gmGeostats (version 0.11.3)

LMCAnisCompo: Create a anisotropic model for regionalized compositions

Description

Creates a (potentially anisotropic) variogram model for variation-variograms

Usage

LMCAnisCompo(
  Z,
  models = c("nugget", "sph", "sph"),
  azimuths = rep(0, length(models)),
  ranges = matrix(1, nrow = length(models), ncol = G),
  sillarray = NULL,
  V = NULL,
  tol = 1e-12,
  G = 2
)

Value

an object of class "LMCAnisCompo" with all provided information appropriately structured, ready to be used for fitting, plotting or prediction.

Arguments

Z

compositional data set, used to derive the compositional dimension and colnames

models

string (or vector of strings) specifying which reference model(s) to use

azimuths

typically a vector providing, for each model, the direction of maximal continuity (measured from North clockwise)

ranges

typically a G-column matrix providing the minimal and maximal ranges, with one row per model (with G specified below)

sillarray

array of sills for each model. It can be null (to be estimated in the future). If specified, provide an appropriate (x,x,K)-array, where K is the number of models given, and x is explained below.

V

depending on what do you give as sillarray, you may need to provide the matrix of logcontrasts, or a string indicating whether the sills are represented in "alr" or "ilr"

tol

tolerance for determination of positive definiteness

G

one of c(1, 2, 3) identifying if we are in 1D, 2D or 3D cases

Examples

data("jura", package="gstat")
Zc = compositions::acomp(jura.pred[,7:9])
LMCAnisCompo(Zc, models=c("nugget", "sph"), azimuths=c(0,45))