Forecasting the Relativistic Electron Flux at Geosynchronous Orbit.

摘要

A neural network, developed to model the temporal variations of relativistic (> 3 MeV) electrons at geosynchronous orbit, has been used to make reasonably accurate day-ahead forecasts of the relativistic electron flux at geosynchronous orbit. This model can be used to forecast days when internal discharges might occur on geosynchronous satellites or satellites operating within the outer Van Allen radiation belt. The neural network (in essence, a nonlinear prediction filter) consists of three layers of neurons, containing 10 neurons in the input layer, 6 neurons in a hidden layer, and 1 output neuron. The network inputs consist of ten consecutive days of the daily sum of the planetary magnetic index, Kp. The output is a prediction of the daily averaged electron flux for the tenth day. The neural network model, together with projections of Kp based on its historical behavior, can be used to make the day-ahead forecasts of the relativistic electron flux at geosynchronous orbit. A significantly better forecast is obtained by modifying the network to include one additional input, the measured daily averaged electron flux for the day prior to the forecast day, and one more neuron in the hidden layer. Both models are described in this report.