Introducing and evaluating a new building-height categorization based on the fractal dimension into the coupled WRF/urban model

Model simulations of the boundary-layer structures and summer rainfall in cities are heavily affected by urban morphology, which are defined by a number of urban canopy parameters (UCPs). In the default Weather Research Forecasting (WRF) model/urban canopy model (UCM) system, UCPs are specified in a lookup table as function of three different urban land-use types, which cannot totally capture the heterogeneity in building heights (BHs). In this article, a method of using a fractal dimension (FD) was developed to define new categories of urban BH for characterizing the vertical urban canopy morphology. Combined with the default three urban land-use categories in WRF/UCM system, the BH categorization improves the description of the urban canopy characteristics. WRF/UCM multi-day simulations showed that the new method produced lower air temperatures and higher 10-m wind speeds in densely urbanized areas in Beijing, and performed better in capturing the temperature variations and spatial distributions of the 10-m wind speed. A rainfall case was also simulated by WRF/UCM using the two methods, and the new BH method yielded a more accurate simulation of precipitation. The difference of 24-h total precipitation between WRF simulations using the two methods can reach 40 mm, indicating a high sensitivity to urban surface representations. Results showed that the introduction of FD-based BH categorization is a simple and yet efficient way to improve WRF/UCM performance in the areas where detailed information on BHs is not available.