Sorvari, Jaana
(Finnish Environment Institute, 2010)
Monographs of the Boreal Environment Research 37
Land contamination is a significant environmental problem requiring systematic management actions. Defining the type and scale of the actions requires information on the risks involved. The numerous methods available for conducting risk assessment (RA) vary in terms of complexity, level of detail, conservatism, and outcomes. Thus, selecting suitable methods requires information on their applicability in Finnish conditions and at the specific site. On the other hand, it is generally accepted that current contaminated land management (CLM) should not only focus on minimizing site-specific risks, but should also consider overall environmental effects and socio-cultural and socio-economic aspects. Multi-Criteria Analysis (MCA) could then be used as a tool for integrating multidimensional data and generating aggregated information on the consequences of different risk management (RM) options, such as environmental, social, and economic impacts. Nonetheless, such approaches have very seldom been applied in CLM in Finland, probably partly due to a lack of tools specifically developed or modified for Finnish conditions.This research studied the application and suitability of different RA methods for assessing risks and identifying RM needs at some typical contaminated sites in Finland and demonstrated the use of MCA, the emphasis being on soil contamination. The studied RA approaches comprised qualitative rating and quantitative methods that were based on using environmental benchmarks, uptake and exposure models, and multimedia software. To derive estimates of ecological risks, the so-called TRIAD procedure that uses chemical studies, bioassays, and ecological studies was also applied and combined with MCA in order to account for the performance of the study methods, i.e. their ability to depict ecological risks at a study site. Qualitative rating and the statistical Monte Carlo technique provided additional means for uncertainty analysis. A separate study applying the Metaplan technique, interviews, a questionnaire, and a literature survey showed that a lack of suitable assessment tools was one of the key barriers to eco-efficient CLM in Finland. An MCA-based decision support tool (DST) adapting the Multi-Attribute Value Theory (MAVT) was therefore developed for case-by-case determination of the preferred RM option and tested with some typical Finnish contaminated sites.Many of the conclusions of the research are overarching and applicable to RA methods in general. Fistly, it appered that care must be taken in applying different models and software tools in site-specific RA, since some of their components are not straightforwardly suitable for Finnish conditions or for certain contaminants. These problems often relate to specific contaminant transport pathways. Moreover, the lack of verified data on the parameter values representative of Finnish conditions is an issue. The prevailing practice of using complicated software programs with ample data demands as the first and primary tools in human health risk assessment is not supported by this research, since it appeared that even simple tools and calculations can often provide adequate information on risks for decision-making. In ecological risk assessment (ERA), the usefulness of the approach founded on uptake and exposure models is reduced by the high uncertainties involved, particularly since the applicability of these models in Finnish conditions could not be verified. The accuracy and reliability of ecological risk estimates can be enhanced by applying the TRIAD methodology, although the procedure includes some pitfalls that need to be acknowledged. Combining TRIAD with MCA proved to be a feasible means to quantitatively study the performance of separate ERA methods. MCA thereby complements mechanical statistical analysis, such as Monte Carlo simulation, and increases the reliability of the final integrated risk estimates. In practice, a lack of data on the statistics of the input variables can restrict the use of statistical tools. The MAVT-based DST turned out to be efficient in facilitating discussion between different interest groups and experts and in identifying the preferred RM option in the common situation where risks are not the only factors relevant in decision-making. In practice, additional factors, such as the temporal scope of RM actions and some sustainability components that were not comprehensively included in the DST, might need to be considered.