Spatio-temporal structures of flow and pollutant in the atmospheric surface layer over real urban surface

Abstract

The air quality in cities, deteriorated by the complex urban geometries, has become one major threat to the well-being of citizens. Pollutant transport and fresh air entrainment are essential for promoting ground-level air quality, which is tightly associated with the atmospheric surface layer (ASL) developing just over the urban surface. However, the massive construction in cities alters the natural geometries that complicate the ASL flow. On the one hand, the streets and buildings slow down the prevailing wind and degrade the horizontal ventilation in the near-ground regions. On the other hand, the vertical transport of momentum and pollutants is governed by the ASL turbulence structures, which are primarily initiated by the urban architecture and surface. The in-depth investigation of the turbulence structures in real urban ASL helps to elucidate the momentum transport and pollutant dispersion mechanism within and benefit air quality management. In this connection, this study utilizes the Large-eddy Simulation (LES) to investigate the spatio-temporal turbulence structures in the real urban ASL (in both Hong Kong and ShangHai). In particular, the signals processing techniques, including wavelet analysis, empirical mode decomposition (EMD), and Hilbert-Huang transform, are implemented to study the multi-scale nature of ASL turbulence structures, including the energy spectrum, momentum and pollutant transport contribution from different turbulence structures, and the scale interaction between large- and small-scale eddies. Despite the difference in urban surface, large-scale motions, usually in the form of detached eddies, are long-lasting energetic eddies that contribute significantly to the momentum and pollutant transport. However, the small-scale turbulence structures, mostly initially by the buildings and attached to the solid surface, exhibit less transport efficiency while constituting a large portion of turbulence kinematic energy (TKE). Furthermore, the LSMs are found to modulate the amplitude of small-scale turbulence. Moreover, the presence of amplitude modulation, regardless of positive or negative, promotes both the TKE and momentum/pollutant transport on the small scales. Meanwhile, the proper orthogonal decomposition (POD), one data compression technique, is applied to the velocity fields to extract the dominant flow patterns in the real urban ASL. The traveling structures (detached eddies) are captured by the first two dominant modes that are up to kilometers in size. In addition, the first 80 out of 2000 modes extract more than 90% of the original TKE, indicating the great potential of POD in data compression for ASL flow. Finally, we tested the potential application of the predictive model in the real urban boundary layer, which aims to reconstruct the near-ground turbulence statistics by probing in the outer layer with pre-determined calibration procedures. The result illustrates the difficulty in predicting a specific location over the highly heterogeneous urban surface, while the ensemble-averaged statistics illustrate similar accuracy to the smooth-wall counterparts. The in-depth description of turbulence structures and advanced understanding of momentum/pollutant transport mechanisms benefit the practice of environmental impact assessment, real-time air quality monitoring, and sustainable city management, which can be easily extended to other megacities with high urbanization and dense buildings, such as Beijing, Tokyo, and Seoul.