Spatio-temporal distribution and peak prediction of energy consumption and carbon emissions of residential buildings in China

With the acceleration of urbanization and the improvement of people's living standards, the carbon emissions of residential buildings (BCE) in China have been increasing, hindering the achievement of carbon peaking and carbon neutrality goals. It has become increasingly urgent and important to analyze and grasp the dynamic trends of BCE. Given the notable regional variations in carbon emissions, it is necessary to delve into the distribution characteristics of BCE and the underlying factors influencing them. First, based on the emission factor method, the energy consumption and carbon emissions calculation models are established. Next, the main influencing factors are obtained by improving Kaya's constant equation. Meanwhile, the GA-PPC-Kmeans model is established to divide BCE into different regions. Then, the STIRPAT-Ridge model is built to predict the peak carbon emissions. Finally, a combination of static and dynamic methods is used to forecast peaks under three scenarios, which are low-carbon (LC) scenario, baseline (BA) scenario, and high-carbon (HC) scenario. The results show that five pivotal factors significantly impact BCE in China, which are energy carbon emission intensity (ECEI), energy intensity (EI), economic density (ED), residential housing level (RHL) and population size(P). Furthermore, BCE is categorized into five distinct regions: low-carbon demonstration area, energy efficiency improvement area, carbon concentration and innovation area, high-carbon optimization area, and carbon poverty development area. Under BA scenario, the peaks for BCE in these regions converge primarily around 2040 and 2045. This study provides a regional research perspective for China's residential buildings to reach the peak of carbon emissions, and serves as a reference for formulating carbon emission reduction plans in different regions. [Display omitted] •The GA-PPC-Kmeans model is established to divide China's BCE into five regions.•The spatio-temporal distribution of BCE in China decreases from east to west.•The STIRPAT-Ridge model is established to forecast BCE.•Dynamic and static scenarios are combined for long-term forecasting.•The peak time of BCE in China is concentrated in 2040 and 2045.