The module addresses oceanic mesoscale eddy analysis by providing the tools for pre-processing FESOM2 data and training DL models.
Release Notes
The development of the Python packages is currently underway. In this version, a pre-processing pipeline has been developed, with both a complete notebook and a demonstration one. They allow interpolations from unstructured to regular grids and vice versa to prepare the dataset for the training, which has both a demonstrative notebook with plots and a Python script to be run with Slurm on the chosen cluster. Both use the PyTorch Geometric framework to run the deep learning process, and both are natively built to make use of GPUs, if the underlying system offers them. Custom loss functions and a Graph U-Net model were customised for this task.
Future Plans
The next versions will improve the prediction capabilities of the trained network. It will be possible to have a tool that can produce eddy segmentation masks directly on unstructured climate grids, and do it faster than the classic physics-based algorithms that only work on regular matrix data. Additional usage examples will be included in the documentation.