Target cleanup levels at the site of a former manufactured gas plant in northern Italy: deterministic versus probabilistic results.

A quantitative comparison was performed between allowable soil concentrations calculated with deterministic and probabilistic risk analysis methods for a large and heavily contaminated industrial site, former Manufactured Gas Plant in Italy. The remediation costs to comply with soil standards issued by the regulatory agency were higher than the commercial value of the site. Deterministic cleanup targets based on a Risk Based Corrective Action (RBCA) procedure allowed a five-fold reduction of the volume of soil to be decontaminated. The site-specific, probabilistic analysis shows that the deterministic clean up standards were for volatile organic compounds lower than those required to meet the probabilistic criteria, but were higher, and thus more conservative, for polycyclic aromatic hydrocarbons and inorganic substances. The use of probabilistic methods shows that deterministic standards may hide significantly different level of conservatism in relationship to the uncertainty and variability present in each exposure parameter. The convenience of probabilistic calculation in avoiding over-remediation efforts is not sure as probabilistic results may be more stringent than deterministic ones, and the remediation costs are related not only to the single specie allowable concentration but also on the contaminants mix found on site. In-depth knowledge of the uncertainty implied in risk assessment calculations validates its implementation in policies for soil remediation of large industrial urban areas.