[Test] SEPFLOW: A High-Resolution Multivariate Time-Series Dataset for Solar Energetic Particle Forecasting

Solar energetic particle (SEP) events pose significant risks to space-based infrastructure and human space exploration, highlighting the need for reliable forecasting methods. We introduce SEPFLOW, a high-resolution multivariate time-series dataset spanning 28 years (1998–2025) to support data-driven SEP forecasting. SEPFLOW integrates solar wind and interplanetary magnetic field measurements from OMNI with GOES soft X-ray and energetic proton observations. The dataset provides synchronized measurements at 5-minute resolution over 24-hour windows preceding solar flare onset, along with SEP and non-SEP event labels. By providing a comprehensive, consistently aligned observational resource, SEPFLOW aims to facilitate reproducible research, comparative model evaluation, and the development of advanced machine learning approaches for space weather forecasting.