Anisotropic Spartan random field models for geostatistical analysis

Spartan random fields (RF’s) were recently introduced as computationally fast and parametrically frugal models for geostatistical applications. We present a method based on the covariance tensor identity that allows incorporating anisotropic spatial correlations in the Spartan models and the efficient calculation of anisotropic parameters. The method is applied to determine the anisotropy parameters of lignite quality data from a mine in West Macedonia, Greece. 1