compute the empirical variogram or covariance function in a 3D case study
gsi.EVario3D(
X,
Z,
Ff = rep(1, nrow(X)),
maxdist = max(dist(X[sample(nrow(X), min(nrow(X), 1000)), ]))/2,
lagNr = 15,
lags = seq(from = 0, to = maxdist, length.out = lagNr + 1),
dirvecs = t(c(1, 0, 0)),
angtol = 90,
maxbreadth = Inf,
minpairs = 10,
cov = FALSE
)An empirical variogram for the provided data. NOTE: avoid using directly gsi.* functions! They
represent either internal functions, or preliminary, not fully-tested functions. Use variogram instead.
matrix of Nx3 columns with the geographic coordinates
matrix or data.frame of data with dimension (N,Dv)
for variogram, matrix of basis functions with nrow(Ff)=N (can be a N-vector of 1s; should include the vector of 1s!!); for covariance function, a (N,Dv)-matrix or a Dv-vector giving the mean values
maximum lag distance to consider
number of lags to consider
if maxdist and lagNr are not specified, either: (a) a matrix of 2 columns giving minimal and maximal lag distance defining the lag classes to consider, or (b) a vector of lag breaks
matrix which rows are the director vectors along which variograms will be computed (these will be normalized!)
scalar, angular tolerance applied (in degrees; defaults to 90??, ie. isotropic)
maximal breadth (in lag units) orthogonal to the lag direction (defaults to Inf, ie. not used)
minimal number of pairs falling in each class to consider the calculation sufficient; defaults to 10
boolean, is covariance (TRUE) or variogram (FALSE) desired? defaults to variogram
Other gmEVario functions:
as.gmEVario.gstatVariogram(),
gsi.EVario2D(),
ndirections(),
plot.gmEVario(),
variogramModelPlot()