Modelling concentration fluctuations and individual exposure in complex urban environments

One of the key problems in coping with deliberate or accidental atmospheric releases, which in many cases are short or/and result in high concentrations, is the ability to reliably predict the individual exposure during the event. Furthermore, for consequence assessment and countermeasures application, it is more realistic to rely on the maximum expected dosage rather than on the actual dosage. Recently, Bartzis et al. (2008) have introduced an approach relating maximum dosage to parameters such as concentration variance and turbulence integral time scale. The need for an estimation of these parameters poses new challenges to CFD models. In the CFD RANS model ADREA, new approaches have been implemented recently, where the parameterization of the dispersion of a pollutant emitted from a point source depends not only on the parameters of turbulence, but also on the pollutant travel times. In this study the new methodology is tested against MUST and FLADIS field experimental data, which consist of high resolution concentration time series enabling the production of short term dosage data. The present comparisons further strengthen the evidence that the applied methodology is capable of dealing properly with complex transient dispersion phenomena.