Divergent urbanization-induced impacts on global surface urban heat island trends since 1980s

• Published in: Remote sensing of environment Vol. 295; p. 113650
• Main Authors: Li, Long, Zhan, Wenfeng, Hu, Leiqiu, Chakraborty, TC, Wang, Zhihua, Fu, Peng, Wang, Dazhong, Liao, Weilin, Huang, Fan, Fu, Huyan, Li, Jiufeng, Liu, Zihan, Du, Huilin, Wang, Shasha
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.09.2023; Elsevier
• Subjects: Data reconstruction; ENVIRONMENTAL SCIENCES; Land surface temperature; Landsat data; Surface urban heat island; Thermal remote sensing; Landsat data; Thermal remote sensing; Land surface temperature; Data reconstruction; Surface urban heat island
• ISSN: 0034-4257; 1879-0704
• DOI: 10.1016/j.rse.2023.113650

Urbanization experiences different speeds and forms under diverse development stages across the globe. However, urbanization-induced impacts on long-term surface urban heat island intensity (Is) trends across global cities and the regulators of such impacts remain understudied Here we estimate interannual trends in daytime Is (i.e., urban-rural differences in surface temperatures) across 511 major cities for 1985–2020 using annual averages calculated using reconstructed land surface temperature data derived from >250,000 Landsat thermal images. Our study reveals that the global mean Is growth rate is 0.156 °C/decade. We further examine Is change associated with per 1% impervious land growth (denoted as β) in each city throughout the research period and during different periods. The global mean β is 0.018 ± 0.025 °C/% (mean ± 1 standard deviation) for the whole period, with greater values in humid than in arid climates; and the β may change during different periods, e.g., it has more than tripled when urban impervious land exceeds 30%, indicating the spatiotemporally divergent impacts of urbanization on Is trends across global cities. The spatial variations in β across global cities are well correlated with rural vegetation abundance and precipitation but not with urban population. Among these three factors, rural vegetation abundance possesses the greatest standardized regression coefficient of partial least-squares model, signifying the critical role of biome background in regulating β. The finding implies that future urbanization over densely vegetated regions should be more carefully and strategically planned due to the greater urbanization-induced surface warming effect. •Landsat land surface temperatures across global 511 cities were reconstructed.•Global surface urban heat island trends since the 1980s were analyzed.•Urbanization-induced impacts on surface urban heat island trends were evaluated.•Urbanization-induced impacts are greater in cities with humid climates.•Rural vegetation abundance regulates urbanization-induced impacts.