Geostatistical modelling of heavy metal contamination in the topsoil of Guadiamar river margins (S Spain) using a stochastic simulation technique

This paper presents a methodology to evaluate and map environmental hazard and characterize environmental impact of heavy metals in Guadiamar river margins (South of Spain) contaminated by the Aznalcollar mining spill. This methodology accounts for uncertainties in the mapping of the probability of soil contamination with different heavy metals and is based on a co-simulation technique using direct sequential simulation of a multivariate set of variables. Each variable is simulated based on primary data the heavy metal concentrations in the soil - and on secondary data, consisting of a previously simulated map of one of the heavy metals. Using several realizations of the simulated multivariate set, a "hazard" index was calculated for each pixel of the study area, based on the simultaneous proportions (joint probabilities) of different levels of all metals. Finally, the intersection of the hazard map, based on the joint dispersion of all contaminants, with the environmental impact map for the different ecosystems, resulted in environmental hazard maps for the river margins. The performances of the multivariate sets of co-simulated variables were compared considering two extreme alternatives: i) the soil is considered in need of treatment if all 5 heavy metals simultaneously exceed their concentration limit value at the same location; ii) the soil is considered in need of treatment if at least one heavy metal exceeds its concentration limit at the same location. (c) 2006 Elsevier B.V. All rights reserved.