Improving deep learning precipitation nowcasting by using prior knowledge

TL;DR

This paper enhances deep learning precipitation nowcasting by integrating physical prior knowledge through a PhyCell, aiming to improve interpretability and capture high-frequency features often missed by traditional models.

Contribution

The study introduces a physical prior via a PhyCell into a PhyDNet model, combining physical equations with deep learning for better precipitation prediction.

Findings

PhyCell learns the intended physical dynamics.

Training PhyDNet remains driven by loss optimization.

Prediction capabilities are comparable to existing models.

Abstract

Deep learning methods dominate short-term high-resolution precipitation nowcasting in terms of prediction error. However, their operational usability is limited by difficulties explaining dynamics behind the predictions, which are smoothed out and missing the high-frequency features due to optimizing for mean error loss functions. We experiment with hand-engineering of the advection-diffusion differential equation into a PhyCell to introduce more accurate physical prior to a PhyDNet model that disentangles physical and residual dynamics. Results indicate that while PhyCell can learn the intended dynamics, training of PhyDNet remains driven by loss optimization, resulting in a model with the same prediction capabilities.