Sparse-Observation Atmospheric Thermal Forecasting with Physics-Informed Neural Networks for Climate-Aware Digital Twins

This research evaluates physics-informed neural networks for atmospheric potential-temperature forecasting under incomplete thermal observations. The PINN is constrained by a pressure-coordinate thermodynamic advection-source equation, with a diabatic-source closure fitted from the preceding 12-hour period and frozen before future-time training. The method leverages hourly ERA5 reanalysis data for training and evaluation.