Using ARMA and a linear time series model, an algorithm for multivariate ionospheric TEC forecasting over low-latitude GNSS stations has been developed

Abstract

The model combines a linear time series model with Autoregressive and Moving Average (ARMA) methods. It utilizes data from the Bengaluru International GNSS Service (IGS) station during the 24th solar cycle (2009-2016). Various factors, including geomagnetic activity (Ap), solar Extreme Ultraviolet (EUV) irradiance (F10.7), periodic oscillations (annual, semi-annual, terannual, and biennial), and long-term trends, are considered as input parameters along with real-time TEC observations. The model investigates the impact of these factors on TEC and uses ARMA for forecasting each factor. The forecasted factors are then combined to obtain the forecasted TEC values, which show good agreement with observed GPS-TEC. The study reveals that the semi-annual variation has higher magnitudes during the High Solar Activity (HSA) period, while the geomagnetic effect on TEC is relatively low. The proposed model is considered valuable for characterizing low-latitude ionospheric variations under different space weather conditions.