vgm: Generate, or Add to Variogram Model

vgmR Documentation

Generate, or Add to Variogram Model


Generates a variogram model, or adds to an existing model. print.variogramModel prints the essence of a variogram model.


vgm(psill = NA, model, range = NA, nugget,, anis, kappa = 0.5, ..., covtable,
	Err = 0)
## S3 method for class 'variogramModel'
print(x, ...)
## S3 method for class 'variogramModel'
plot(x, cutoff, ..., type = 'l')



(partial) sill of the variogram model component, or model: see Details


model type, e.g. "Exp", "Sph", "Gau", "Mat". Calling vgm() without a model argument returns a data.frame with available models.


range parameter of the variogram model component; in case of anisotropy: major range


smoothness parameter for the Matern class of variogram models


nugget component of the variogram (this basically adds a nugget compontent to the model); if missing, nugget component is omitted

the variogram model to which we want to add a component (structure)


anisotropy parameters: see notes below


a variogram model to print or plot


arguments that will be passed to print, e.g. digits (see examples), or to variogramLine for the plot method


if model is Tab, instead of model parameters a one-dimensional covariance table can be passed here. See covtable.R in tests directory, and example below.


numeric; if larger than zero, the measurement error variance component that will not be included to the kriging equations, i.e. kriging will now smooth the process Y instead of predict the measured Z, where Z=Y+e, and Err is the variance of e


object of class variomodel, see geoR


maximum distance up to which variogram values are computed


plot type


If only the first argument (psill) is given a character value indicating a model, as in vgm("Sph"), then this taken as a shorthand form of vgm(NA,"Sph",NA,NA), i.e. a spherical variogram with nugget and unknown parameter values; see examples below. Read fit.variogram to find out how NA variogram parameters are given initial values for a fitting a model, based on the sample variogram. Package automap gives further options for automated variogram modelling.


If a single model is passed, an object of class variogramModel extending data.frame.

In case a vector ofmodels is passed, an object of class variogramModelList which is a list of variogramModel objects.

When called without a model argument, a data.frame with available models is returned, having two columns: short (abbreviated names, to be used as model argument: "Exp", "Sph" etc) and long (with some description).

as.vgm.variomodel tries to convert an object of class variomodel (geoR) to vgm.


Geometric anisotropy can be modelled for each individual simple model by giving two or five anisotropy parameters, two for two-dimensional and five for three-dimensional data. In any case, the range defined is the range in the direction of the strongest correlation, or the major range. Anisotropy parameters define which direction this is (the main axis), and how much shorter the range is in (the) direction(s) perpendicular to this main axis.

In two dimensions, two parameters define an anisotropy ellipse, say anis = c(30, 0.5). The first parameter, 30, refers to the main axis direction: it is the angle for the principal direction of continuity (measured in degrees, clockwise from positive Y, i.e. North). The second parameter, 0.5, is the anisotropy ratio, the ratio of the minor range to the major range (a value between 0 and 1). So, in our example, if the range in the major direction (North-East) is 100, the range in the minor direction (South-East) is 0.5 x 100 = 50.

In three dimensions, five values should be given in the form anis = c(p,q,r,s,t). Now, $p$ is the angle for the principal direction of continuity (measured in degrees, clockwise from Y, in direction of X), $q$ is the dip angle for the principal direction of continuity (measured in positive degrees up from horizontal), $r$ is the third rotation angle to rotate the two minor directions around the principal direction defined by $p$ and $q$. A positive angle acts counter-clockwise while looking in the principal direction. Anisotropy ratios $s$ and $t$ are the ratios between the major range and each of the two minor ranges. The anisotropy code was taken from GSLIB. Note that in it is reported that this code has a bug. Quoting from this site: “The third angle in all GSLIB programs operates in the opposite direction than specified in the GSLIB book. Explanation - The books says (pp27) the angle is measured clockwise when looking toward the origin (from the postive principal direction), but it should be counter-clockwise. This is a documentation error. Although rarely used, the correct specification of the third angle is critical if used.”

(Note that anis = c(p,s) is equivalent to anis = c(p,0,0,s,1).)

The implementation in gstat for 2D and 3D anisotropy was taken from the gslib (probably 1992) code. I have seen a paper where it is argued that the 3D anisotropy code implemented in gslib (and so in gstat) is in error, but I have not corrected anything afterwards.


Edzer Pebesma


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

Deutsch, C.V. and Journel, A.G., 1998. GSLIB: Geostatistical software library and user's guide, second edition, Oxford University Press.

For the validity of variogram models on the sphere, see Huang, Chunfeng, Haimeng Zhang, and Scott M. Robeson. On the validity of commonly used covariance and variogram functions on the sphere. Mathematical Geosciences 43.6 (2011): 721-733.

See Also

show.vgms to view the available models, fit.variogram, variogramLine, variogram for the sample variogram.


vgm(NA, "Sph", NA, NA)
vgm(, "Sph") # "Sph" is second argument: NO nugget in this case
vgm(10, "Exp", 300)
x <- vgm(10, "Exp", 300)
vgm(10, "Nug", 0)
vgm(10, "Exp", 300, 4.5)
vgm(10, "Mat", 300, 4.5, kappa = 0.7)
vgm( 5, "Exp", 300, = vgm(5, "Exp", 60, nugget = 2.5))
vgm(10, "Exp", 300, anis = c(30, 0.5))
vgm(10, "Exp", 300, anis = c(30, 10, 0, 0.5, 0.3))
# Matern variogram model:
vgm(1, "Mat", 1, kappa=.3)
x <- vgm(0.39527463, "Sph", 953.8942, nugget = 0.06105141)
print(x, digits = 3);
# to see all components, do
vv=vgm(model = "Tab",  covtable = 
	variogramLine(vgm(1, "Sph", 1), 1, n=1e4, min = 0, covariance = TRUE))
vgm(c("Mat", "Sph"))
vgm(, c("Mat", "Sph")) # no nugget

gstat documentation built on April 6, 2023, 5:21 p.m.