Abstract. In Geographie Information Systems there are many applications where the objeets of study can only be described by sampling information at selected point locations and by subsequent interpolation. Quite often inappropriate methods are used for the interpolation of sparse environmental data which do not take into account the characteristics of the data themselves. This paper presents the case for more widely using Kriging-methods in environmental data approximation. It gives a short theoretical overview of Kriging and the interpretation of variograms and goes on with a discussion of more advanced concepts (Co-Kriging, disjunetive Kriging), some examples and the Illustration of the use of Kriging within a GIS environment.