Spatial correlation network structure and influencing factors of agricultural ecological efficiency under carbon constraints in China

Abstract

Agricultural ecological efficiency (AEE) is essential for coordinating agricultural development with carbon reduction and ecological sustainability. This study evaluates the spatiotemporal evolution of AEE in 31 provincial-level units in China from 2007 to 2024 under carbon constraints. A super-efficiency slacks-based measure (SBM) model incorporating undesirable outputs is used to measure AEE, while a modified gravity model and social network analysis (SNA) are employed to examine the interprovincial spatial correlation network. A spatial Durbin model (SDM) is further applied to identify the direct and spillover effects of key influencing factors.The results show that China’s overall AEE improved during the study period, although substantial regional heterogeneity persisted. The national mean increased from 0.738 in 2007 to 0.952 in 2024. The eastern region experienced the most pronounced improvement, whereas the northeastern region maintained the highest period average. The central and western regions also improved, particularly during the later years. Spatially, the provincial distribution of AEE evolved from a mixed pattern of low, medium, and high efficiency in 2007 to a predominantly high-efficiency pattern in 2024. The kernel density analysis further confirms an overall upward shift in AEE in recent years. However, the remaining low-efficiency tail and persistent regional differences indicate that improvements were uneven across provinces. From 2007 to 2024, the spatial correlation network of China’s provincial AEE remained fully connected. The interprovincial AEE network remained fully connected, with a slight increase in network ties and density and a substantial decline in network hierarchy. Henan consistently occupied a prominent position, while several provinces, including Shandong, Shaanxi, Hubei, Ningxia and Gansu, performed important connecting or bridging roles. Economic development, urbanization, education, informatization, rural residents’ income, and technological innovation all exerted significant positive direct effects on AEE. Informatization generated a positive spatial spillover effect, whereas economic development produced a negative spatial spillover effect; the other factors showed no significant spillover effects. These findings provide empirical evidence for designing spatially targeted and regionally coordinated policy frameworks to advance the sustainable enhancement of AEE, with broader implications for agricultural climate governance in developing economies undergoing rapid urbanization.