ABSTRACT
This paper presents a preliminary application of Spartan Spatial Random Fields (SSRFs) to a real geostatistical data set. The SSRFs bypass calculation and fitting of the experimental variogram and thus provide a computationally fast alternative to the classical geostatistical approach. The study focuses on the concentration of the heavy metal chromium (Cr) in the Jura region (Switzerland). A map of Cr concentration on a square prediction grid is generated based on a set of irregularly spaced measurements. A new estimation method that minimizes the SSRF interaction functional is applied instead of the commonly used kriging estimators.
