A long-range dispersion model evaluation study with Chernobyl data

The Chernobyl accident outlined the need for improving the capability of the real-time estimate of long-range atmospheric dispersion and provided an opportunity to test numerical models against the radiological data collected over Europe. The APOLLO model, which became recently operational in the Accidental Release Impact Evaluation System (ARIES) at ENEA-DISP, has been tested based on meteorological and radiological data made available during and after the joint IAEA/CEC/WMO Atmospheric Transport Model Evaluation Study (ATMES). In the present paper a model evaluation and a sensitivity test with respect to the mixing depth and horizontal diffusion parameterizations are presented. The quantities involved in the validation are Cs-137 air concentrations paired in space and time, time-integrated concentrations and time of arrival of the cloud at each locality. The results show that the treatment of space and time variability of the mixing depth improves the model estimates, and that a linear trend in time of the horizontal dispersion coefficient gives better results than a square-root trend in terms of scattering between observed and computed values. (Author)