Method of modeling and forecasting of ionospheric data based on the combination of wavelet transform and autoregressive-integrated moving average models

• download

   The work is focused on the development of technologies and software
   systems for the study of the time variations of ionospheric parameters and
   detection anomalies that can occur during periods of ionospheric
   disturbances. The paper describes a method of modeling of calm
   (background) variations of the critical frequency of the F2 layer of the
   ionosphere. The method is a combination of multiply-scaled analysis and
   methods of autoregressive-integrated moving average (ARIMA). This method
   allows identifying typical behavior of ionospheric parameters, making
   forecast with five-hour increments and detecting anomalies which occur
   during periods of ionospheric disturbances. Using multiply-scaled analysis
   allows us to decompose the time series into components and simplify the
   complex structure of the data. Computational algorithms developed on the
   basis of the method provide a selection of informative components and
   noise reduction; these informative components are modeled by ARIMA
   methods. Forecasting and analysis of residual errors of the model provides
   detection of anomalies which can occur during periods of ionospheric
   disturbances. The proposed method can be used to fill data gaps with
   respect to diurnal and seasonal variation.

   To test the method, we used data of the critical frequency of the F2 layer
   of the ionosphere (registered at the station Paratunka, IKIR FEB RAS), as
   well as data of the total electron content in the ionosphere over
   Kamchatka. We detected some peculiarities associated with solar and
   seismic activity on Kamchatka using the method we developed.