Gaussian Mixture Model based classification for analyzing longitudinal outdoor thermal environment data to evaluate comfort conditions in urban open spaces

Abstract

Global warming and the urban heat island effect have led to a decline in the quality of urban open spaces. Although many studies have been conducted on the thermal environment of urban open spaces, there is a lack of a detailed analysis of the time dimension. This study is to propose a method for analyzing long-term time series data on thermal environment, to more comprehensively evaluate the comfort of urban open spaces. We employed thermal environment data from urban open spaces in Melbourne and conducted a time series comfort analysis using a Gaussian mixture model. The research findings indicate that urban open spaces surrounded by low-rise buildings exhibit the highest thermal comfort, whereas those encompassed by a combination of buildings and green spaces exhibit the lowest comfort level. More importantly, we discovered similarities in the comfortable time periods across the five types of urban open spaces, with the majority falling between 6:00 AM and 10:00 AM. These discoveries can assist in achieving a more objective evaluation of the thermal environment in urban open spaces and provide scientific guidance for urban planning and landscape design, thus enhancing the quality of life for urban residents.