Sensitivity analysis of WRF-CFD-based downscaling methods for evaluation of urban pedestrian-level wind

Abstract

Computational fluid dynamics (CFD) techniques are widely adopted for predicting pedestrian-level wind (PLW). However, the lack of on-site measurement data is the primary impediment to establishing a reliable inflow wind profile. We propose a downscaling method that enables accurate modeling of PLW without the need for on-site measurements. The downscaling method involves three stages of Weather Research and Forecasting (WRF)-CFD simulations conducted in Meteodyn software. The WRF model is utilized to generate a time series of mesoscale data of mesoscale cells covering the microscale domain. The microscale CFD model consists of two nested CFD models, a parametric model and a full-information model, to ensure a smooth transition of the downscaled information. The physical-statistics method is employed to couple the mesoscale and microscale wind flow information. The sensitivity of the 6 downscaled schemes with different configurations is evaluated. To eliminate the effects of high-rise buildings, 3 potential mesoscale data heights are examined as inputs for the CFD simulations. The accuracy of the proposed downscaling method is validated using long-term on-site measurement data. We recommend utilizing mesoscale data at a height of 200 m as an input to the CFD model for PLW modeling in complex urban environments.