Regional geochemical mapping constrained by airborne radiometric data and satellite imagery

We employ fuzzy Gustafson-Kessel cluster analysis for the rapid and objective integration of Landsat imagery and airborne radiometric data. Fuzzy cluster analyses describe the structural heterogeneity present in all input data sets in a fuzzy sense using a set of membership information. We analyze this membership information to identify 46 locations for geochemical sampling in our 3500 km2 survey area located in South Africa. Consecutively, we interpolate the zinc concentration measured at these selected locations using the membership information obtained from cluster analysis of the satellite imagery and airborne radiometric data. Finally, we compare the zinc concentrations predicted by our interpolated map to more than 2500 additional zinc concentration measurements in the survey area and find good structural agreement.