Multivariate geostatistical and natural attenuation model approach for remediation of chlorinated compounds

Modelling spatial variability of environmental parameters through the application of modern techniques of geostatistics has permitted to acquire elements apt to evaluate the possibility of using Natural Attenuation as a remediation technique for groundwater contaminated by chlorinated compounds. Moreover, a precise delineation of hazardous areas in a polluted site is strongly based on accurate predictions of contaminant concentrations, a task that is complicated by the presence of censored data and highly positively skewed distributions. The present paper proposes a geostatistical approach that includes different methods for the study and the control of the ongoing activity of Natural Attenuation in an abandoned industrial area. The considered variables in this application are the values of concentration in groundwater of some aliphatic chlorinated carcinogenic compounds (PCE and TCE), measured in numerous piezometers, which show highly skewed distributions, characterized by a considerable number of values below the detection limit of the instrument. Indicator Kriging has been applied and the areas with 90% probability of exceeding the detection limit have been considered, inside which the variables were transformed through Gaussian Anamorphosis. The application of geostatistics has put into evidence the existence of a correlation between them not only from the chemical point of view, but also from the point of view of their spatial distribution: the shared presence of "hot spot" areas might be imputed to the existence of localized hazard sources. In those areas it is possible to detect coherence between the concentrations of the contaminants in study and the distribution of the parameters that control the phenomenon, which might be considered a clue that the degrading chemical activity is in progress. Successively, some punctual considerations have been done on the basis of the concentration values of contaminants coming from a following sampling plan on the same area, in order to verify this result. At the same time, an analytical model has been implemented to quantify the Natural Attenuation.