Autocorrelation method for temporal interpolation and short-term prediction of ionospheric data

An autocorrelation method is developed for temporal interpolation and short-term prediction of ionospheric characteristics. The ionospheric data are considered as a realization of a periodic process with randomly dispersed measured values superimposed on it. The autocorrelation function or its normalized autocorrelation coefficients are determined from the measured data over a period of 20–30 days, and on that basis an autocorrelation model is obtained. This model is then used to interpolate the missing values in the monthly tables of ionospheric characteristics, here called “gaps.” The interpolation at a given hour is performed by calculating weighting coefficients for the neighboring measured values. The procedure selects those measurement values around the gap which have the highest autocorrelation coefficients. The model can be used to extrapolate (predict) the data, treating the prediction period (usually 24 hours) as a gap placed at the end of the available data. The method also calculates the so-called prediction error, which is found to be close to the standard deviation of the measured data. The interpolation and prediction error are estimated to be less than 12% in the case of foF2.