Spatiotemporal evolution, regional disparity, and driving factors of county-level rice production carbon efficiency: A case study of Jiangxi Province, China.

Accurately quantifying the carbon efficiency of rice production (RCE) and elucidating its spatiotemporal evolution, regional disparities, and driving factors hold significant theoretical and practical implications for advancing agricultural green transformation and achieving sustainable development. Utilizing panel data from 85 counties in Jiangxi Province, China (2012-2022), this study employs a super-efficiency slack-based measure (Super-SBM) model incorporating undesirable outputs to estimate RCE. Spatial visualization via ArcGIS, kernel density estimation, Theil index decomposition, and geographical detector are applied to explore spatiotemporal patterns, regional heterogeneity, and driving mechanisms. The findings reveal that: (1) RCE exhibits a fluctuating upward trend with dynamic convergence characteristics, yet substantial improvement potential remains relative to the optimal production frontier. (2) A "central-high, peripheral-low" spatial distribution pattern dominates, accompanied by significant spatial autocorrelation and stable agglomeration features. (3) The overall Theil index initially declines before rising, with intra-regional disparities constituting the primary contributor to total differences. (4) Spatial differentiation is jointly driven by industrial and input-level factors, with distinct dominant drivers and interaction types across regions. Accordingly, we recommend formulating region-specific low-carbon policies, prioritizing key drivers, and enhancing multi-factor synergistic effects to achieve balanced regional development and facilitate agricultural green transformation.