Comparison of the ESTE CBRN Model with the Joint Urban 2003 Experiment

Abstract

To model urban airflow and dispersion near buildings and in street canyons, several Gaussian diffusion models, and diagnostic urban models and methods based on computational fluid dynamics (CFD), are developed to provide a fast and adequate response when hazardous material is released. The ESTE CBRN software tool computes the urban atmospheric variables using the Reynolds-averaged Navier–Stokes equations with the k–ε closure model. The dispersion is modelled by an Eulerian model linked to the calculated flow field. The ESTE CBRN results are compared with the Joint Urban 2003 Street Canyon Experiment for the instantaneous puff releases conducted within the Urban Dispersion International Evaluation Exercise project. Urban airflow is simulated with a steady-state approach where the time-averaged velocity is evaluated from local measurements conducted outside downtown Oklahoma City, USA. Twenty anemometer measurements inside the downtown area are used to verify the CFD calculation, after which 22 individual puff releases are considered. The modelled velocity and sampler responses are in moderate agreement with the measurements. Many compared variables, such as the mean wind speed and puff arrival time, are generally well reproduced, and fulfil the urban modelling criteria. The turbulent kinetic energy is, in general, underestimated by 30–40%. The modelled time series of sampler responses fulfil many, but not all, of the urban criteria, mainly due to differences in puff trajectories and a lower dispersion intensity. Additionally, the impact of applying various three-dimensional meshes on the predicted sampler responses is tested.