Atmospheric dispersion using a Lagrangian stochastic approach: Application to an idealized urban area under neutral and stable meteorological conditions

Abstract

We present an adaptation of the Lagrangian stochastic dispersion model of the computational fluid dynamics (CFD) open source code Code_Saturne to simulate atmospheric dispersion of pollutants in complex urban geometries or around industrial plants. The wind is modeled within the same code with an Eulerian RANS (Reynolds-averaged Navier-Stokes equations) approach and thus involves the solution for the ensemble-mean velocity field and turbulent moments, using eddy viscosity or Reynolds stress turbulence models adapted to the atmosphere and complex geometries. The Lagrangian stochastic model used for the dispersion of the particles within this flow field is the simplified Langevin model, which pertains to the approaches referred to as PDF (Probability Density Function) methods. This formulation of model has not been widely used in atmospheric applications, despite interesting theoretical and computational benefits. Therefore, its usage must be validated on different atmospheric cases. In this paper, we present the validation of the model with a field experiment, considering atmospheric stratification and buildings: the MUST (Mock Urban Setting Test) campaign, conducted in Utah’s desert, USA.