Studies on urban ventilation indicate that urban ventilation performance is highly dependent on urban morphology. Some studies have linked local-scale urban ventilation performance with the local climate zone (LCZ) that is proposed for surface temperature studies. However, there is a lack of evidence-based studies showing LCZ ventilation performance and affirming the reliability of using the LCZ classification scheme to demonstrate local-scale urban ventilation performance. Therefore, this study aims to analyse LCZ ventilation performances in order to understand the suitability of using the LCZ classification scheme to indicate local-scale urban ventilation performance. This study was conducted in Shenyang, China, with wind information at 16 weather stations in 2018. The results indicate that the Shenyang weather station had an annual mean wind speed of 2.07 m/s, while the mean wind speed of the overall 16 stations was much lower, only 1.44 m/s in value. The mean wind speed at Shenyang weather station and the 16 stations varied with seasons, day and night and precipitation conditions. The spring diurnal mean wind was strong with the speeds of 3.56 m/s and 2.21 m/s at Shenyang weather station and the 16 stations, respectively. The wind speed (2.21 m/s at Shenyang weather station) under precipitation conditions was higher than that (1.75 m/s at Shenyang weather station) under no precipitation conditions. Downtown ventilation performance was weaker than the approaching wind background, where the relative mean wind speed in the downtown area was only 0.53, much less than 1.0. The downtown ventilation performance also varied with seasons, day and night and precipitation conditions, where spring diurnal downtown ventilation performance was the weakest and the winter nocturnal downtown ventilation performance was the strongest. Moreover, the annual mean wind speed of the 16 zones decreased from the sparse, open low-rise zones to the compact midrise zones, indicating the suitability of using LCZ classification scheme to indicate local-scale urban ventilation performance. The high spatial correlation coefficients under different seasons, day and night and precipitation conditions, ranging between 0.68 and 0.99, further affirmed that LCZ classification scheme is also suitable to indicate local-scale urban ventilation performance, despite without the consideration of street structure like precinct ventilation zone scheme.
Feature Importance Analysis for Local Climate Zone Classification Using a Residual Convolutional Neural Network with Multi-Source Datasets