Naserikia Marzie, Hart Melissa A, Nazarian Negin, Bechtel Benjamin, Lipson Mathew, Nice Kerry A
Australian Research Council Centre of Excellence for Climate Extremes, University of New South Wales, Sydney, Australia.
Australian Research Council Centre of Excellence for Climate Extremes, University of New South Wales, Sydney, Australia.
Sci Total Environ. 2023 Dec 20;905:167306. doi: 10.1016/j.scitotenv.2023.167306. Epub 2023 Sep 22.
Due to the scarcity of air temperature (T) observations, urban heat studies often rely on satellite-derived Land Surface Temperature (LST) to characterise the near-surface thermal environment. However, there remains a lack of a quantitative understanding on how LST differs from T within urban areas and what are the controlling factors of their interaction. We use crowdsourced air temperature measurements in Sydney, Australia, combined with urban landscape data, Local Climate Zones (LCZ), high-resolution satellite imagery, and machine learning to explore the influence of urban form and fabric on the interaction between T and LST. Results show that LST and T have distinct spatiotemporal characteristics, and their relationship differs by season, ecological infrastructure, and building morphology. We found greater seasonal variability in LST compared to T, along with more pronounced intra-urban spatial variability in LST, particularly in warmer seasons. We also observed a greater temperature difference between LST and T in the built environment compared to the natural LCZs, especially during warm days. Natural LCZs (areas with mostly dense and scattered trees) showed stronger LST-T relationships compared to built areas. In particular, we observe that built areas with higher building density (where the heat vulnerability is likely more pronounced) show insignificant or negative relationships between LST- T in summer. Our results also indicate that surface cover, distance from the ocean, and seasonality significantly influence the distribution of hot and cold spots for LST and T. The spatial distribution for T hot spots does not always overlap with LST. We find that relying solely on LST as a direct proxy for the urban thermal environment is inappropriate, particularly in densely built-up areas and during warm seasons. These findings provide new perspectives on the relationship between surface and canopy temperatures and how these relate to urban form and fabric.
由于气温（T）观测数据稀缺，城市热研究常常依赖卫星反演的地表温度（LST）来表征近地表热环境。然而，目前仍缺乏对城市区域内LST与T之间差异以及二者相互作用控制因素的定量认识。我们利用澳大利亚悉尼的众包气温测量数据，结合城市景观数据、局部气候区（LCZ）、高分辨率卫星图像和机器学习，来探究城市形态和结构对T与LST相互作用的影响。结果表明，LST和T具有不同的时空特征，它们之间的关系因季节、生态基础设施和建筑形态而异。我们发现，与T相比，LST的季节变异性更大，且在城市内部的空间变异性更显著，尤其是在温暖季节。我们还观察到，与自然LCZ相比，建成环境中LST与T之间的温差更大，尤其是在温暖的日子。与建成区相比，自然LCZ（树木大多密集且分散的区域）显示出更强的LST - T关系。特别是，我们观察到建筑密度较高的建成区（热脆弱性可能更明显）在夏季LST与T之间呈现不显著或负相关关系。我们的结果还表明，地表覆盖、与海洋的距离和季节性显著影响LST和T的热点和冷点分布。T热点的空间分布并不总是与LST重叠。我们发现，仅依靠LST作为城市热环境的直接替代指标是不合适的，尤其是在高密度建成区和温暖季节。这些发现为地表温度和冠层温度之间的关系以及它们与城市形态和结构的关系提供了新的视角。
