Implementation of VARMA Model for Ionospheric TEC Forecast over an Indian GNSS Station

Abstract

Accuracy in positioning services of the Global Navigation Satellite System (GNSS) is majorly affected due to ionospheric signal delays. The forecasting of ionospheric delays is tough and challenging low-latitude regions due to rapid temporal variations in ionospheric electron density irregularities. Hence, in this paper a non-stationary signal decomposition technique based on Variational Mode Decomposition (VMD), combined with Auto Regressive Moving Average (ARMA) called VMD-ARMA (VARMA) model is presented to forecast the ionospheric delay values 1 hour ahead. The performance of the proposed VARMA ionospheric TEC forecasting algorithm is tested during geomagnetic storms that occurred in June 2013. Three months GNSS data i.e., from 1 April 2013- 30 June 2013 is logged using GNSS Ionospheric Scintillation and TEC Monitor (GISTM) receiver located at Koneru Lakshamaiah Education Fondation, (K L E F), Guntur station (geographic: 16.37°N, 80.44°E), India. It is found that the VARMA model is 2-3% more efficient than the ARMA model in providing good forecasting accuracy during storm conditions. The forecasting results demonstrate that the VARMA version can be useful to forecast the ionospheric TEC variations at low-latitude regions during disturbed ionospheric space weather conditions also.