gmGeostats (version 0.11.3)

gsi.gstatCokriging2compo: Reorganisation of cokriged compositions

Description

Produce compositional predictions out of a gstat::gstat() prediction

Usage

gsi.gstatCokriging2compo(COKresult, ...)
# S3 method for default
gsi.gstatCokriging2compo(COKresult, ...)
# S3 method for data.frame
gsi.gstatCokriging2compo(
  COKresult,
  V = NULL,
  orignames = NULL,
  tol = 1e-12,
  nscore = FALSE,
  gg = NULL,
  ...
)
# S3 method for default
gsi.gstatCokriging2rmult(COKresult, ...)
# S3 method for data.frame
gsi.gstatCokriging2rmult(COKresult, nscore = FALSE, gg = NULL, ...)

Value

an (N,D)-object of class c("spatialGridAcomp","acomp")

with the predictions, together with an extra attribute "krigVar" containing the cokriging covariance matrices in an (N, D, D)-array; here N=number of interpolated locations, D=number of original components of the composition

Arguments

COKresult

output of a gstat::predict.gstat() cokriging, typically of class "data.frame", sp::SpatialPointsDataFrame(), sp::SpatialGridDataFrame() or sp::SpatialPixelsDataFrame()

...

further arguments needed for nscore (deprecated)

V

string or matrix describing which logratio was applied ("ilr", "alr", or a matrix computing the ilr corrdinates; clr is not allowed!)

orignames

names of the original components (optional, but recommended)

tol

for generalized inversion of the matrix (rarely touched!)

nscore

boolean, were the data normal score-transformed? (deprecated)

gg

in the case that normal score transformation was applied, provide the gstat object! (deprecated)

Methods (by class)

• default: Reorganisation of cokriged compositions

• data.frame: Reorganisation of cokriged compositions

• default: Reorganisation of cokriged multivariate data

• data.frame: Reorganisation of cokriged multivariate data

See Also

image_cokriged.spatialGridRmult() for an example