Ordinary least squares modelling of urban heat island intensity based on landscape composition and configuration: A comparative study among three megacities along the Yangtze River.

• The effects of landscape composition on LST increased from 2000 to 2019. • The effects of landscape configuration on LST differs in different cities. • Reorganizing landscape configuration was applicable in relieving UHI. • The distributions of SUHI in three cities are roughly parallel to the Yangtze River. The urban heat island (UHI) phenomenon, arising from rapid urbanization, has attracted both social and academic attention because of its effects on citizens' daily life. This paper investigated the relationship between landscape composition and configuration, population and land surface temperature (LST) in three megacities along the Yangtze River, aiming at providing useful information to reduce the effect of UHI. A correlation analysis and ordinary least squares (OLS) models were used to discover the relationship between these indices. The results showed that the effects of landscape composition, especially impervious surfaces (IS), on the LST increased obviously from 2000 to 2019. The correlation coefficients of the density of IS increased from 0.608 to 0.839 in Nanjing, from 0.368 to 0.777 in Wuhan, and from 0.416 to 0.874 in Chongqing, respectively. On contrast, the effects of landscape configuration on LST differs in different cities. In Nanjing, small and scattered patches of green spaces showed the highest capacity in cooling the surrounding areas. In Wuhan, small patches of waterbodies owned the highest cooling effect. While in Chongqing, small and scattered patches of forest with irregular shapes showed the highest cooling effect. The findings proved that reorganizing landscape configuration was applicable in relieving UHI. [ABSTRACT FROM AUTHOR]