Multi-factor dynamic correlation prediction and analysis of carbon peaking for building sector: A case study of Shaanxi province

Abstract

The factors influencing carbon emissions in the construction sector are numerous, and the relationships between these factors are complex. Previous studies on carbon peaking have often overlooked the dynamic changes between influencing factors and limited the number of variables to simplify the computation of predictive models. Based on the goal of carbon peaking, this study explores the relationships between internal factors within the construction industry and establishes a network of factor correlation. Furthermore, this network is embedded into an improved STIRPAT model, and a multi-factor dynamic correlation prediction model is constructed by incorporating scenario analysis. Taking Shaanxi Province, China, as a case for empirical analysis, the study explores carbon-peaking solutions for the building sector under different development scenarios. The findings indicate that carbon emissions in Shaanxi's building sector continuously increased during the study period, reaching 213 MtCO₂ in 2020. Through factor screening, 12 driving factors were found to be significantly related to carbon emissions, all showing positive correlations, with the urbanization rate contributing the most to emissions. The dynamic association prediction model constructed had an accuracy of 0.996. Using this model, nine carbon emission scenarios were predicted, with optimizing the energy structure identified as the critical pathway, achieving a 5.01% reduction in emissions. A comprehensive strategy could achieve a 12.49% reduction and meet the carbon peaking target. Finally, the study proposes policy recommendations for the coordinated management of emissions reductions in cities and the construction industry, contributing to the development of sustainable cities and societies.