Assessing the amount of accidental releases from dose rate values

Inverse simulation methods allow a rough estimate of the source term to be made from the numerous measured values and from the use of an atmospheric dispersion model. When a nuclear accident occurs, measurement campaigns are quickly launched that allow the knowledge of dose rate values in different parts of the territory and their evolution over time. A more accurate inverse method has been developed by IRSN (French Institute for Radiation protection and Nuclear Safety). The strategy of this method is to constrain the inverse problem in order to make it converge to an optimal source term. This method relies on 3 steps: first, determining all the nuclides that contribute most to the dose rate signal. For instance in the case of the Fukushima accident about 10 nuclides were selected. Second step: identifying on a time scale the release periods and the third step is a mere calculation step that gives for each nuclide the amount of hour-by-hour releases that represent best the measured values. The method has been applied to Fukushima accident data, the results are acceptable but including other measured values than dose rates would be beneficial. (A.C.)