An urban large-eddy-based dispersion model for marginal gridresolutions: CAIRDIO v1.0

Abstract. The capability for high spatial resolutions is an important feature of accurate numerical models dedicated to simulate the large spatial variability of urban air pollution. On the one hand, the well established mesoscale chemistry transport models have their obvious short-comings attributed to their extensive use of paramterizations. On the other hand, obstacle resolving computational fluid dynamic models, while accurate, still often demand too high computational costs, to be applied on a regular and holistic basis. The major reason for the inflated numerical costs is the required horizontal resolution to meaningfully apply the obstacle discretization, which is most often based on boundary-fitted grids, like e.g. the marker-and-cell method. Here we present a large-eddy-simulation approach that uses diffusive obstacle boundaries, which are derived from a simplified diffusive interface approach for moving obstacles. The diffusive interface approach is well established in two-phase modeling, but to the author’s knowledge has not been applied in urban boundary layer simulations so far. Our dispersion model is capable of representing buildings over a wide range of spatial resolutions, including marginally coarse resolutions inaccessible for standard methods. This opens up a very promising opportunity for application of accurate air quality simulations and forecasts on entire mid-sized city domains. Furthermore, our approach is capable of incorporating the influence of the land orography by the additional optional use of terrain-following coordinates. We validated the dynamic core against a set of numerical benchmarks and a standard high-quality wind-tunnel data set for dispersion-model evaluation.