Much research has been done characterizing surficial contaminants using geostatistics or other spatial estimation technique. This chapter examines the use of a non-Euclidean distance metric combined with geostatistical techniques to model the surficial distribution of dioxins in Imerman Park near Midland, Michigan. This chapter also examines the applicability of geostatistics to small data sets. An overview of the dioxin sampling in Midland, MI, will be examined, followed by a brief overview of geostatistical theory, variogram modeling, and the use of non-Euclidean distance metrics to capture the geologic processes. Preliminary results of a case study evaluating the surficial dioxin distribution in Imerman Park downstream from the Dow Midland plants will then be presented, comparing the use of a flood plain non-Euclidean distance norm versus a Euclidean distance norm.