Physicists urgently teach neural networks to predict fatal cosmic weather.
A solar storm can begin far from Earth, but the consequences quickly reach quite terrestrial systems: power grids, railways, satellite communications and navigation. The streams of the solar wind and geomagnetic disturbances rock the magnetosphere and the ionosphere, which is why noticeable changes appear in the electric and magnetic field. For scientists, the problem is complicated by the fact that some of these signals are similar to the weaker traces of natural disasters, such as large eruptions or other processes that also reach the upper layers of the atmosphere.
And this has long been an abstract danger. On February 3, 2022, a moderate space weather event led to a major loss of SpaceX satellites. Shortly after launch, the company lost 38 of 49 Starlink devices. The event showed an unpleasant thing: even the most powerful geomagnetic storm can significantly hit the road infrastructure if the forecast was not accurate enough.
The General Assembly of the European Union of Earth Sciences EGU26 was presented with the Swarm-AWARE project, which was launched by the European Space Agency. The full name of the project is Swarm Investigation of Space Weather and Natural Hazards Effects. The team wants to learn to more accurately separate the electromagnetic traces of space weather from signals related to natural hazards on Earth.
The project is based on Swarm satellites. These devices measure the Earth’s magnetic field, plasma density and temperature, electric fields, and other parameters of the near-Earth environment. Satellite data alone is not enough, so the researchers combine Swarm measurements with ground observations and Copernicus Sentinel-5P data. Georgios Balasis of the National Observatory of Athens explained that such a set of sources helps to understand where the ionosphere reacts to solar activity, and where the signal is associated with an earthly event.
The ionosphere is high in the atmosphere and contains charged particles. Radio communication, navigation signals and part of satellite systems depend on the state of this layer. When the solar wind changes the magnetic field of the planet, disturbances occur in the ionosphere. Similar changes can come from below: a powerful eruption, a shock wave or atmospheric oscillations can rise to the upper layers of the atmosphere and leave an electromagnetic trail. To mix these sources, the infrastructure forecasts lose accuracy.
A good test for Swarm-AWARE was the eruption of the Hunga-Tong volcano in 2022. The explosion ejected a huge amount of water from the South Pacific to the stratosphere and generated waves that reached the upper atmosphere. In the ionosphere after that there were strong changes in density. The waves also triggered electric fields that went along the lines of the Earth’s magnetic field and almost immediately manifested themselves on the opposite side of the Pacific Ocean. Smar magnetometers recorded these disturbances.
For researchers, the example of Hong Kong is important as a reference case. The volcano was on Earth, but the signal rose so high that the satellites saw the consequences in the near-Earth environment. A similar picture can confuse monitoring systems if solar activity simultaneously operates. Swarm-AWARE should help to disassemble a complex mixture of signals and show which source caused a specific change.
The project team will apply machine learning and advanced time series analysis. Simply put, algorithms will look for patterns in long chains of satellite and terrestrial measurements: where the signal grows, how quickly spreads, whether it coincides with a solar storm, a volcanic event, or another natural process. Such an approach should improve understanding of how space weather affects infrastructure, communications and early warning systems.
The ultimate goal of Swarm-AWARE is not limited to scientific publications. If researchers learn to distinguish more confidently in sources of electromagnetic perturbations, satellite operators, energy, transport services and monitoring centers will be able to rate risks faster in almost real time. For infrastructure, the difference is fundamental: one thing is temporary noise in data, another is the beginning of an event that can disable satellites, disrupt communication or create problems for power grids.
A solar storm can begin far from Earth, but the consequences quickly reach quite terrestrial systems: power grids, railways, satellite communications and navigation. The streams of the solar wind and geomagnetic disturbances rock the magnetosphere and the ionosphere, which is why noticeable changes appear in the electric and magnetic field. For scientists, the problem is complicated by the fact that some of these signals are similar to the weaker traces of natural disasters, such as large eruptions or other processes that also reach the upper layers of the atmosphere.
And this has long been an abstract danger. On February 3, 2022, a moderate space weather event led to a major loss of SpaceX satellites. Shortly after launch, the company lost 38 of 49 Starlink devices. The event showed an unpleasant thing: even the most powerful geomagnetic storm can significantly hit the road infrastructure if the forecast was not accurate enough.
The General Assembly of the European Union of Earth Sciences EGU26 was presented with the Swarm-AWARE project, which was launched by the European Space Agency. The full name of the project is Swarm Investigation of Space Weather and Natural Hazards Effects. The team wants to learn to more accurately separate the electromagnetic traces of space weather from signals related to natural hazards on Earth.
The project is based on Swarm satellites. These devices measure the Earth’s magnetic field, plasma density and temperature, electric fields, and other parameters of the near-Earth environment. Satellite data alone is not enough, so the researchers combine Swarm measurements with ground observations and Copernicus Sentinel-5P data. Georgios Balasis of the National Observatory of Athens explained that such a set of sources helps to understand where the ionosphere reacts to solar activity, and where the signal is associated with an earthly event.
The ionosphere is high in the atmosphere and contains charged particles. Radio communication, navigation signals and part of satellite systems depend on the state of this layer. When the solar wind changes the magnetic field of the planet, disturbances occur in the ionosphere. Similar changes can come from below: a powerful eruption, a shock wave or atmospheric oscillations can rise to the upper layers of the atmosphere and leave an electromagnetic trail. To mix these sources, the infrastructure forecasts lose accuracy.
A good test for Swarm-AWARE was the eruption of the Hunga-Tong volcano in 2022. The explosion ejected a huge amount of water from the South Pacific to the stratosphere and generated waves that reached the upper atmosphere. In the ionosphere after that there were strong changes in density. The waves also triggered electric fields that went along the lines of the Earth’s magnetic field and almost immediately manifested themselves on the opposite side of the Pacific Ocean. Smar magnetometers recorded these disturbances.
For researchers, the example of Hong Kong is important as a reference case. The volcano was on Earth, but the signal rose so high that the satellites saw the consequences in the near-Earth environment. A similar picture can confuse monitoring systems if solar activity simultaneously operates. Swarm-AWARE should help to disassemble a complex mixture of signals and show which source caused a specific change.
The project team will apply machine learning and advanced time series analysis. Simply put, algorithms will look for patterns in long chains of satellite and terrestrial measurements: where the signal grows, how quickly spreads, whether it coincides with a solar storm, a volcanic event, or another natural process. Such an approach should improve understanding of how space weather affects infrastructure, communications and early warning systems.
The ultimate goal of Swarm-AWARE is not limited to scientific publications. If researchers learn to distinguish more confidently in sources of electromagnetic perturbations, satellite operators, energy, transport services and monitoring centers will be able to rate risks faster in almost real time. For infrastructure, the difference is fundamental: one thing is temporary noise in data, another is the beginning of an event that can disable satellites, disrupt communication or create problems for power grids.
