14:30 〜 14:45
[015] GIS-based urban characterization for urban wind prediction
キーワード：urban climate model, urban wind, microclimate
Rapid urbanization leads to urban feature changes, that also impacted to microclimate changes. These changes better to be monitored and understood, thus we may improve urban planning strategies and building design strategies in future. The vertical city weather generator (VCWG) is one of the latest urban microclimate models by integrating Monin-Obukhov similarity theory and vertical diffusion model to imitate climate process from meso- to micro-scale of urban climate. However, this model assumed the homogeneous urban settings to derive key urban morphological variable values. Instead, real urban contexts have variety building and street configurations that may affects the dynamics of wind circulation in surrounding area.
Thus, this paper aims to improve wind speed prediction by using actual urban characteristics on the basis of the geographical information system (GIS) data instead of using the hypothetical homogeneous urban settings. Urban morphology variables were reflected by calculating actual plan area density (𝜆𝑝) and frontal area index (𝜆𝑓) located at 233 urban areas in Seoul Metropolitan City, Republic of Korea. Results show that we could get more accurate 𝜆𝑝 and 𝜆𝑓 owing to GIS data. Prediction improvements were made in term of lower bias and lower RMSE compared to homogeneous-derived VCWG model, especially in summer period. However, the magnitude of improvements was limited and vary depends on location and season. Further research is identifying and improving the model’s performance based on other parameter values that affected VCWG model.
Thus, this paper aims to improve wind speed prediction by using actual urban characteristics on the basis of the geographical information system (GIS) data instead of using the hypothetical homogeneous urban settings. Urban morphology variables were reflected by calculating actual plan area density (𝜆𝑝) and frontal area index (𝜆𝑓) located at 233 urban areas in Seoul Metropolitan City, Republic of Korea. Results show that we could get more accurate 𝜆𝑝 and 𝜆𝑓 owing to GIS data. Prediction improvements were made in term of lower bias and lower RMSE compared to homogeneous-derived VCWG model, especially in summer period. However, the magnitude of improvements was limited and vary depends on location and season. Further research is identifying and improving the model’s performance based on other parameter values that affected VCWG model.