Multi-scale impacts of 2D/3D urban building pattern in intra-annual thermal environment of Hangzhou, China

• Published in: International journal of applied earth observation and geoinformation Vol. 104; p. 102558
• Main Authors: Lu, Huimin, Li, Fei, Yang, Gang, Sun, Weiwei
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15.12.2021; Elsevier
• Subjects: autumn; Hangzhou; heat island; Land surface temperature; Machine learning; Multi-dimensional urban morphology; regression analysis; spatial data; spring; summer; surface temperature; temporal variation; winter; Multi-dimensional urban morphology; Land surface temperature; Hangzhou; Machine learning
• DOI: 10.1016/j.jag.2021.102558
• ISSN: 1569-8432

•The spatial dependency variations of seasonal thermal environment are investigated.•The relative contribution variations of 2D/3D building metrics are revealed.•The threshold values of dominant building metrics are discovered.•The sensitivity of dominant metrics to analytical scales are examined. Understanding the relationship between multi-dimensional architectural pattern and land surface temperature (LST) is vital for alleviating the urban heat island. Although prior researches have showed that 2D/3D architectural pattern can significantly affect LST, the comprehensive effect of 2D/3D architectural pattern on thermal environment is still controversial and obscure due to the impact of architectural pattern on LST would vary under different observational scales and seasons, making it necessary to examine the temporal variation and scale-dependent of the relationship between them. This study takes Hangzhou as a study area to probe the relationship between seasonal LST and 2D/3D building metrics across ten analytical scales. Major findings include: (1) From spring to winter, the spatial dependency of LST becomes weaker with the increasing observational scale. (2) The regression analysis results show that the mean architecture projection area (MAPA), building coverage ratio (BCR), floor area ratio (FAR) and mean architecture height (MAH) are the most dominant metrics, with the range of average relative contributions are 10.96–16.68%, 14.61–45.20%, 7.90–17.05% and 5.80–14.23% in four seasons. Moreover, the relative contributions of chosen metrics on seasonal LST exhibit consistency in spring, summer and autumn. (3) The marginal effect analysis results reveal that the threshold values of these four dominant metrics are sensitive to analytical scale and would reduce with the analytical unit increases in four seasons. These findings indicate that controlling the building density, building height and floor area ratio of built-up areas in a reasonable range according to the spatial scale of planning unit, advocating more dispersed arrangement of the buildings and smaller architectural base area in Hangzhou can be favorable in ameliorating thermal environment.