Uncertainties of atmospheric dispersion calculations for emergency preparedness

Date: 22.10.2015

Time: 13:30 - 14:30

Place: Nordlys, DMI

Contact: Cathrine Fox Maule

Disaster, Death and Destruction

Atmospheric dispersion model calculations for radioactive isotopes released from a nuclear accident provide information on possible contamination levels and radiation hazards thereby facilitating decisions on protective actions. This is implemented for emergency management through the use of Decision Support Systems (DSSs). 

Recent developments in numerical weather prediction modelling include probabilistic forecasting techniques addressing the inherent uncertainties. This approach may be taken over by atmospheric dispersion modelling. Today, however, most current DSSs do not take uncertainties into account, but merely allow for presentation of a single deterministic plume dispersion or deposition pattern.

In the research project ‘Meteorological Uncertainty of atmospheric Dispersion model results’ (MUD), the uncertainties of atmospheric dispersion model calculations are investigated as well as means for  incorporating the uncertainties into DSSs,  allowing for the presentation of uncertainties to decision makers in a comprehensible manner. The MUD methodology has been implemented operationally in the Danish setup providing long-range atmospheric dispersion modelling for the Danish Emergency Management Agency (in Danish: Beredskabsstyrelsen).

In the current project ‘Fukushima Accident: UNcertainty of Atmospheric dispersion modelling’ (FAUNA), the MUD methodology is applied to the Fukushima Daiichi nuclear accident, and the influence of meteorological uncertainties on real-time assessments of atmospheric dispersion and deposition is being investigated, imitating real-time emergency management. The objective of the project is to examine how uncertainty estimates can be presented to experts as well as to decisions makers in a manner that meets requirements of both the experts using the DSS and the decision makers relying on practical decision support.

In the seminar, the methodology developed in MUD will be described, and results of MUD and FAUNA presented.