Bi-temporal Monitoring the Spatial Pattern and Variations of the Surface Urban Heat Island in three Chinese Coastal Megacities: A Comparative Study of Guangzhou, Hangzhou, and Shanghai.

The side-effect of booming urbanization on the ecosystem and climate system has been continuously exacerbating. The coastal metropolises are located at the interface between land and ocean, unavoidably influenced by multiple aspects of the terrestrial environments, aquatic ecosystems, and urban developments. Thus, the environmental health of coastal metropolis should be more concerned. In this study, targeting Guangzhou, Hangzhou, and Shanghai, an attempt was made to evaluate the spatiotemporal patterns and variations of surface urban heat island (SUHI) in three coastal metropolises of China based on Landsat-derived land surface temperatures (LST) and land cover data. The results indicate that overall, within a nearly 15-year interval, the extents of hot spots in three metropolises were significantly expanded, the spatial patterns of SUHI have been transformed from monocentric to polycentric high-LST clusters, which were identical to the trend of urban expansion. However,  these three metropolises possess distinct features in terms of the thermal layouts and land cover/use composition. Although the total area of SUHI hot spots in Shanghai has surged, the intensity of some hot spots has been a shrink. Besides, the interactions and associations between SUHI and urban development were investigated using spatial regression analysis. The urban composition and configuration considerably affected the intensity of SUHI. Terrain morphology constrained the SUHI. Prolific population growth had a continuing effect on SUHI formation. The proportion of forests displayed a consistently critical influence on easing the adverse of SUHI. Additionally, it is essential to appropriately consider the impacts of water in the comparative analysis of different thermal environments. However, water might be treated as a time-invariant factor and have a limited effect on the bi-temporal comparison for each metropolis. These findings suggest the policy-makers and urban planners should balance and optimize the land cover/use configurations with accommodating the increasing population, reasonably maximize the reservations of the greenbelt and green space under improving the utilization of urban infrastructures and constructions.