New AI model detects hidden signs of solar eruptions hours before they emerge

A new AI model called EarlyDetect has been developed to detect early signs of solar eruptions, providing warnings up to nine hours before they occur. The model, which uses machine learning, analyzes acoustic power maps and magnetic field measurements from NASA's Solar Dynamics Observatory (SDO). The study, published in the Journal of Geophysical Research: Machine Learning and Computation, reports that EarlyDetect can identify precursor signals in the sun's acoustic activity and magnetic field. The model is not yet operational but has shown promising results in forecasting active-region emergence.

The model was developed by researchers from New Jersey Institute of Technology (NJIT), Princeton University, and NASA's Ames Research Center. The team used observations from SDO's Helioseismic and Magnetic Imager (HMI) to analyze hourly acoustic power maps and magnetic field measurements. The model uses a Transformer architecture, similar to large language models like ChatGPT. After training, the model identified precursor signals an average of 9.24 hours before active regions became visible, outperforming both a standard Transformer model and a previous benchmark approach.

The study highlights that machine learning can help predict when solar active regions will emerge, potentially allowing satellite communications and power grid companies to prepare for solar storms. However, the model is not yet ready for real-time forecasting, and it still needs validation across more solar events. The team has also released the Solar Active Region Emergence Dataset (SolARED), a publicly available collection of solar active region observations, to help other researchers develop new prediction approaches. The dataset provides a shared resource for both the machine learning and heliophysics communities.

AI-generated summary of the original article — see the source for the full story.