An urban large-eddy-simulation-based dispersion model for marginal grid resolutions: CAIRDIO v1.0

Abstract

Abstract. The ability to achieve high spatial resolutions is an important feature for numerical models to accurately represent the large spatial variability of urban air pollution. On the one hand, the well-established mesoscale chemistry transport models have their obvious shortcomings due to the extensive use of physical parameterizations. On the other hand, obstacle-resolving computational fluid dynamics (CFD) models, although accurate, are still often too computationally intensive to be applied regularly for entire cities. The major reason for the inflated computational costs is the required horizontal resolution to meaningfully apply obstacle discretization, which is mostly based on boundary-fitted grids, e.g., the marker-and-cell method. In this paper, we present the new City-scale AIR dispersion model with DIffuse Obstacles (CAIRDIO v1.0), in which the diffuse interface method, simplified for non-moving obstacles, is incorporated into the governing equations for incompressible large-eddy simulations. While the diffuse interface method is widely used in two-phase modeling, this method has not been used in urban boundary-layer modeling so far. It allows us to consistently represent buildings over a wide range of spatial resolutions, including grid spacings equal to or larger than typical building sizes. This way, the gray zone between obstacle-resolving microscale simulations and mesoscale simulations can be addressed. Orographic effects can be included by using terrain-following coordinates. The dynamic core is compared against a standard quality-assured wind-tunnel dataset for dispersion-model evaluation. It is shown that the model successfully reproduces dispersion patterns within a complex city morphology across a wide range of spatial resolutions tested. As a result of the diffuse obstacle approach, the accuracy test is also passed at a horizontal grid spacing of 40 m. Although individual flow features within individual street canyons are not resolved at the coarse-grid spacing, the building effect on the dispersion of the air pollution plume is preserved at a larger scale. Therefore, a very promising application of the CAIRDIO model lies in the realization of computationally feasible yet accurate air-quality simulations for entire cities.