ABSTRACT
The method of kriging was designed to provide an estimator for a phenomenon at positions where it was not observed. The determination of the interpolated value is not carried out by some arbitrary algorithm, but takes into account the statistical characteristics of the phenomenon. These characteristics are derived from the available observations by inspecting the correlation of values with respect to the spatial, temporal or spatio-temporal distance of the observations. The plotting of the halved variance against the distance for all possible observation pairs is called the experimental variogram. The general trend of this point cloud can be represented by a mathematical function, the theoretical variogram. There are multiple types of theoretical variograms to choose from with associated parameters that need to be adjusted to the actual data. Since kriging is a statistical approach, it also provides an estimation of uncertainty for each interpolation, the kriging variance. This feature is unique amongst interpolation methods and it can be exploited for multiple purposes. This chapter provides a survey of the general ideas behind the method of kriging. The most commonly used concepts and approaches are introduced and also represented graphically.
