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
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
|
\name{RMwhittlematern}
\alias{RMwhittle}
\alias{RMkbessel}
\alias{RMmatern}
\alias{RMhandcock}
\alias{Sobolev}
\alias{whittle-matern}
\title{Whittle-Matern Covariance Model}
\description{
\command{\link{RMmatern}} is a stationary isotropic covariance model
belonging to the Matern family.
The corresponding covariance function only depends on the distance
\eqn{r \ge 0}{r \ge 0}
between two points.
The Whittle model is given by
\deqn{C(r)=W_{\nu}(r)=2^{1- \nu}
\Gamma(\nu)^{-1}r^{\nu}K_{\nu}(r)}{C(r)=W_{\nu}(r)=2^{1- \nu}
\Gamma(\nu)^{-1}r^{\nu}K_{\nu}(r)}
where \eqn{\nu > 0}{\nu > 0} and \eqn{K_\nu}{K_\nu} is the modified
Bessel function of second kind.
The Matern model is given by
\deqn{C(r) = \frac{2^{1-\nu}}{\Gamma(\nu)} (\sqrt{2\nu}r)^\nu
K_\nu(\sqrt{2\nu}r)}{C(r) = 2^{1- \nu} \Gamma(\nu)^{-1} (\sqrt{2\nu}
r)^\nu K_\nu(\sqrt{2\nu} r)}
The Handcock-Wallis parametrisation is given by
\deqn{C(r) = \frac{2^{1-\nu}}{\Gamma(\nu)} (2\sqrt{\nu}r)^\nu
K_\nu(2 \sqrt{\nu}r)}{C(r) = 2^{1- \nu} \Gamma(\nu)^{-1} (2\sqrt{\nu}
r)^\nu K_\nu(2\sqrt{\nu} r)}
}
\usage{
RMwhittle(nu, notinvnu, var, scale, Aniso, proj)
RMmatern(nu, notinvnu, var, scale, Aniso, proj)
RMhandcock(nu, notinvnu, var, scale, Aniso, proj)
}
\arguments{
\item{nu}{a numerical value called \dQuote{smoothness parameter};
should be greater than 0.}
\item{notinvnu}{logical. If \code{FALSE} then in the definition
of the models \eqn{\nu} is replaced by \eqn{1/\nu}.
This parametrization seems to be more natural.
Default is, however, \code{TRUE} according with the definitions in
literature.
}
\item{var,scale,Aniso,proj}{optional arguments; same meaning for any
\command{\link{RMmodel}}. If not passed, the above
covariance function remains unmodified.}
}
\details{
The three models
are alternative parametrizations of the same covariance function.
The Matern model or the Handcock-Wallis parametrisation
should be preferred as they seperate the
effects of the scaling parameter and the shape parameter.
The Whittle-Matern model is the model of choice if the smoothness of a
random field is to be parametrized: the sample paths of a Gaussian
random field with this covariance structure are \eqn{m}{m} times
differentiable if and only if \eqn{\nu > m}{\nu > m} (see Gelfand et
al., 2010, p. 24).
Furthermore, the fractal dimension (see also \command{\link{RFfractaldim}})
\emph{D} of the Gaussian sample paths
is determined by \eqn{\nu}{\nu}: We have
\deqn{D = d + 1 - \nu, \nu \in (0,1)}{D = d + 1 - \nu, 0 < \nu < 1}
and \eqn{D = d}{D = d} for \eqn{\nu > 1}{\nu > 1} where \eqn{d}{d} is
the dimension of the random field (see Stein, 1999, p. 32).
If \eqn{\nu=0.5}{\nu=0.5} the Matern model equals \command{\link{RMexp}}.
For \eqn{\nu}{\nu} tending to \eqn{\infty}{\infty} a rescaled Gaussian
model \command{\link{RMgauss}} \eqn{C(r) = -r^2}
appears as limit of the above Handcock-Wallis parametrisation.
For generalizations see section \sQuote{See Also}.
}
\note{
The Whittle-Matern model is a normal scale mixture.
}
\value{
The functions return an object of class \code{\link[=RMmodel-class]{RMmodel}}.
}
\references{
Covariance function
\itemize{
\item Chiles,
J.-P. and Delfiner, P. (1999)
\emph{Geostatistics. Modeling Spatial Uncertainty.}
New York: Wiley.
\item Gelfand, A. E., Diggle, P., Fuentes, M. and Guttorp,
P. (eds.) (2010) \emph{Handbook of Spatial Statistics.}
Boca Raton: Chapman & Hall/CRL.
\item Guttorp, P. and Gneiting, T. (2006) Studies in the
history of probability and statistics. XLIX. On the Matern
correlation family. \emph{Biometrika} \bold{93}, 989--995.
\item Handcock, M. S. and Wallis, J. R. (1994) An approach to
statistical spatio-temporal modeling of meteorological fields.
\emph{JASA} \bold{89}, 368--378.
\item Stein, M. L. (1999) \emph{Interpolation of Spatial Data --
Some Theory for Kriging.} New York: Springer.
}
Tail correlation function (for \eqn{\nu \in (0,1/2]}{0 < \nu \le 1/2})
\itemize{
\item Strokorb, K., Ballani, F., and Schlather, M. (2014)
Tail correlation functions of max-stable processes: Construction
principles, recovery and diversity of some mixing max-stable
processes with identical TCF.
\emph{Extremes}, \bold{} Submitted.
}
}
\me
\seealso{
\itemize{
\item \command{\link{RMexp}}, \command{\link{RMgauss}} for special
cases of the model (for \eqn{\nu=0.5}{\nu=0.5} and
\eqn{\nu=\infty}{\nu=\infty}, respectively)
\item \command{\link{RMhyperbolic}} for a univariate
generalization
\item \command{\link{RMbiwm}} for a multivariate generalization
\item \command{\link{RMnonstwm}}, \command{\link{RMstein}} for anisotropic (space-time) generalizations
% \item \command{\link{}} for
\item \command{\link{RMmodel}},
\command{\link{RFsimulate}},
\command{\link{RFfit}} for general use.
}
}
\keyword{spatial}
\keyword{models}
\examples{\dontshow{StartExample()}
RFoptions(seed=0) ## *ANY* simulation will have the random seed 0; set
## RFoptions(seed=NA) to make them all random again
x <- seq(0, 1, len=100)
model <- RMwhittle(nu=1, Aniso=matrix(nc=2, c(1.5, 3, -3, 4)))
plot(model, dim=2, xlim=c(-1,1))
z <- RFsimulate(model=model, x, x)
plot(z)
\dontshow{FinalizeExample()}}
|
