Physics-Informed Neural Networks improve N2O flux predictions over classical models.
problem Predicting N2O flux emissions from agricultural processes.
method Constructed a rigorously derived physics residual from DayCent models and trained an MLP-based PINN on agricultural data.
result Physics-Informed Neural Networks consistently outperform classical models in predicting N2O flux emissions.