Three-Layer Nested Wrf-Cfd-Based Downscaling Method for Accurately Modeling Pedestrian Wind in an Urban Context

Abstract

Computational fluid dynamics (CFD) techniques have received widespread acceptance in predicting pedestrian wind. The primary impediment to establishing a reliable inflow wind profile is the lack of on-site measurement data. We propose a downscaling method for accurately modeling pedestrian wind in the absence of on-site measurement data. The downscaling method consists of three layers of nested WRF (Weather Research and Forecasting)-CFD numerical simulations. The WRF model in layer one was used to acquire mesoscale data across the study site and to define the boundary conditions for the nested microscale domains. Another two layers of CFD calculations were performed at microscales via parametric modeling and full-information modeling to ensure a smooth transition of the downscaled information. To eliminate the effects of high-rise buildings, three potential mesoscale data heights were examined as CFD inputs. The accuracy of the downscaling method was validated using long-term on-site measurement data. The microscale simulation results using mesoscale and measurement data as the inlet boundary conditions were compared, and recommended to employ mesoscale data at a height of 200 m as an input to the CFD model for pedestrian wind modeling in complicated urban contexts.

Keywords: Downscaling method, three-layer nested WRF-CFD model, Pedestrian wind, Urban context, On-site measurement validation