Wind speed estimation in urban areas based on the relationships between background wind speeds and morphological parameters

This study proposes a method for estimating wind speeds. The method determined the rate of change in wind speed associated with the morphological parameters in the urban or suburban districts by conducting multiple regression analyses. Then, it estimated wind speeds by combining the change rate and inflow wind speeds in Gangnam-gu (GD), Yangcheon-gu (YD), and Pyeongtaek City (PD). Compared to the local data assimilation and prediction system (LDAPS), the estimation method decreased the root mean square error (RMSE) of wind speeds at the automatic weather stations for 1 year (56.2% in GD and 25.6% in YD). For the selected period (4–13 Feb. 2017), the RMSEs in GD (0.66 ​m ​s−1) and PD (0.80 ​m ​s−1) of the wind speeds estimated by the proposed methods were comparable to, or even better than, those (0.72 ​m ​s−1 in GD and 1.00 ​m ​s−1 in PD) of the computational fluid dynamics (CFD) model coupled to LDAPS. The wind speed distribution estimated by the proposed method was reasonably well correlated with that simulated by LDAPS-CFD coupled model (R ​= ​0.69–0.92). We need to improve the method for districts with spatially uneven morphological parameters. [Display omitted] •A new method to estimate wind speeds in urban areas is proposed.•Urban morphological parameters are taken into account.•The proposed method well reproduced the measured wind speeds.•Estimated wind speeds are spatially well correlated with those simulated by a CFD model.