The use of computers in emergency situations and the choice of dispersion model

ABSTRACT

The range of atmospheric dispersion models available for use in emergency situations varies from simple straight Gaussian plume simulations to complex computer codes attempting to allow for site-specific topographical features where the restriction of running in real time becomes significant. Various factor influence the choice of model; there are rather different needs and constraints for (a) training exercises, (b) early, (c) intermediate and (d) recovery phases of an actual accident. In selecting appropriate dispersion models it is necessary to define the objectives - for example what spectrum of accidents is to be considered and what surrounding area, and how will the model be linked with a monitoring programme (specifically the availability of early measurements and the evaluation of source terms). The model must predict the quantities of interest to the user, and results must be displayed in an unambiguous and convenient form. Thus spatial distributions at a specific time may be an asset as well as integrated doses, and deposited activity (dry and wet) as well as air concentrations. There will also be site-specific questions related to the complexity of the environs and the availability of extra meteorological information (e.g. vertical wind soundings or precipitation data). Nor must the constraints of time and resources and simplicity for the user be forgotten. (AU)