Software
Solar Flare Records Dashboard
A web application that brings together solar flare records from NOAA, LMSAL, and NCEI Science Quality Data in one place. Data from the three sources are updated daily through asynchronous collection.
Software & Data
Software, datasets, and tools for research and discovery.
Dataset · 2026
SuryaBench
A collaborative collection of machine-learning-ready Solar Dynamics Observatory observations and benchmarks for heliophysics. Includes datasets for solar flare forecasting, active-region segmentation and emergence, coronal field extrapolation, solar wind forecasting, and EUV spectra modeling, with preparation code and baseline implementations.
Related publications
Software · 2025
Surya: Heliophysics Foundation Model
A collaborative foundation model trained on Solar Dynamics Observatory imagery. The implementation includes inference and adaptation workflows for applications such as solar flare forecasting, active-region segmentation, solar wind prediction, and EUV spectra modeling.
License: Apache-2.0
Related publications
Software · 2025
TAMAG
A Python library for transforming and augmenting solar magnetograms. TAMAG provides preprocessing and augmentation tools for machine learning workflows in solar physics and space weather forecasting.
License: GPL-3.0-or-later
Related publications
Dataset · 2024
Bitmap Filtered Line of Sight HMI Active Region Patches with Augmentations
Hourly HMI/SDO line-of-sight magnetograms of solar active regions from 2010–2018, filtered using SHARP bitmaps and standardized to 512 × 512 pixels. Includes flare labels and augmented images for training and evaluating solar flare prediction models.
Dataset · 2023
Data-Driven Labels for Solar Flare Prediction
Solar flare labels derived from GOES X-ray measurements, including relative flux increases and cumulative intensity measures. Includes companion code for exploring alternatives to conventional peak-flux labels in solar flare forecasting.
Related publications
- Enhancing solar flare prediction with innovative data‑driven labels (2023) · Read paper
- Beyond traditional flare forecasting: A data‑driven labeling approach for high‑fidelity predictions (2023) · Read paper
- An innovative solar flare metadata collection for space weather analytics (2023) · Read paper
Software · 2024
GEO-CLOAK
An end-to-end geomagnetic forecasting pipeline developed through FDL Heliolab. It combines SHEATH solar-wind forecasts with DAGGER-CL ground-level magnetic-field predictions, using continual learning to adapt as new observations arrive.
License: MIT