Classifying the potential for soil organic carbon gain under regenerative agriculture

Abstract Regenerative agriculture is pivotal for mitigating climate change, with no-tillage practices on cropland being generally effective at raising soil organic carbon (SOC). Yet, our understanding of the compound impact of soil and environmental factors on SOC gain potential after transitioning to no-till practices is still developing. Using imbalanced machine learning classification, here we quantify key thresholds to hierarchically classify SOC gain potential by switching from conventional tillage to long-term no-tillage with residue retention. Our findings reveal that antecedent SOC level exerts the primary influence, with a reduced gain potential for antecedent SOC exceeding 50 tonnes per hectare. Wet climate (Dryness Index <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>&lt;</mml:mo> </mml:mrow> </mml:math> 1.5) and low productivity (net annual primary productivity <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>&lt;</mml:mo> </mml:mrow> </mml:math> 5.5 tonnes per hectare) could further lessen the effectiveness of SOC sequestration. These key thresholds identify vast areas across Africa, Australia, South Asia, Southern Europe, and parts of North and South America as high-potential croplands for carbon sequestration and offer guidelines for assessing the reliability of regenerative agriculture in local and regional contexts.