A sequential calibration approach based on the ensemble Kalman filter (C-EnKF) for forecasting total electron content (TEC)

Mona Kosary, Ehsan Forootan, Saeed Farzaneh, Maike Schumacher

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review


Ionospheric models are applied for computing the Total Electron Content (TEC) in ionosphere to reduce its effects on the Global Navigation Satellite System (GNSS)-based Standard Point Positioning (SPP) applications. However, the accuracy of
these models is limited due to the simplified model structures and their dependency on the calibration period. In this study, we present a sequential Calibration approach based on the Ensemble Kalman Filter (C-EnKF) to improve TEC estimations. The advantage of C-EnKF, over the frequently implemented state-of-the-art, is that a short period of GNSS network measurements
is needed to calibrate model parameters. To demonstrate the results, the International Reference Ionosphere (IRI)-2016 model is used as reference and the Vertical TEC (VTEC) estimates from 53 IGS (the International GNSS Service) stations in Europe are applied as observation. The C-EnKF is applied to calibrate four selected model parameters (i.e., I G12, URSI(771), URSI(1327) and URSI(1752) related to the ionospheric activity as well as height and density peak-modelling in the F2 layer), which are identified by performing a sensitivity analysis. The calibrated model, called ‘C-EnKF-IRI’, is localized within Europe and can be used for near-real time TEC estimations and forecasting of the next day (at least). Validation against the dual frequency GNSS measurements of three IGS stations indicates that during September 2017, the accuracy of forecasting VTECs is improved up to 64.87% compared to IRI-2016. The electron density (N e) profiles of C-EnKF-IRI are validated against those of COSMIC products, which indicates ∼38.1% improvement during days with low (K p = 3) and high (K p = 8) geomagnetic activity. Applying the forecasts of VTECs in SPP experiments shows similar performance as the 11-days delayed IONEX data, i.e., 51%, 52% and 79%, improvements in estimating ionospheric contributions compared to the usage of the original IRI-2016, Klobuchar and NeQuick-G models, respectively. The TEC forecasts of C-EnKF-IRI are found to be of the same quality of the IONEX final TEC products in SPP applications.
TidsskriftJournal of Geodesy
Udgave nummer4
StatusUdgivet - 20 apr. 2022


Dyk ned i forskningsemnerne om 'A sequential calibration approach based on the ensemble Kalman filter (C-EnKF) for forecasting total electron content (TEC)'. Sammen danner de et unikt fingeraftryk.