Fog modelling during the ParisFog campaign : predictive approach and spatial heterogeneity effect

In fog or low clouds modeling, the accurate comprehension of the interaction among the turbulence, the microphysics and the radiation is still an important issue in improvement of numerical prediction quality. The improvement of fog modeling is important both in forecasting in transportation and in industrial domain by reason of their discharges atmospheric (cooling tower, smog . . . ). The 1D version of Code_Saturne has been used for the numerical simulation with the observational data from the ParisFog campaign, which took place at the SIRTA site during 2006-2007 winter. The comparison between the simulation and observation shows that the model is able to reproduce correctly the fog evolution from its formation to its dissipation. The sensitivity analysis of the behavior of the different parameterizations shows that the fog dynamic is sensible to the turbulence closure, the fog water content to the sedimentation processes and the fog droplet spectrum to the nucleation scheme. The performance of a long-period simulation in forecasting mode shows that the robustness of the model and the contribution of the coupling by nudging and a mesoscale model in 36 hours advance. The 3D version of Code_Saturne allows us to study the effect of spatial heterogeneity on the fog formation. Firstly, the simulations have been performed within a homogeneous horizontal domain with RANS mode. And then, the surface roughness in different type of surface and the building area will be taken into account.