Urban stations used in previous urban warming studies were mostly located in urban fringes (i.e., newly urbanized areas) rather than located in urban cores. However, these studies focused more on urbanization contribution to regional warming rather than on warming within cities. It is estimated that urbanization-induced warming accounts for 20% to 50% of the overall observed warming in areas that have experienced fast urbanization 13, 14. This approach has been applied to China where rapid urbanization has occurred in the past decades and where the density of weather station network is high 13. Some studies have examined the equal-rate assumption by isolating the urban warming signal from surface air temperature (SAT) data, i.e., by directly separating contributions of urbanization from other forcing factors based on linear or non-linear models 11, 12. Furthermore, although the urban effect has been considered in projections of heat exposure in the future 9, many studies make an implicit assumption that urban temperatures will increase at an equal-rate as rural temperatures 6, 10.

Despite the prevalence of the UHI and the increasing recognition of the need for climate monitoring in urban environments 3, a great majority of the assessments of heat-related mortality 6 and loss of workplace productivity 7 in cities are still based on temperature data collected by non-urban and peri-urban weather stations 8. This problem will become more severe in the future because of global climate change and urban population growth in cities 4, 5. Urban residents can experience greater heat exposure during heatwave events than the general population because of the urban heat island (UHI) 1, 2, the phenomenon of higher temperatures over urban land than over the surrounding rural land 3.

Global cities