R: Empirical variogram

asr_varioGram {asreml}

R Documentation

Empirical variogram

Description

Calculates the empirical variogram from regular or irregular one- or two-dimensional data.

Usage

asr_varioGram(
  x,
  y,
  z,
  composite = TRUE,
  model = c("empirical"),
  metric = c("euclidean", "manhattan"),
  angle = 0,
  angle.tol = 180,
  nlag = 20,
  maxdist = 0.5,
  xlag = NA,
  lag.tol = 0.5,
  grid = TRUE,
  ...
)

Arguments

x

Numeric vector of x coordinates, may also be a matrix or data frame with two or three columns. If ncol(x) is three, the columns are taken to be the x and y coordinates and the response (z), respectively. If ncol(x) is two, the columns are taken to be the x coordinates and the response, respectively. In this case the y coordinates are generated as rep(1, nrow(x)).

y

Numeric vector of y coordinates.

z

The response vector.

composite

To be used for data on a regular grid. If TRUE, the average of the variograms in quadrants (x, y) and (x, -y) is returned. Otherwise, both variograms are returned and identified as quadrants 1 and 4 (default = TRUE).

model

At the present, it can only be "empirical".

metric

The distance between (x, y) points. Options are: "euclidean" or "manhattan" (default = "euclidean").

angle

A vector of directions. Angles are measured in degrees anticlockwise from the x axis (default = 0).

angle.tol

The angle subtended by each direction. That is, an arc angle +/- angle.tol/2 (default = 180, which gives an omnidirectional variogram).

nlag

The maximum number of lags (default = 20).

maxdist

The fraction of the maximum distance to include in the calculation. The default is half the maximum distance in the data: 0.5.

xlag

The width of the lags. If missing, xlag is set to maxdist/nlag.

lag.tol

The distance tolerance (default = 0.5).

grid

If FALSE, forces polar variograms if (x, y) specifies a regular grid (default = TRUE).

...

Additional arguments, for example core_version.

Details

For one-dimensional data the y coordinates need not be supplied and a vector of ones is generated. The function identifies data on a complete regular array and in such cases only computes polar variograms if grid = FALSE. The data are assumed sorted with the x coordinates changing the fastest; the data are sorted internally if this is not the case.

Value

A data frame including the following components:

x: The original x coordinates.
y: The original y coordinates.
gamma: The variogram estimate.
distance: The average distance for pairs in the lag.
np: The number of pairs in the lag.
angle: Direction if not a regular grid.

[Package asreml version 4.2.0.480 Index]