Prediction Model of the Geosynchronous Electron Fluxes at a Wide Energy Range Based on a Neural Network Scheme

The orbit in the range 2 to 7 Re (earth radii), which include the geosynchronous orbit, is known to be filled with particles of various energies. High flux levels of energetic electrons can cause irreparable damage to the instruments and equipment on satellites. Significant problems in satellite systems due to flux enhancement have promoted development of electron flux prediction model. In this study, we adopted a neural network technique to prediction the electron flux in a geosynchronous orbit. Solar wind data and geomagnetic indices are used for input parameter of neural network. As a result, we present combinations of solar wind and geomagnetic indices that show highest prediction efficiency. Our prediction model can predict the typical substorm-associated energetic (~40-400 keV) and relativistic-energy (> 0.8 MeV, > 2 MeV) electron fluxes up to 24 hours ahead with a reasonably good prediction efficiency.