Probabilistic prediction of phosphate ion adsorption onto biochar materials using a large dataset and online deployment
Published in Chemosphere, 2025
Recommended citation: Iftikhar, S., Ishtiaq, R., Zahra, N., Ruba, F., Lam, S.-M., Abbas, A., & Jaffari, Z. H. (2025). Probabilistic prediction of phosphate ion adsorption onto biochar materials using a large dataset and online deployment. Chemosphere, 370, 144031. https://doi.org/10.1016/j.chemosphere.2024.144031
Phosphate washing into rivers and lakes is a leading trigger of algal blooms, and biochar — a charcoal-like material made from waste plant matter — can capture it cheaply, though how well it works varies enormously with how it is made and used. Pooling 2,952 measurements from the published literature, this work trained models that not only predict how much phosphate a given biochar will remove but also report how confident they are in each figure. The best explained 95% of the variation for materials it had never encountered, and showed that how the treatment is run carries the most weight, at 43% of the influence, ahead of how the biochar was manufactured at 29% and what it is made of at 28%. The team then released the model as a free web application, so a candidate material can be screened without running a single experiment.