NASA Heliolab Releases Machine Learning Dataset for Ionospheric Forecasting

This paper introduces a new machine learning-ready dataset for ionospheric forecasting, developed as part of the 2025 NASA Heliolab. The dataset integrates diverse ionospheric and heliospheric measurements, including data from the Solar Dynamic Observatory, solar wind parameters, geomagnetic activity indices, and Global Ionospheric Maps of Total Electron Content (GIM-TEC). It also incorporates geospatially sparse data from GNSS receivers and crowdsourced Android smartphone measurements. The dataset is temporally and spatially aligned into a single, modular structure, supporting both physical and data-driven modeling approaches. The authors demonstrate the utility of the dataset by training and benchmarking several spatiotemporal machine learning architectures for forecasting vertical TEC under various conditions. This work provides a valuable resource for researchers and operational forecasters, enabling exploration of ionospheric dynamics and broader Sun-Earth interactions. The availability of this open-access dataset is expected to accelerate progress in ionospheric forecasting and improve the accuracy of space weather predictions.

_Context: This intelligence report was compiled by the DailyOrbitalWire Strategy Engine. Verified for Art. 50 Compliance._