Spatio-temporal association of fossil fuel CO2 emissions from crop production across US counties

• We study the distribution of CO2 emissions from crop production across US counties.
• Significant global/local spatial autocorrelations are found in crop CO2 emissions.
• Decomposition method reveals the origin of spatial differences in crop productivity.
• Variance of crop productivity is sensitive to output/area and to spillover effects.
• Spatial dependence matters for CO2 emissions measurements and mitigation policies.

This article offers a spatio-temporal analysis of the distribution of CO2 emissions, the main cause for greenhouse gases, due to agricultural activities across US counties. Based on a novel database, we investigate how crop production output (measured in carbon) relates to CO2 emitted in the production and transportation process of the inputs needed for crop production. Various spatial statistics are used to highlight the clusters of counties with similarities in the levels and growth of output per area, input per area and productivity. At the same time, significant levels of heterogeneity are highlighted for all variables. A decomposition method allows us to uncover that the origin of interregional differences in productivity differs across the clusters of counties. Our results indicate that future mitigation policies should not fail to recognize interregional differences in the location, spatial extent and origin of carbon emissions due to the crop production process.