Skip to contents

Linear model of coregionalization and cokriging

Source: R/LmcCokriging.R
LmcCovariance.Rd

LmcCovariance: cross-covariance matrix \(C(h) = \sum_s B_s \rho_s(h)\) of a linear model of coregionalization (a single structure is the intrinsic coregionalization model). LmcCokriging: ordinary cokriging (one unbiasedness constraint per variable, as gstat) or simple cokriging (known means) of heterotopic data in any dimension, with the cokriging weights. Identical to the Python arm morie.fn.lmckrige.

Usage

LmcCovariance(h, lmc)

LmcCokriging(z, coords, var, new_coords, lmc, target = 0L, means = NULL)

Arguments

h

Distance.

lmc

List of structures, each a list with model (Exp, Gau, Sph, Mat, Nug), range, optional kappa and a symmetric p by p matrix B.

z

Observations.

coords

Matrix of observation coordinates (any dimension).

var

Variable of each observation (0-based).

new_coords

Matrix of target coordinates.

target

Variable to predict (0-based).

means

Known means (simple cokriging), or NULL.

Value

Matrix (LmcCovariance) or list with prediction, variance and weights.

References

Wackernagel, H. (2003). Multivariate Geostatistics, 3rd edn. Springer.

Pebesma, E. J. (2004). Multivariable geostatistics in S: the gstat package. Computers and Geosciences 30, 683-691.

Examples

lmc <- list(list(model = "Exp", range = 2, B = matrix(c(1, 0.6, 0.6, 1), 2)))
LmcCovariance(1, lmc)
#>           [,1]      [,2]
#> [1,] 0.6065307 0.3639184
#> [2,] 0.3639184 0.6065307
LmcCokriging(c(1, 2, 1.5, 0.5), rbind(c(0, 0), c(2, 0), c(1, 0), c(0, 1)), c(0, 0, 1, 1),
             rbind(c(1, 1)), lmc)$prediction
#> [1] 1.426632