WSA+ (Neural Enhancement of the Traditional Wang–Sheeley–Arge Solar Wind Relation)

The training dataset and the WSA+ model checkpoint (wsaplus.pt) are hosted here.

Along with scripts for training, evaluating, and visualizing the WSA+ model — a neural enhancement of the traditional Wang–Sheeley–Arge (WSA) solar wind relation. 

The workflow is divided into two main stages:

1. Optimization — per Carrington Rotation (CR) fitting of WSA empirical parameters using in-situ solar wind observations.
2. Generalization — training a neural network to predict optimized WSA speed maps directly from magnetogram-derived features.

WSA+ model can be freely installed using: pip install wsaplus

For  more information, refer to this preprint: https://doi.org/10.48550/arXiv.2509.06181