Quantification of the effect of spatially varying environmental contaminants into a cost model for soil remediation

In this study we investigated the effects of spatial variability of soil contaminants on cost calculations for soil remediation. Most cost models only provide a single figure, whereas spatial variability is one of the sources to contribute to the uncertainty. A cost model is applied to a study site of 18 ha containing a former gasworks in the Rotterdam harbour. The site was contaminated by heavy metals, PAH and mineral oil. Two sets of environmental thresholds were applied, one for identifying the severeness of contamination and one to decide upon the future use of excavated soil. Three remediation scenarios were compared. Geostatistical simulations were applied, both on individual contaminants and on indicator variables derived from these. As it turns out, spatial uncertainty causes 2¿5% uncertainty in the final cost estimates. Another source of uncertainty is the direction of application of the cost model: a least-case approach starts with the lowest threshold value, followed by increasingly higher values, whereas a worst-case approach starts with the highest threshold value followed by decreasing values. Using a worst-case approach yielded cost estimates that were 6¿8% higher than cost estimates by a least-case approach. We concluded that 8¿13% of the uncertainty in cost estimates could be explained by spatial variation of soil contaminants and litholog