Diversity and uniformity: An entropy-based diagnosis of sectoral roles in China’s embodied carbon flow network

Understanding the structural characteristics of carbon flows is important for designing effective climate mitigation strategies. This study proposes an entropy-based framework to analyze the intersectoral embodied carbon flow network of China, using the 2017 multi-regional input–output data. We decompose information entropy into two components: structural entropy, which captures the diversity of sectoral emissions, and interaction entropy, which reflects the balance of embodied carbon flows across intersectoral linkages. Results show that traditionally high-emission sectors, such as Power and Water and Metal Production, display both high structural entropy and high interaction entropy. In contrast, sectors such as Electronics and Nonmetal Products exhibit high structural entropy but low interaction entropy, indicating internally diverse emissions are not evenly distributed across intersectoral exchanges. In addition, sectors such as Food and Tobacco exhibit moderate structural entropy but relatively high interaction entropy, indicating that concentrated emission sources can still give rise to balanced carbon flows across intersectoral linkages. Uncertainty and robustness analyses further show that entropy metrics reliably explain emission responses to sector-specific changes and remain stable over time. These findings highlight the importance of entropy in identifying structural patterns of carbon exchange and supporting both centralized and decentralized decarbonization strategies.