[Kriging] Solve kriging system from coviariance terms (instead of semivariance)

[SimonMolinsky]

Is your feature request related to a problem? Please describe.

• Literature shows equations to solve the kriging system from covariance terms, especially the Poisson Kriging system in here: Goovaerts, P. Kriging and Semivariogram Deconvolution in the Presence of Irregular Geographical Units. Math Geosci 40, 101–128 (2008). https://doi.org/10.1007/s11004-007-9129-1. The current version of the package (0.3.0) uses semivariance terms, but it should be diversified. TheoreticalVariogram class generates a covariogram, so the first step is done.
• Covariance calculations allow to the creation of a new set of tests to control the output of a Kriging.

Describe the solution you'd like At the beginning: add additional kriging flows for area-to-area PK and area-to-point PK where covariance terms are used instead of semivariance. All equations are available in the article cited above.

Describe alternatives you've considered n/a

Additional context Goovaerts, P. Kriging and Semivariogram Deconvolution in the Presence of Irregular Geographical Units. Math Geosci 40, 101–128 (2008). https://doi.org/10.1007/s11004-007-9129-1: pages 108-110