Role of land use in China’s urban energy consumption: based on a deep clustering network and decomposition analysis

Land use can affect energy consumption by changing the economic and social structure of cities. Thus, the optimization of land use patterns is key to promoting energy sustainability. In this study, we explored the spatiotemporal evolution of China’s urban energy consumption and its driving factors from the role of land use, with the application of high-precision nighttime light images and land use data acquired from remote sensing satellites. A deep clustering network in deep learning was used for clustering analysis of urban energy consumption. The results indicated that the economic and structural effects of land use were the primary driving factors of the increasing urban energy consumption, whereas the decrease in the energy intensity (caused by technological progress) restrained the growth rate of energy consumption. With the exception of economically developed cities, generally, the contribution of the population size to the temporal increase in energy consumption was relatively small. The spatial difference in urban energy consumption was mainly due to the between-group differences among the diversified cluster groups, which were strongly influenced by land urbanization and population size. These conclusions can help the Chinese government formulate differentiated urban energy policies.