The complex network transfer pathways and evolutionary patterns of embodied carbon emissions in China's agricultural industry Chain: An empirical analysis based on multi-node flow characteristics

Against the backdrop of the intertwining challenges of global climate change and agricultural sustainable development, agriculture serves not only as a fundamental industry but also as a significant source of greenhouse gas emissions. As a major agricultural nation, China still lacks a clear understanding of the carbon flow processes within its agricultural industrial chain, which hampers the formulation of targeted emission reduction strategies. To systematically reveal the transfer structure and evolutionary patterns of agriculture-related carbon emissions and support the low-carbon transition of the industry in alignment with the “dual carbon” goals, this study develops an integrated “input-output and complex network” coupling framework and conducts an analysis based on six years of input-output tables. Key findings reveal a three-phase evolution of agricultural carbon emissions (growth, peak, and plateau), with 71.55 % of emissions concentrated in four sectors: agriculture, food processing, food manufacturing, and fertilizer production. The construction (S31) and basic chemical raw materials manufacturing (S16) are identified as the primary sectors for embodied carbon inflow, while electricity and heat production and supply (S21) and petroleum and nuclear fuel processing (S15) serve as the core sectors for embodied carbon outflow. Food processing (S5) and specialty chemical products manufacturing (S17) play critical intermediary roles. The “fertilizer manufacturing → agriculture” pathway shows the highest carbon transfer volume, while “agriculture → food processing” remains a stable high-carbon-flow route. The “pesticide manufacturing → agriculture” pathway has intensified significantly since 2012. Notably, the livestock sector achieved reduced embodied carbon transfer despite production scale expansion. At the upstream stage of the industrial chain, agriculture (S1) and the food processing industry (S5) exhibit a clear transition in the carbon emission structure of the high-carbon sectors they drive during production—shifting from reliance on petroleum-based fuels toward electricity as the dominant energy source. At the downstream stage, driven by consumption demand from food-related industries, the resulting carbon emissions are highly concentrated in the paper and paper products industry (S29), accounting for more than 80 % of the total. Community detection identifies stable modular structures, revealing the systematic dependencies of embodied carbon flows. The study concludes by proposing enhanced technology adoption and optimized intermediate input management as key policy recommendations for agricultural carbon mitigation.