The development and validation of the Inhomogeneous Wind Scheme for Urban Street (IWSUS-v1)

Abstract

Abstract. The layout of urban buildings shows significant heterogeneity, which leads to the significant spatial inhomogeneity of the wind field in and over the canopy of urban street canyons. However, most of the current urban canopy models do not fully consider the heterogeneity of the urban canopy. Large discrepancies thus exist between the wind speeds simulated by the current urban canopy models and those observed in the street canyon. In this study, a parameterization scheme for wind fields, Inhomogeneous Wind Scheme for Urban Street (IWSUS), is developed to better characterize the heterogeneity of the urban canopy. We use a computational fluid dynamics method to generate the IWSUS scheme and compare it with observations of the wind profile and turbulent flux in and over the street canyon for validation. In IWSUS, the wind speed vertical profiles at six representative positions located in a typical street canyon (i.e., the windward or leeward side of a long straight street or the inflow or outflow end) are parameterized separately. The wind profile by IWSUS thus can better describe the horizontal heterogeneity of the urban near-surface wind field, e.g., the dynamic drag effect of buildings in the lower atmospheric layer over the urbanized land use. The validation based on observations shows that the performance of simulation results by IWSUS is better than that by the exponential–logarithmic (exp-log) law widely used in the current urban schemes. We consider typical building arrangement and specific street orientations in IWSUS for wind field simulations, which can better match the distribution characteristics of street canyons around the observation point in the street canyon. The averaged wind profiles and turbulence energy fluxes in the model grids of urban areas by IWSUS are also nearer to the observations than those by the exp-log law. The normalized mean errors (NMEs) between the simulated and the observed vertical average wind speed are 49.0 % for IWSUS and 56.1 % for exp-log law in the range from the ground to 4 times the average height of the buildings and 70 % for IWSUS and 285.8 % for exp-log law in the street canyon (range from the ground to building top). This study proves that the accuracy of simulations of land surface processes and near-ground meteorological processes over the urban canopy can be improved by fully considering the heterogeneity of the urban canopy layout structures and the inhomogeneity of wind field distributions in and over the street canyon. IWSUS is expected to be coupled with mesoscale atmospheric models to improve the accuracy of the wind field, land surface energy budget, meteorological and atmospheric chemistry simulations.