The sun does not only send light to Earth, but also the solar wind which is a plasma with energetic particles and an embedded magnetic field that interacts with the magnetosphere that surrounds the Earth. As a result, energy is continuously put into the magnetosphere. Such energy input into a system can not continue forever, but the system has to relax to its ground state by releasing the energy. The almost explosive energy release by the magnetosphere is called a substorm and it leads to the acceleration of particles, generation of electric and magnetic fields, heating of the plasma, and the creation of bright dynamic aurora. SSL has been operating ground-based aurora cameras and magnetometers since 2007 that observed tens of thousands of substorms (~80 TB of data). We want to create a system that classifies aurora images and magnetometer measurements and determines characteristic features that precede substorms. The final goal will be a data science approach to the prediction of substorms similar to weather predictions that predict thunderstorms or hurricanes.

Fall 2022
Physical Science/Engineering
Technical Area(s)
Machine Learning (ML)