Objective  Explore the spatial distribution and influencing factors of soil organic carbon (SOC) in cultivated land in Fujian Province.

  Method  Based on the data generated from over 30,000 survey sites on farmland in Fujian, Pearson correlation coefficient and random forest model were employed to derive key factors affecting the SOC. The geodetector model was used to analyze the spatial SOC distribution in the province.

  Result  The data on SOC of the province in 2008 ranged between 0.12 and 67.28 g·kg⁻¹ with a spatial pattern of being low in the southeast coastal areas and high in the west and central regions. The geodetector model was shown to render the most comprehensive and objective analysis among the three models tested. It concluded the climate-related conditions to be the major factors affecting the spatial differentiation of SOC on the farmland with top 6 rankings of: annual precipitation (0.168 5)＞annual average temperature (0.167 7)＞altitude (0.144 9)＞climate type (0.135 9)＞soil type (0.082 4)＞landform type (0.073 1). The interactive detectors further revealed the interaction between the annual precipitation and annual average temperature to exert the greatest influence on the SOC spatial differentiation (0.194 1), while the annual precipitation and soil type (0.192 3) and the annual precipitation and cultivated land use type (0.1918) followed.

  Conclusion  Multiple factors affected the SOC on the farmland in Fujian in the past. For improving the spatial utilization efficiency and bettering the agriculture production layout on the land, it seemed imperative that all various factors highlighted in this study be taken into serious considerations.