2025 Solar-HL-Solar_wind

Solar wind is the continuous outflow of plasma from the Sun in the interplanetary space. It forms the background environment in which all other extreme space weather phenomena propagate, such as coronal mass ejections and solar energetic particles. Such phenomena can cause disastrous consequences to space- and ground-based technological systems and operations by damaging space electronics, exposing astronauts to high radiation risks, degrading power grids at Earth, among others. Additionally, the fast component of the solar wind which we usually call “High-speed-streams” can cause medium-scale geomagnetic storms. Therefore, reliably forecasting the solar wind at Earth will not only lead to better predictions of such medium-scale storms but it will also lead to reliable predictions on the arrival time and intensity of the extreme phenomena so that we can protect the human technological infrastructure.

Goal	Description

Repo structure

├── assets              # images for this readme
├── experiments         # configuration files for different trials
├── notebooks           # visualisation/testing ipynb
├── scripts             # entrypoint and highest level executors
├── src/hl-solar-wind   # python package
│   ├── ablation        # models without backbone integration
│   ├── benchmarks      # metrics for comparison
│   ├── datasets        # dataloaders/modules
│   ├── finetuning      # modules for finetuning
│   ├── io              # data input and transformation module
│   ├── models          # model components
│   ├── pretraining     # modules for pretraining
└── └── visualisation   # various graphing utilities

Datasets

Name	Description	Granularity & Source
National Solar Observatory Global Oscillation Network Group (GONG) – ADAPT Synoptic Maps Hoeksema et al. 1983, Arge et al. 2010	Data-driven photospheric magnetic field maps generated from ground-based full-disk magnetograms: GONG – Six ground stations providing nearly continuous 1-minute cadence line-of-sight magnetograms. ADAPT – Ensemble-based data assimilation using surface flux transport and randomization to produce 12 realizations per day. Maps are in Carrington coordinate system with time-dependent assimilation of observed magnetic flux.	180 x 360 synoptic maps (latitude × longitude) in sine-latitude projection, 1-hour cadence for assimilation. ADAPT – Hickmann et al. 2015 GONG – Harvey et al. 1996 Access/download: gong.nso.edu/adapt/maps/gong/ Machine learning-ready versions available on request

Models

Backbones

Name	Paper

Heads

Name	Paper

Setup

Installation

Install Miniforge from here and then run the following commands to create the hl-solar-wind-env environment:

mamba env create -f environment.yml

mamba activate hl-solar-wind-env

Next, install the package:

pip install -r requirements.txt

or if you want development dependencies as well:

pip install -r requirements.txt -r requirements-dev.txt

Optional, but highly recommended

Install pre-commit by running the following command to automatically run code formatting and linting before each commit:

pre-commit install

Usage

To run any task we assume execution inside a virutal machine and Hydra configurations, these are kept in the experiments directory. The entry point is main.py and args will select a configuration:

python scripts/main.py --config-name=default

CLI overrides are still possible with this selection but be aware of some shells not escaping quotes or square brackets:

python scripts/main.py --config-name=default experiment.seed=37

Pre-training

python scripts/main.py --config-name=pretrain_tiny

Citation

@software{HLSOLARWIND,
    title           = {{Helio Solar Wind}},
    repository-code = {https://github.com/},
    year            = {2025}
}

Acknowledgement

This work is the research product of the : This has been funded and supported by NASA under **Grant award No **. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Aeronautics and Space Administration (NASA). The research and its outputs have been designed, managed and delivered by Trillium Technologies Inc (https://trillium.tech). Trillium is a research and development company with a focus on intelligent systems and collaborative communities for planetary stewardship, space exploration and human health. Trillium aspires to ensure that the latest tools and techniques in Artificial Intelligence (AI) and Machine Learning (ML) are applied to developing open science for all Humankind.