Enhancing Global Cropland Soil Organic Carbon Stability Through Integrated Management Practices.

The ratio of mineral-associated organic carbon to particulate organic carbon (MAOC:POC) can serve as an indicator for assessing relative soil organic carbon (SOC) stability, which is closely linked to climate change mitigation and sustainable agriculture. However, the potential for enhancing relative SOC stability in global croplands, and the corresponding optimized integrated management strategies, remain unclear. Here, we developed a machine learning model that captured the interactions among cropland management practices, soil, and climate factors using a compiled global data set of 2,029 field observations, enabling systematic geospatial analysis of MAOC:POC across croplands worldwide. We found that cropland management practices, including tillage, fertilization, and irrigation, exert strong control over the MAOC:POC, and identified grid-specific optimal integrated combinations that achieve the highest ratio without compromising total SOC. Compared with the current global mean MAOC:POC of 6.32, implementing these strategies could increase this ratio to 9.58, representing a 51.5% improvement. Our findings provide the first global spatial diagnosis of relative SOC stability and demonstrate the potential of context-adapted strategies, offering actionable insights for carbon-smart and climate-resilient agriculture.