Methods for radiological assessment in the near-field during the early phase of an accidental release of radioactive material, using an incomplete database

Following a major accidental release of radionuclides to the atmosphere, appropriate countermeasures must be taken to mitigate the consequences to the population and the environment. Time constraints will force the decision making process to be initially model based. Poor knowledge of the main model parameters can lead to erroneous and cost-ineffective decisions. To circumvent these difficulties, a real-time assimilation database containing monitoring and accident-specific data is proposed. Comparison of this dynamic database with a dispersion/deposition model and the application of physical principles can help to reduce the uncertainty in the early radiological assessment. Methods based on fuzzy set theory to infer information from quantities that are imprecise, incomplete or not totally reliable are proposed. Practical steps to implement an uncertainty reduction module in the European emergency decision support system RODOS are discussed.