Connect some coordinates to a multivariate data set (of hard data, of predictions or of simulations); currently, the coordinates are stored in an attribute and the dataset is given a complex S3 class. This functionality will change in the future, to make use of package "sp" classes.
spatialGridRmult(coords, data, dimcomp = 2, dimsim = NA)
coordinates of the locations
(observed or predicted) rmult or matrix data set; or else array of simulated rmult /real-valued multivariate data
which of the dimensions of data
does correspond to the
variables?
if data
contains simulations, which of its dimensions does
run across the realisations? leave it as NA if data
has observations or predictions.
A (potentially transposed/aperm-ed) matrix of class c("spatialGridAcomp","acomp") with the coordinates in an extra attribute "coords".
image_cokriged.spatialGridRmult()
for an example; gsi.gstatCokriging2rmult()
to
restructure the output from gstat::predict.gstat()
confortably