MODELING, PREDICTION AND FORECAST OF THE IONOSPHERIC CRITICAL FREQUENCY foF2

The temporal variations of the ionospheric critical frequency foF2 is a typical example of a time series in which deterministic variations at various time scales and nonstationary stochastic variations are involved. In a series of journal and conference papers we have obtained various modelling, prediction and forecast algorithms for foF2 variations over Europe, based on data from about 15 ionosonde stations over 41 years. We recovered the well known results that variations in the monthly medians obey a linear or parabolic regression model in terms of the smoothed sunspot number R12. A model involving a trigonometric expansion in the harmonics of the annual variation modulated by R12 is constructed. For prediction of the monthly medians, we used this model with a “sliding window”, based on 48 months of observation to predict the foF2 for the forthcoming month, with an error of about %3-4. Statistical properties of the deviations from monthly medians were studied and we used a one-parameter feedback to forecast the hourly values of foF2 was used.