Find the "best" variogram parameters for a given geodata object. Several different models are tested; the one with the lowest least squares error is chosen.
krig_auto_params(geodata, trend = "cte", breaks = krig_breaks(2, 320, 30))
geodata | a list containing element |
---|---|
trend | specifies the mean part of the model.
See documentation of |
breaks | a vector with values to define the variogram binning. Only
used when |
A list of the best fitted variogram parameters. The following
description is adapted from geoR::variog()
-- which you should see
for more details:
nugget: value of the nugget parameter. An estimated value if
fix.nugget = FALSE
or a fixed value if fix.nugget = TRUE
.
sill and range: First and second elements of cov.pars` -- a two elements vector with estimated values of the covariance parameters sigma^2 and phi, respectively.
kappa: Fixed value of the smoothness parameter.
model: Name of the correlation function (see cov.model
).
minVM: The minimum fit error.
vg: The variogram.
The default breaks argument is set to have more points where the exponential curve rises the most and fewer at large distances. This means that the curve fitting is not overly biased by points beyond the effective maximum range.