Carbon emissions and government interventions in urban agglomerations of China: An integrated GWR and neural network approach

Abstract

The dual-carbon target that aims to achieve peak carbon and carbon neutrality before 2060 is a pivotal strategy for China's green and low-carbon development. As major contributors to China's economy and its carbon emissions, urban agglomerations play a critical role in reducing carbon emissions, and government interventions have attracted considerable attention. In this context, this study focuses on the spatiotemporal nonstationary characteristics of factors influencing carbon emissions in 19 urban agglomerations in China. Using the geographically weighted regression (GWR) model and its extensions, we analyzed the impacts of two government intervention factors—fiscal expenditure and green cover—on carbon emissions. A comparison of multiple models revealed that the geographically and temporally neural network weighted regression (GTNNWR) model best captures the spatiotemporal nonstationary relationships. Our findings indicate that increased government fiscal expenditure generally reduces carbon emissions, with stronger effects in the northern cities we studied. Urban green cover has significant negative impacts in the core cities of most urban agglomerations. However, these impacts may reverse in a very small portion of cities, possibly due to differences in development stages. The results provide insights for the government to formulate carbon reduction strategies.