Application of SST to forecast ionospheric delays using GPS observations

Abstract

The temporal variations of ionospheric total electron content (TEC) due to space weather events compromise the precision of the global navigation satellite system (GNSS) measurements. The vulnerability of GNSS services to ionospheric delay can be studied using TEC datasets. Hence, in this study, the ionospheric TEC measured during 1 January 2013–31 December 2013 from global positioning system receiver located at a low‐latitude station, KL University, Vaddeswaram, Guntur, India (16.31°N, 80.37°E) has been processed and analysed using synchrosqueezing transform (SST). SST is a wavelet‐based adaptive time–frequency analysis tool. The performance of SST with an autoregressive moving average (ARMA) model is analysed and validated. Its potential in decomposing and forecasting ionospheric delay for an hour in advance has been exploited. It is observed that the hybrid ARMA with SST, i.e. the SST‐ARMA model is 4% more accurate in forecasting ionospheric variations than the model ARMA and 46% more accurate than the wavelet transform‐ARMA model for the 91 days of TEC dataset (April–June 2013). The forecast results during various geomagnetic conditions of 24th solar maximum period (2013) indicate that SST‐ARMA has the potential to be a useful tool to set up early warning systems for ionospheric disturbances.