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 atmospher e. appropriate countermeasures must be taken to mitigate the consequen ces to the population and the environment. Time constraints will force the decision making process to be initially model based. Poor knowled ge of the main model parameters can lead to erroneous and cost-ineffec tive decisions. To circumvent these difficulties. a real-time assimila tion database containing monitoring and accident-specific data is prop osed. Comparison of this dynamic database with a dispersion/deposition model and the application of physical principles can help to reduce t he uncertainty in the early radiological assessment. Methods based on fuzzy set theory to infer information from quantities that are impreci se, incomplete or not totally reliable are proposed. Practical steps t o implement an uncertainty reduction module in the European emergency decision support system RODOS are discussed.