In this activity, the ensemble-statistical methodology developed in the NKS-B activity MUD will be applied to the Fukushima Daiichi NPP accident. The activity addresses real-time forecasting of atmospheric dispersion and deposition of radionuclides released from a nuclear installation taking into account the meteorological uncertainties. Under certain weather conditions, these uncertainties can be large, and, as demonstrated in MUD, consequently also the uncertainties of the dispersion model results.
The objective of the study is to apply the MUD methodology to a realistic setting of the Fukushima accident, and to investigate the implications of the uncertainty estimates for the emergency management. Thus, for the first time a study will be carried out on the influence of meteorological uncertainties on real-time assessments of geographical areas affected by radioactivity from the Fukushima accident.
A meteorological ensemble forecasting system will be set up on DMI’s supercomputer for the period of concern and for a geographical domain covering Japan and surroundings. For the full period, two-day meteorological forecasts will be generated four times a day, as would be the case in real time.
For selected dates and times in the release period, the Danish and the Norwegian long-range atmospheric dispersion models will be run based on data of the meteorological ensemble assuming that a realistic source term is available in near real time. Corresponding ensemble-statistical parameters will be calculated, e.g. percentiles of the concentration and deposition fields. The predictions will be made available to the ARGOS decision-support system for display and dose modelling. Thereby, the activity will imitate real-time emergency management taking into account estimates of the uncertainty of dispersion model results.
Finally, the results of the activity will be described in the form of an NKS report as well as published in a peer-reviewed journal.