawn, instead of all graphs.
• When selecting a plot window the associated file in the spreadsheet is now displayed.

3.3.2.9 (2016-01-24)
• Fixed Page Size dialog display failure.
• Fixed Replicates symbol picker display.
• Fixed several User Manual images in Chapter 6.
• Added Watson confidence cone data to Log.
• Watson confidence cones now rotate on projection rotation.
• Bingham confidence cones now rotate on projection rotation.
• Fixed preferences initialization bug.
• Preferences dialog now reopens on last panel.
• Preferences dialog now reopens with last data type selected.
• Weighting now works with Fisher confidence cones.
• Weighting now works with Watson confidence cones.
• Added translucent display of data polygons on hidden hemisphere.
• Fixed divide by zero bug.
• Fixed additional Windows and Linux control display issues.
• No longer asks twice on quitting with modified file.
• Added translucent display of tangent lines on hidden hemisphere.
• Status bar now indicates if a file is modified.
• Can now display kinematic beachball and tangent lines at same time.

3.3.1.13 (2016-01-17)
• Fixed some dialog control display issues in Linux.
• Added translucent display of contour lines on hidden hemisphere.
• Added translucent display of confidence ellipses on hidden hemisphere.
• Fixed XML creator to correctly translate [&”<>] characters in file paths. Project files are validated XML.
• Project files now save open file paths.
• Added commands to save and open project files.
• Fixed bug causing crash when closing a tabbed window.
• Fixed bug that was naming “untitled” “untitled 1”.
• Fixed some tab orders and control spacing.
• Modified status bar display to include calculated angle for two selected data points. Clicking on the status bar
toggles between line and plane units.

3.3.0.8 (2016-01-11)
• Corrected the name of the Calculate Line Orthorectify option to Orthonormalize.
• Changed Orthonormalize algorithm from QR decomposition to polar decomposition, to give the closest
orthogonal frame. The difference should be minimal.
• Fixed sorting in Weight, Error, Rake, Sense, and Alpha columns.
• Tweaked the tangent-normal wrapping algorithm used in bootstrap covariance calculation. The difference
should be minimal.
• Plot windows now save their placement on the screen based on name.
• Added Cluster Partitioning, see User Manual Chapter 10.
• Added hidden small circle display.

3.2.1.4 (2016-01-02)
• Fixed numerous control display issues in Linux.
• Cleaned up control alignments in Preferences dialog.
• Fixed Beachball to rotate with projection.

95
• Modified small circle bootstrapping work with psi close to 90.
• Added small circle bootstrap estimation of confidence in psi.
• Added drawing of small circle bootstrap confidence rings about psi.
• Added Fibonacci Sphere command.
• Added opacity options to plot symbols, rays, and great circles on hidden hemisphere.
• Numerous updates to the user manual.

3.2.0.7 (2015-12-29)
• Added best-fit small circles.
• Modified labels in Columns Dialog to show current data angles, e.g., Strike.
• Fixed control spacing in Linux Calculate Lines dialog.
• Rotate Projection dialog now includes Vector Mean and Small Circle Pole.
• Added circular confidence cones for bootstrap means.
• Added Alpha field and plotting of small circle data.
• Added option to plot hidden directed data, on opposing hemisphere, with an unfilled symbol.
• Added Help Orient Home Page command.
• Fixed NaN entry and polyline markers, that got broken in 3.1.0.
• Cleaned up Log file, added Fisher statistics.

3.1.0.27 (2015-12-14)
• Added Sense column for kinematic analysis to allow undirected (downward) lineation entry, previously it was
necessary to input directed (upward or downward) fault lineations.
• Fixed weighting error in orientation field.
• Fixed orientation map page size setting.
• Added Weight column for data point weighting, 1 is the default value, 2 counts a data point twice, etc..
• Decreased non-splined points to 91 in spherical net small circles. Use Restore defaults to implement.
• Fixed duplicate drawing of spherical net small circles.
• Spherical net great circles in SVG are now computed as cubic Bezier paths.
• Added Windows system tray icon
• Added encapsulated PostScript (EPS) vector graphics export.
• Added AutoCAD drawing exchange format (DXF) vector graphics export.
• Added hexagon and pentagon symbols.
• Fixed arrowheads in kinematic analysis plots.
• Clarified directional versus orientation data definitions.
• Added Rake column for the rake of a line in the containing plane.
• Added Error column giving the angular error between a line and the containing plane. Read only.
• Fixed data entry to not delete invalid entry over existing value.
• Fixed several data entry display issues.
• Added Calculate Line command to calculate line in plane angles from projection, trend, plunge or rake.
• Rotate Data command now rotates only selected items.
• Modified status bar to display maximum and minimum eigenvectors for two or more data points.
• Fixed command clicking on plots to update multiple selections in the spreadsheet.
• Removed Yahoo! Maps linking, they shut down this service.
• Fixed bug setting X and Y fields when locating points.
• Added orthorectification option to Calculate Line command.
• Added bootstrapping analysis with elliptical confidence cones and polygon regions (hulls).
• Fixed Log Window Copy command.

3.0.2.1 (2015-09-28)
• Fixed Preferences Dialog list spacing.
• Optimized contour line segment joining for vector export.

96
• Optimized data grid scrolling.
• Fixed Linux About display.
• Optimized messaging.
• Fixed crash when non-Roman Unicode characters are in user name. A workaround is to run Orient from a
thumb drive with only Roman characters in file and folder names.
• Work on User Manual.

3.0.1.0 (2015-05-30)
• Fixed bug importing dip direction.

3.0.0.77 (2015-05-01)
• First release of Orient 3. Complete rewrite with numerous new features. Compiled, tested, and debugged on
Macintosh, Windows, and Linux.

2.1.2 (2012-10-31)
• Last release of Orient 2.

2.0.0.7 (2006-07-23)
• First release of Orient 2. Complete rewrite with Macintosh, Windows, and Linux versions.

1.6.1 (1995-02-28)
• Last release of Orient 1.

1.0.0 (1986)
• First release of Orient 1, Microsoft DOS. Introduced modified Kamb contouring, triangular fabric plots, and
automated domain analysis (see Vollmer, 1985, 1988, 1990, 1993, 1995). Available by download (modem)
from COGSnet, sponsored by the Computer Oriented Geological Society, Denver, CO.

97
This page intentionally left blank

98
99