Prediction of Soil Organic Carbon using Machine Learning Techniques and Geospatial Data for Sustainable Agriculture.

The article focuses on predicting soil organic carbon (SOC) using digital soil mapping (DSM) and machine learning techniques to support sustainable agriculture in the Dhamtari district of Chhattisgarh, India. Four machine learning models—Random Forest (RF), Support Vector Machine (SVM), Adaptive Boosting (AdaBoost), and k-Nearest Neighbour (kNN)—were applied to multisource environmental covariates including topography, climate, and remote sensing data at 30 m spatial resolution. Among these, the Random Forest model demonstrated the best predictive performance, achieving the highest coefficient of determination (R²) and lowest root mean square error (RMSE) in both cross-validated and non-cross-validated scenarios. The resulting high-resolution SOC maps can aid farmers in precision fertilization and soil management, promoting improved crop yield and environmental sustainability. Future work aims to extend this approach to other soil nutrients and incorporate deep learning and hyperspectral remote sensing data for enhanced accuracy.