fit.vgmModel: Fit variogram using point data

In landmap: Automated Spatial Prediction using Ensemble Machine Learning

Description

Fit variogram using point data

Usage

## S4 method for signature 'formula,data.frame,SpatialPixelsDataFrame'
fit.vgmModel(
  formulaString.vgm,
  rmatrix,
  predictionDomain,
  cov.model = "exponential",
  dimensions = list("2D", "3D", "2D+T", "3D+T"),
  lambda = 0.5,
  psiR = NULL,
  subsample = nrow(rmatrix),
  ini.var,
  ini.range,
  fix.psiA = FALSE,
  fix.psiR = FALSE,
  ...
)

Arguments

formulaString.vgm	formula.
rmatrix	data.frame with coordinates and values of covariates.
predictionDomain	SpatialPixelsDataFrame.
cov.model	covariance model type used by the geoR package.
dimensions	optional 2D or 3D dimensions.
lambda	transformation value used by the geoR package.
psiR	range parameter used by the geoR package.
subsample	number of subset of original samples.
ini.var	initial variance (sill) used by the geoR package.
ini.range	initial range parameter used by the geoR package.
fix.psiA	setting used by the geoR package.
fix.psiR	setting used by the geoR package.
...	optional arguments to pass to the geoR package.

Value

Fitted variogram model

Note

Extends variogram fitting functionality from the geoR package. Can be used for 2D or 3D point data sets, with and without trend variables. Models need to be in the form zinc ~ dist and only numeric variables are allowed. Often reports Singular matrix. Covariates may have different orders of magnitude. if the covariates are perfectly aligned.

Author(s)

Tom Hengl

Examples

library("geoR")
library(rgdal)
demo(meuse, echo=FALSE)
vgm = fit.vgmModel(zinc~dist, as.data.frame(meuse), meuse.grid["dist"], lambda=1)
plot(variog(vgm$geodata))
lines(vgm$vgm)

landmap documentation built on Oct. 14, 2021, 5:24 p.m.