JCIEE 10(4), 291-299 (2000), Lee and Juang, Application of Disjunctive Kriging on Delineation of Heave-Metal Contaminated Soils
Journal of the Chinese Institute of Environmental Engineering, Vol. 10, No. 4, pp. 291-299 (2000)
Application of Disjunctive Kriging on Delineation of Heave-Metal Contaminated Soils
Dar-Yuan Lee and Kai-Wei Juang
Graduate Institute of Agricultural Chemistry National Taiwan University Taipei, Taiwan
Kriging, non-linear geostatistical model, heavy-metal contamination, risk of decision-making
The spatial distribution of pollutant is an essential information to the delineation of a contaminated site. Recently, kriging has been used to estimate the spatial distribution of pollutant. Based on the spatial distribution, the hazardous areas needed for remediation can be determined. However, the kriging estimates have uncertainty due to the measurement errors and data variation. Therefore, there are risks in determining hazardous areas needed for remediation only based on the kriging estimates. Thus, for the delineation of a contaminated site, the risks of false delineation should be taken into account. Disjunctive kriging is a non-linear geostatistical approach. One can use the disjunctive kriging estimator not only to estimate the pollutant concentration but also to obtain the conditional probability of the pollutant concentration above a given cutoff value. The conditional probability of the pollutant concentration being above the cutoff value can be coupled with the estimated pollutant concentration to obtain the risks of false decision-making. The purpose of this study is to investigate the feasibility of disjunctive kriging on delineation of heavy-metal contaminated soils. A real data set of soil Cd concentrations in a contaminated site in Taoyuan County, Taiwan was studied. The results of this study demonstrate that the normalized transformation used in disjunctive kriging can prevent the spatial structure from the interference of great variation and high skewness of original data. Moreover, under a given cutoff value (for example, Cd = 10 mg/kg) , the estimates of Cd concentration and the conditional probabilities of the Cd concentration, being less than or above 10 mg/kg at unsampled locations, obtained from the disjunctive kriging estimation were used to calculate the risks of false decision-making. The risks of false decision-making obtained by this approach are useful for decision-makers who need to delineate hazardous areas in a heavy metal contaminated site.