A neural network approach to predict the ionospheric scintillation Wbmod model variables

Abstract

The ionospheric scintillation can be explained as the fluctuations in the phase and intensity of electromagnetic rays after crossing the ionosphere. Rino’s theory was proposed in 1979 to quantify this scintillation, and subsequent models appeared to predict its characteristics. One of them is the WideBand ionospheric scintillation Model (WBMOD) from 1984. This study aims to provide a neural network that emulates the behavior of WBMOD model, by learning from phase and intensity scintillation data gathered from several ESA projects. By using the Rino’s power-law phase-screen ionospheric scintillation theory, the values of the height-integrated irregularities strength (C k L) and the slope of its PSD (q) can be obtained from the physically measured S 4 and σ ϕ . So, using this data, two neural networks are presented which obtains results that fit well with the WBMOD expected values.