Reconstruction or modeling of the ionospheric electron density has steadily been developed through various techniques and computerized ionospheric tomography (CIT) is one of them. There are over 80 Global Navigations Satellite System (GNSS) stations over South Korea and it can be accessed through Korea Astronomy and Space Science Institute (KASI). Especially, KASI data base maintains Receiver independent exchange format (RINEX) files for every 10 minutes. Although slant total electron contents (STEC) from about 80 stations and 10 minutes interval dataset are not enough to finely conduct CIT, using empirical orthonormal functions (EOFs) can reduce not only the number of voxels for vertical directions but also the time required to calculate. It allows to reconstruct the ionospheric condition for near real time (NRT) by means of computerized tomography. We derive EOFs according to solar activity, seasons, local time and Korean region from International Reference Ionosphere-2016 model (IRI-2016) which is one of the most popular empirical model for the ionosphere. After simplifying 3-dimensional geometry matrix of the ionospheric electron density using EOFs, multiplicative algebraic reconstruction technique (MART) algorithm with an initial condition of IRI-2016 is used to perform the CIT. The CIT results are already validated by using ionosonde NmF2 (ionospheric peak density at F2 layer) at two South Korean stations in statistical. It shows improvement of the ionospheric electron density calculation. Moreover, whole processes for the CIT is less than 10 minutes and it means that our results show the possibility for near real time reconstruction of the ionosphere. Furthermore, the results for special case during St. Patrick’s day geomagnetic storm in 2015, which is one of the strong geomagnetic storm in solar cycle 24, are selected to compare with IRI-2016 and NeQuick2 models in this study. STEC calculated by Tomography, IRI-2016 and NeQuick2 models are compared with observed STEC at ground stations for consecutive 3 days (before, during and after day of geomagnetic storm) during 10 minutes interval. The CIT results approach to actual observation more than both IRI-2016 and NeQuick2 models. IRI-2016 and NeQuick2 models show low correlation coefficients and linear fitting slopes during night time. After geomagnetic storm, the gap between models and observation increase and it is very remarkable during recovery phase of geomagnetic storm. Even though the CIT is greatly dependent to the initial condition, the CIT results during geomagnetic storm show better. In this study, we will present our tomography techniques and results.

Keywords: CIT, EOF, MART

Speaker Junseok Hong*