Project objectives:
A NICOLE-sponsored study to evaluate the risk assessment models currently in use throughout Europe was completed in 2003. Eleven industrial members of NICOLE
(Akzo Nobel, BNFL, BP, Fortum, ICI, JM Bostad, Powergen, SecondSite, Shell, Solvay and Total) plus the NICOLE organization itself sponsored the effort, which
was undertaken by Arcadis Geraghty and Miller, in their Newmarket, UK office. Additionally, the Netherlands-based SKB organization assembled an independent
team of experts to critically review the work.
Over the last half dozen or so years, the philosophy of using a risk-based approach for addressing contaminated sites has found widespread acceptance throughout
Europe. However, individual countries within the EU have developed ‘national models’ for environmental risk assessment which vary with legislative requirements
and are at different levels of development. Since risk-based clean-up targets will vary with each model, the credibility of risk assessment could be undermined
if there is a lack of understanding of why the results differ.
The aim of the Risk Assessment Model study is to benchmark and compare the different models, clearly explain the reasons behind any disparities, and generally
show that the fate and transport components of most models are conservative when compared against actual field data. It was not the intention to rank the models
or show one to be superior to another.
|
Project
Summary:
The study is limited to models that meet three criteria: 1) can it quantify human health (as opposed to ecological) risk; 2) is it applicable for site-specific
risk assessments; and 3) is it commonly used in Europe. The latter criteria allow commercial codes to be considered. Ten models that met these standards were
evaluated, including:
• CLEA (UK)
• JAGG (Denmark)
• P20 (UK)
• RBCA Toolkit (Commercial)
• RISC (Commercial)
• Risc-Human (Netherlands)
• ROME (Italy)
• SFT 99:06 (Norway)
• UMS (Germany)
• Vlier-Humaan (Belgium-Flanders)
The evaluation involved running the models for a generic (hypothetical) data set and five case studies where ground truth data, such as off-site monitoring
wells or on-site vapour probes, were available. Results for receptor point concentrations, dose and risk levels were compared and sensitivity to key input
parameters identified.
|
Professor Paul Bardos