How to generate power variogram?

[Martin20494]

Hi team,

Thank you for a great package! I have a question about variogram/covariance model. Can gstools generate power variogram (as here)? (Sorry I also had the same question on gitter)

[MuellerSeb]

Hi there,

Power variogram are not directly supported, since we only support bounded variogram models. (There is an open issue for unbounded models: #192)

BUT: A special feature of GSTools are truncated power law variograms. These variograms are an approximation of power law variogram with an upper cut of bound. You could have a look there, if that already helps you.

See the publication by Di Federico and Neuman (1997) for theoretical background.

Hope that helps. Cheers,
Sebastian

[Martin20494]

Hi Mueller,

Thank you so much for your reply.

I have tried the truncated power law variograms, but unfortunately, it does not have the same shape as my estimated variogram to fit. I attached the code and the results as below. Do you still have any other suggestion?

# Package
import pandas as pd
import gstools as gs

# Read file
file_points = pd.read_csv(<path to file>, index_col=0)

# Estimate variogram
bin_center, gamma = gs.vario_estimate((file_points.x, file_points.y), file_points.z)

# Fit variogram
var = gs.TPLStable(
    dim=2,
    var=3,
    len_low=0,
    len_scale=1750,
    nugget=0.0001,
    hurst=0.5
)

var.fit_variogram(bin_center, gamma)

# Plot original variogram
ax = var.plot(x_max=max(bin_center))
ax.scatter(bin_center, gamma)