A global urban heat island intensity dataset: Generation, comparison, and analysis

• Published in: Remote sensing of environment 2024-10, Vol.312, p.114343, Article 114343
• Main Authors: Yang, Qiquan, Xu, Yi, Chakraborty, TC, Du, Meng, Hu, Ting, Zhang, Ling, Liu, Yue, Yao, Rui, Yang, Jie, Chen, Shurui, Xiao, Changjiang, Liu, Renrui, Zhang, Mingjie, Chen, Rui
• Format: Article
• Language: English
• Subjects: air temperature; anthropogenic activities; canopy; climate; data collection; environment; Estimation method; heat island; Rural definition; Spatiotemporal variation; Urban thermal environment; Urban warming trend; urbanization; weather
• ISSN: 0034-4257
• DOI: 10.1016/j.rse.2024.114343

The urban heat island (UHI) effect, a phenomenon of local warming over urban areas, is the most well-known impact of urbanization on climate. Globally consistent estimates of the UHI intensity (UHII) are crucial for examining this phenomenon across time and space. However, publicly available UHII datasets are limited and have several constraints: (1) they are for clear-sky surface UHII, not all-sky surface UHII and canopy (air temperature) UHII; (2) the estimation methods often neglect anthropogenic disturbance, introducing uncertainties in the estimated UHII. To address these issues, this study proposes a new dynamic equal-area (DEA) method that can minimize the influence of various confounding factors on UHII estimates through a dynamic cyclic process. Utilizing the DEA method and leveraging various gridded temperature data, we develop a global-scale (>10,000 cities), long-term (over 20 years by month), and multi-faceted (clear-sky surface, all-sky surface, and canopy) UHII dataset. Based on these estimates, we provide a comprehensive analysis of the UHII and its trends in global cities. The UHII is found to be greater than zero in >80% of cities, with global annual average magnitudes around 1.0 °C (day) and 0.8 °C (night) for surface UHII, and close to 0.5 °C for canopy UHII. Furthermore, an interannual upward trend in UHII is observed in >60% of cities, with global annual average trends exceeding 0.1 °C/decade (day) and over 0.06 °C/decade (night) for surface UHII, and slightly surpassing 0.03 °C/decade for canopy UHII. Notably, there exists a positive correlation between the magnitude and trend of UHII, suggesting that cities with stronger UHII tend to experience faster growth in UHII. Additionally, discrepancies in UHII are found between different temperature data, stemming not only from distinctions in data types (surface or air temperature) but also from differences in data acquisition times (Terra or Aqua), weather conditions (clear-sky or all-sky), and processing methodologies (with or without gap filling). Overall, our proposed method, dataset, and analysis results have the potential to provide valuable insights for future urban climate studies. The UHII dataset is publicly available at https://doi.org/10.6084/m9.figshare.24821538. •A new dynamic equal-area (DEA) method is proposed to estimate UHII.•Our UHII dataset spans over 20 years and covers 10,000+ global cities.•Our UHII dataset includes surface (clear-sky and all-sky) and canopy UHIIs.•T