Stochastic model of radionuclide migration from natural and engineered repositories

The overall stochastic model includes two types of evaluations. The first involves the identification of the sequence of events which lead to the repository conditions necessary for the initiation of radionuclide migration and evaluation of their probabilities of occurrence. The second considers the radionuclide transport into the surrounding media as a function of time and space. The seven major steps in the current model are: geology and hydrology; repository attributes; probability of distribution of attributes; event paths; probability of time to event; updating attribute probabilities, and simulation of sequential events. After the occurrence of each extreme event, some of the repository characteristics will change. Thus, an important provision in the model is the continuous updating of probability distributions of attributes. The Bayesian statistical techniques provide the necessary tools for this modification of parameter probabilities. Currently computerized procedures are being tested for the event path model. Their sensitivity are being evaluated using partial data from existing natural and inadvertent repositories. The event path model forms the basis for the radionuclide migration model. At any given time period of forecast, migration or retention of radionuclides will depend on the respository characteristics altered by the occurrence of extreme events which have taken place during that time. Thus, a migration model has to be considered in conjunction with the event path model. This aspect of the overall stochastic model is currently under investigation