Anthropogenic controls over soil organic carbon distribution from the cultivated lands in Northeast China

[Display omitted] •Quantifying the potential human factors on SOC content in topsoil of agroecosystems.•Mapping the spatial distribution of SOC content in agroecosystems of Northeast China.•Evaluating the predictive performance and application potential of the predictive model. Both natural and anthropogenic variables affect soil C distribution and its pool, however studies about anthropogenic influence on soil C distribution are very limited in the literature. This study investigated anthropogenic effects on soil organic carbon (SOC) changes in the cultivated lands of Northeast China. A total of 196 topsoil samples (0–30 cm) were collected, and analyzed for SOC content, and 12 environmental variables (natural and anthropogenic) were selected as SOC predictors. Natural factors included elevation, slope gradient, slope aspect (SA), topographic wetness index (TWI), mean annual temperature, mean annual precipitation, and normalized difference vegetation index, while population (POP), gross domestic product (GDP), distance to the socioeconomic center, distance to roads, and reclamation period (PER) represented anthropogenic variables. Three different boosted-regression trees models with different combination of SOC predictors were constructed, and the model performance was evaluated with 10-fold cross-validation. We found that the model that included all predictors had the best performance, followed by the model with topography and climate variables, and the model with only anthropogenic variables. However, adding the anthropogenic variables in the model greatly improved its performance. Results showed that PER, POP and GDP were the key environmental variables affecting SOC content in the topsoil agroecosystems in Northeast China. This study suggests that anthropogenic variables should be selected as the main environmental variable in predicting of SOC content in agroecosystem with a higher human influence. We believe that the accurate prediction and mapping of SOC content in the topsoil agroecosystem will help formulate farmland soil management policies and promote soil carbon sequestration.