1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
|
\name{taylor.diagram}
\alias{taylor.diagram}
\title{ Taylor diagram }
\description{ Display a Taylor diagram}
\usage{
taylor.diagram(ref,model,add=FALSE,col="red",pch=19,pos.cor=TRUE,
xlab="",ylab="",main="Taylor Diagram",show.gamma=TRUE,ngamma=3,
gamma.col=8,sd.arcs=0,ref.sd=FALSE,grad.corr.lines=c(0.2,0.4,0.6,0.8,0.9),
pcex=1,normalize=FALSE,mar=c(5,4,6,6),...)
}
\arguments{
\item{ref}{numeric vector - the reference values.}
\item{model}{numeric vector - the predicted model values.}
\item{add}{whether to draw the diagram or just add a point.}
\item{col}{the color for the points displayed.}
\item{pch}{the type of point to display.}
\item{pos.cor}{whether to display only positive (\samp{TRUE}) or all
values of correlation (\samp{FALSE}).}
\item{xlab,ylab}{plot axis labels.}
\item{main}{title for the plot.}
\item{show.gamma}{whether to display standard deviation arcs around
the reference point (only for \samp{pos.cor=TRUE}).}
\item{ngamma}{the number of gammas to display (default=3).}
\item{gamma.col}{color to use for the gamma arcs (only with pos.cor=TRUE).}
\item{sd.arcs}{whether to display arcs along the standard deviation axes
(see Details).}
\item{ref.sd}{whether to display the arc representing the reference
standard deviation.}
\item{grad.corr.lines}{the values for the radial lines for correlation
values (see Details).}
\item{pcex}{character expansion for the plotted points.}
\item{normalize}{whether to normalize the models so that the reference
has a standard deviation of 1.}
\item{mar}{margins - only applies to the \samp{pos.cor=TRUE} plot.}
\item{...}{Additional arguments passed to \samp{plot}.}
}
\details{
The Taylor diagram is used to display the quality of model predictions
against the reference values, typically direct observations.
A diagram is built by plotting one model against the reference,
then adding alternative model points. If \samp{normalize=TRUE}
when plotting the first model, remember to set it to \samp{TRUE}
when plotting additional models.
Two displays are available. One displays the entire range of correlations
from -1 to 1. Setting \samp{pos.cor} to \samp{FALSE} will produce this
display. The -1 to 1 display includes a radial grid for the correlation
values. When \samp{pos.cor} is set to \samp{TRUE}, only the
range from 0 to 1 will be displayed. The \samp{gamma} lines and the arc at
the reference standard deviation are optional in this display.
Both the standard deviation arcs and the gamma lines are optional in the
\samp{pos.cor=TRUE} version. Setting \samp{sd.arcs} or \samp{grad.corr.lines}
to zero or FALSE will cause them not to be displayed. If more than one value is
passed for \samp{sd.arcs}, the function will try to use the values passed,
otherwise it will call \samp{pretty} to calculate the values.
}
\value{
The values of \samp{par} that preceded the function. This allows the
user to add points to the diagram, then restore the original values. This
is only necessary when using the 0 to 1 correlation range.
}
\references{
Taylor, K.E. (2001) Summarizing multiple aspects of model performance in a
single diagram. Journal of Geophysical Research, 106: 7183-7192.
}
\author{Olivier Eterradossi with modifications by Jim Lemon}
\examples{
# fake some reference data
ref<-rnorm(30,sd=2)
# add a little noise
model1<-ref+rnorm(30)/2
# add more noise
model2<-ref+rnorm(30)
# display the diagram with the better model
oldpar<-taylor.diagram(ref,model1)
# now add the worse model
taylor.diagram(ref,model2,add=TRUE,col="blue")
# get approximate legend position
lpos<-1.5*sd(ref)
# add a legend
legend(lpos,lpos,legend=c("Better","Worse"),pch=19,col=c("red","blue"))
# now restore par values
par(oldpar)
# show the "all correlation" display
taylor.diagram(ref,model1,pos.cor=FALSE)
taylor.diagram(ref,model2,add=TRUE,col="blue")
}
\keyword{misc}
|
