Machine learning-assisted model for predicting biochar efficiency in colloidal phosphorus immobilisation in agricultural soils

Abstract The loss of colloidal phosphorus (P coll ) from agricultural lands significantly contributes to nonpoint source nutrient pollution of receiving waters. This study aimed to develop an advanced machine learning (ML) model to predict the immobilisation efficiency of P coll (IE-P coll ) by biochar in agricultural soils. Six ML algorithms were evaluated using a dataset containing 18 biochar- and soil-related variables. The random forest (RF) algorithm outperformed the others (R 2 = 0.936–0.964, RMSE = 2.536–3.367), achieving superior test performance (R 2 = 0.971, RMSE = 2.276). Key biochar-related parameters, such as oxygen content, total phosphorus content, and application rate were found to be stronger drivers of IE-P coll than most soil parameters. Soil Olsen-P was found to be a more reliable predictor of IE-P coll than the other soil-related parameters. Feature selection techniques narrowed down the original 18 features to the most critical ones, enhancing the performance of the model. A graphical user interface based on the optimised model was developed to provide practical field-based predictions of IE-P coll under varying conditions. This study highlights the strong potential of using biochar as a sustainable soil amendment to enhance P coll immobilisation, thereby reducing non-point source nutrient pollution from agricultural soils. Graphical Abstract