To accurately predict solar eruptions, either through conventional statistical methods or machine learning, appropriate predictors are needed. Dr. Ioannis Kontogiannis, from the Leibniz-Institut für Astrophysik Potsdam (AIP, Germany), writes here about some of those parameters.
The movie shows the evolution of active region 11158 during its first five days of appearance in February 2011, as observed from space by the Helioseismic and Magnetic Imager (HMI) instrument onboard the Solar Dynamics Observatory (SDO) of NASA. As the active region emerges in the solar photosphere it grows in size and complexity, developing regions of closely neighboring opposite magnetic polarities (shown in black and white), which also exhibit strong shearing motions. These features are called magnetic polarity inversion lines (MPILs) and indicate vast amounts of free magnetic energy, available to power flares and coronal mass ejections (CMEs). These phenomena, albeit spectacular, can be hazardous for our technology on Earth and space.
To accurately predict solar eruptions, either through conventional statistical methods or machine learning, we use appropriate predictors. These are parameters that quantify the importance and strength of MPILs and the amount of complexity of the active regions. The evolution of some predictors for AR 11158, as it grows, is also shown in the clip, where flashing letters correspond to flare occurrences (from C to X in increasing strength).
Usually, predictor values increase significantly a few hours before eruptions. Some predictors may refer to physical quantities, such as the amount of electric currents that are injected to the corona or the total magnetic flux of the active region. Others may be proxies of the free magnetic energy, such as the “Ising energy”, which depends on the length of all possible connections between opposite polarity pairs or the “effective connected magnetic field strength” which is a measure of the magnetic flux associated with these connections.
Even though we are in the era of systematic solar observations from space, prediction of solar eruptions is still probabilistic. This is partly because the magnetic information that we have is limited to the photosphere, whereas we know that the energy release takes place at higher atmospheric layers. In the future, more accurate measurements of the magnetic field, extending also to higher atmospheric layers, will open a new window and allow us to probe layers closer to the ones where energy release is taking place. The advanced instrumentation of the new European Solar Telescope can help us understand these eruptions and produce better predictors and more accurate forecasting models.
More information can be found here:
https://arxiv.org/pdf/1708.07087.pdf
https://arxiv.org/pdf/1807.06371.pdf
Data are courtesy of NASA/SDO and the AIA, EVE, and HMI science teams.