Temperature changes after sunset can lead to hot pockets of gas known as ‘plasma balloons’ that form over the Earth’s surface. This causes communication disruptions between satellites. Scientists can now forecast plasma bubble events using new AI models. Sachin Reddy (a University College London PhD student) presented the work at this week’s National Astronomy Meeting (2022).
Shortly after sunset, plasma bubbles (super-heated gas) form in the upper atmospheric and extend into space (upto 9000km above the Earth’s surface). These bubbles grow quickly from a tiny size to a large country in a matter of hours. By blocking and warping satellite radio signals, the bubbles can grow larger and prevent satellites communicating with Earth.
A team of researchers has compiled 8 years worth of data from SWARM’s satellite mission to predict plasma bubbles. The Ionospheric Bubble Index is an automated bubble detector aboard the spacecraft. To determine whether bubbles exist, this compares the magnetic field strength with the density of electrons. If there is a strong correlation between them, it indicates that a plasma-bomb is present.
The satellite is able to fly at an altitude of 460 km (about 30x greater than a commercial aircraft) and through most plasma bubbles. The model incorporates data from SWARM with machine learning techniques to predict when a plasma bubble event will occur.
Results show that there are a variety of plasma bubble events from one season to the next, much like weather. The number of events also increases with increasing solar activity. The model still finds that location is more important than the time of the year in predicting plasma balloons. Most events occur in an area in the Atlantic known as the South Atlantic Anomaly. The AI model predicts events accurately with 91% accuracy across various tests.
Reddy states that “just like the weather forecast on Earth, we need be able predict bubbles to stop major disruptions of satellite services.” Our goal is to be able say that at 8pm tomorrow, there is a 30% chance of bubbles appearing over Horn of Africa. This information is very useful for satellite operators as well as for those who depend on satellite data every single day, like you and I.