vario.transfo {RGeostats}

R Documentation

Performs transformations between experimental variograms

Description

Performs transformations between experimental variograms

Usage

vario.transfo(string, vlist, oper=NA, mZ=NA, vZ=NA, vY=NA)

Arguments

string	Calculation string describing the transformation involving the experimental variograms, for all calculation directions. The user will refer to the variogram contents ("v1" for the first one, "v2" for the second one and so on), to the weight per lag or number of pairs ("w1" for the first one, "w2" for the second one, and so on).
vlist	A list of vario-class. The first experimental variogram should be refered as "v1" (for the variogram contents, "w1" for the variogram weight) in the calculation string. The second experimental variogram should be refered as "v2" (for the variogram contents), "w2" for the variogram weight) in the calculation string.
oper	Prior transformation to be performed on all variograms involved in the transformation string. It can be one of the following options: g12/g1 Divide the cross-variogram by the first simple variogram g12/g2 Divide the cross-variogram by the second simple variogram g12/sqrt(g1*g2) Divide the cross-variogram by the square root of the product of the two simple variograms norm Normalize each simple variogram by its variance and each cross-variogram by the square root of the product of the variances. log Derive the variogram of the Back-transform from the one of the Log-variable (see Details for more information)
mZ,vZ,vY	Constant used to define respectively the mean and variance of raw variable (Z) and the variance of log-transformed variable (Y). These constant are used when oper='log'.

Details

The calculation 'string' must be defined as a string (included between quotes), such as "w1*v1+w2*v2)/(w1+w2)".

Note that the variance (also included in the variogram structure) is also updated using the same string, where this time "v1" stands for the variance of the first variogram and "v2" for the one of the second variogram.

In the multivariate case, this is expanded to the array of variance-covariances. For this calculation, the weights "w1" and "w2" are replaced by 1.

No transformation is performed on the distances. The distances are copied from the first variogram ("v1").

As an example, the normalization of a variogram (called "vario") is obtained using the following syntax: vario.transfo("v1",vlist=vario,oper="norm")

When 'oper' = 'Log', we consider the following case:

Y = log(1 + Z/b)

where Z if the initial variable and Y is the current (log-transformed) variable. We assume that the variogram of Y is calculated and provided as input to this function. The transformation derives the variogram of Z from the one of Y. It requires the following arguments to be provided by the user:

• mZ the mean of the initial variable Z

• vZ the variance of the initial variable Z

• vY the variance of the log-transformed variable Y

Value

The resulting experimental variogram

---