Function ordirgl displays three-dimensional dynamic
ordination graphs which can be rotated and zoomed. This function
works with all ordination results from vegan and all
ordination results known by the scores function. The
orgl-prefixed functions add elements to the ordirgl
graph similarly as ordi-prefixed functions in vegan.
ordirgl(object, display = "sites", choices = 1:3, type = "p", col = "black",
ax.col = "red", arr.col = "yellow", radius, text, envfit, ...)
orglpoints(object, display = "sites", choices = 1:3, radius, col = "black", ...)
orgltext(object, text, display = "sites", choices = 1:3, adj = 0.5,
col = "black", ...)
orglsegments(object, groups, order.by, display = "sites", choices = 1:3,
col = "black", ...)
orglspider(object, groups, display = "sites", w = weights(object, display),
choices = 1:3, col = "black", ...)
orglellipse(object, groups, display = "sites", w = weights(object, display),
kind = c("sd", "se", "ehull"), conf, choices = 1:3, alpha = 0.3,
col = "red", ...)
orglspantree(object, spantree, display = "sites", choices = 1:3,
col = "black", ...)
orglcluster(object, cluster, prune = 0, display = "sites", choices = 1:3,
col = "black", ...)
Function ordirgl returns nothing.
An ordination result or any object known by scores.
Display "sites" or "species" or other
ordination object recognized by scores.
Selected three axes.
The type of plots: "p" for points or "t" for
text labels.
Axis colour (concerns only the crossed axes through the origin).
Colour of biplot arrows and centroids of
environmental variables.
Size of points in the units of ordination scores.
Text to override the default with type = "t".
Fitted environmental variables from envfit
displayed in the graph. Use envfit = NA to suppress display
of environmental variables in constrained ordination.
Text justification passed to text3d.
Factor giving the groups for which the graphical item is drawn.
Order points by this variable within groups.
Weights used to find the average within group. Weights are
used automatically for cca
and decorana results, unless undone by the
user. w=NULL sets equal weights to all points.
Draw ellipse for standard deviations of points
("sd") or standard deviations of their averages
("se") or an ellipsoid hull enclosing all points in the
group ("ehull".
Confidence limit for ellipses, e.g., 0.95. If not given, sd or se ellipses are drawn.
Colour of items. This can be a vector and factors are
interpreted as their internal numerical values. If the function
has a groups argument, vector col is used for each
of these, and for other functions it is matched to points in
ordirgl (see Details below).
Transparency of colour between 0.0 (fully transparent) and 1.0 (non-transparent).
A minimum spanning tree object from vegan
spantree.
Result of hierarchic cluster analysis, such as
hclust or agnes.
Number of upper levels hierarchies removed from the
tree. If prune > 0, tree will be cut into prune + 1
disconnected trees.
Other parameters passed to graphical functions.
Jari Oksanen
Function ordirgl uses OpenGL package rgl
which may not be functional in all platforms.
Function ordirgl plots dynamic graphics using OpenGL with the
rgl package. It clears the graphics device and
starts a new plot. The function was designed for ordination methods
in the vegan package, but it can handle any method known to
vegan scores function, or to any three
column matrix. The orgl-prefixed functions add items to the
opened rgl graphics device.
Function ordirgl uses most default settings of underlying
graphical functions in rgl. It plots only one set of points,
but functions orglpoints and orgltext can add new
items to an existing plot. The points are plotted using
spheres3d and the text using
texts3d which both have their own configuration
switches and their general look and feel can be modified with
material3d. The point size is directly defined
by radius argument in the units of ordination scores in
spheres3d, but ordirgl uses a default
size of 1% of the length of the longest axis, and this can be
further modified by the cex multiplier.
In constrained ordination (cca,
rda, capscale), biplot
arrows and centroids are always displayed similarly as in
two-dimensional plotting function plot.cca.
Alternatively, it is possible to display fitted environmental
vectors or class centroids from envfit in both
graphs. These are displayed similarly as the results of constrained
ordination, and they can be shown only for non-constrained
ordination. The user must remember to specify at least three axes in
envfit if the results are used with these
functions.
Function orglsegments is similar to vegan
ordisegments and connects points by line
segments. This can be useful for regular transects. The colour of
segments can be a vector which corresponds to the groups and
will be recycled.
Function orglspider is similar as vegan
ordispider: it connects points to their
weighted centroid within "groups", and in constrained
ordination it can connect "wa" or weighted averages scores to
corresponding "lc" or linear combination scores if
"groups" is missing. Function orglellipse is similar
as vegan ordiellipse and draws ellipsoids
of standard deviance, standard error or confidence regions for
groups. At least four points are needed to define an
ellipsoid in 3D, and even these will fail if all points are strictly
on 2D. The col argument for both of these functions can be a
vector corresponding to the groups.
Function orglspantree adds a minimum spanning tree from
vegan spantree. This a 3D equivalent of
lines.spantree. Function orglcluster
adds a hierarchic cluster tree from hclust or related
functions. This is a 3D equivalent of
ordicluster. The col argument for both
of these functions can be a vector corresponding to the connected
points. In orglspantree the line colour is a mixture of
colours of joined points, and in orglcluster it is a mixture
of all points in the cluster.
rgl, spheres3d,
text3d, rgl.viewpoint,
envfit. These are 3D dynamic variants of
vegan functions ordiplot,
ordisegments, ordispider
and ordiellipse,
ordicluster and
lines.spantree.
if (interactive() && require(rgl, quietly = TRUE)) {
data(mite, mite.env)
ord <- rda(decostand(mite, "hellinger"))
ordirgl(ord, size=4, col = "yellow")
orgltext(ord, display = "species")
## show groups of Shrub abundance
## ordirgl: col by points
with(mite.env, ordirgl(ord, col = as.numeric(Shrub), scaling = "sites"))
## orglspider & orglellipse: col by groups
with(mite.env, orglspider(ord, Shrub, col = 1:3, scaling = "sites"))
with(mite.env, orglellipse(ord, Shrub, col = 1:3, kind = "se", conf = 0.95,
scaling = "sites"))
}
Run the code above in your browser using DataLab