Geostatistical methods for 3D pollutants mapping in site remediation

Tridimensional (volume) spatial interpolation of soil pollutants concentrations or of other soil properties is a demanding phase of polluted site characterization. In this paper are presented the first steps of a research activity aimed to improve 3D spatialization using a geostatistical approach and information retrieved from satellite remote sensing images. The work done mainly consists in the implementation of a set of software tools for spatial correlation/variability (variogram) compute (experimental) and modeling (theoretical), for the data spatialization using Kriging interpolators, and for the validation (cross validation) of the results. These tools have been implemented in Fortran77 and MatLab® programming languages and are based on the GSLIB (Geostatistical Software Library) library. Besides the methodologies/ procedures above mentioned, an application to a real case is presented. The case study is an industrial area polluted by inorganic compounds and for which measures of Arsenic concentration obtained on samples collected at different locations and depths (196) area available. These data are first geostatistically studied, modeled (spatial variability), 3D interpolated using the ordinary Kriging method and validated.