Identifying critical drivers of transportation carbon emissions: An integrated DEMATEL-random forest approach

The significant impact of greenhouse gases on global warming has drawn widespread attention. This study focuses on the development of the transportation sector and energy consumption across 30 provinces in China from 1997 to 2022, aiming to identify the key drivers of carbon emissions in China’s transportation sector and analyze their causal interactions and spatial heterogeneity. Initially, provincial carbon emissions are estimated based on reallocated energy consumption data. A random forest model is then employed to objectively screen key factors from multidimensional variables. Subsequently, the Decision-Making Trial and Evaluation Laboratory (DEMATEL) approach is utilized to reveal the interaction network among these factors, distinguish their causal attributes, and explore their inter-provincial spatial differentiation. The findings are as follows: (1) Expenditure on research and experimental development, Number of registered scientific and technological achievements, and Total energy consumption are the most crucial factors influencing emissions; (2) Total energy consumption, Green coverage rate of built-up area, and Urbanization level serve as the primary causal drivers within the system; (3) The same factor exhibits significant variations in causal attributes across different provinces, reflecting regional heterogeneity in development stages. This study provides empirical evidence and methodological support for formulating differentiated and precise traffic carbon reduction policies.