The sun may seem a stable luminary in the sky, but in fact it is a turbulent sphere of electrically charged gas. Its activity is formed by complex magnetic forces that create unpredictable solar phenomena that challenge scientists, writes Science Alert.

The focus. Technology has its own Telegram channel. Subscribe so as not to miss the latest and most exciting news from the world of science!

Understanding and forecasting these disturbances is crucial, especially when we are talking about coronal mass emissions (KVM), which can affect earthly technologies and communication systems. A recent study is studying how artificial intelligence can improve forecasting solar events.

Researchers trained machine learning models using decades of sunny data, focusing on the active area AR13664. Their conclusions indicate that AI discovered early warning signs of increased solar activity, which can potentially improve methods for forecasting future solar outbreaks.

KVM is a massive bucks of plasma thrown out of the crown of the sun due to sudden shifts of magnetic power lines. These emissions can move at a speed of hundreds to thousands of kilometers per second, sometimes reaching land in just a few days. Interacting with the magnetosphere of our planet, they can cause geomagnetic storms that disable satellites, GPS networks and power networks, and also create impressive polar radiance.

A group of scientists under the leadership of Sabrina Guastavino from the University of Genoa used artificial intelligence methods to analyze a solar storm, which occurred in May 2024. Their machine learning models processed large historical data to identify complex patterns that traditional methods may not notice.

The study was aimed at predicting solar outbreaks, tracking the formation of a KVM and predicting geomagnetic storms. Their model demonstrated a high level of accuracy, significantly improved forecasts compared to traditional approaches.

The results of the study are of wide importance: accurate forecasts can help soften interruptions in the operation of energy infrastructure, communication networks and satellite communications. In addition, the best prediction of auroral activity can improve the experience of the sky observers who seek to see these natural light phenomena.

Previously, Focus wrote about the unique treasure of Roman and British coins discovered in the Netherlands. This find is the first of its kind, since so far similar treasures have been found only in Britain.

We also talked about the study of the 1900-year-old scroll found in the Jewish desert. Scientists have found that this is a record of a lawsuit, which is the longest legal text of this kind.