Previous Abstract Return to Session B1a Next Abstract

Session B1a: Atmospheric Effects

GNSS TEC Data Assimilation by Using NTCM and Local Ensemble Kalman Filtering for the Next Generation Ionospheric Prediction Services
Matteo Sgammini and Francesco Menzione, European Commission, Joint Research Centre (JRC)

The effect of the Earth’s ionosphere represents a relevant contribution to the Global Navigation Satellite System (GNSS) error budget. Abnormal ionospheric conditions can impose degradation on GNSS system functionality, including integrity, accuracy and availability. Accurate estimation and forecasting of Ionospheric Total Electron Content (TEC) on a global scale is an enabler for novel fast PPP solutions ([1], [2]), and a key element for those single frequency receivers aiming to achieve a positioning accuracy similar to that of dual?frequency [3]. Moreover, spatial/temporal gradient of TEC can be used to monitor irregularities and travelling ionosphere disturbances affecting receivers tracking [4]. In this context, the EU Ionosphere Prediction Service (IPS) project [5] was set up to design and develop a prototype platform to deploy and test TEC nowcast and short/long term forecast [6] solutions to support next generation Galileo services. The system fully inherited achievements obtained in the frame of the Nequick/Neqick-G model development [3], whose TEC modelling was used to leverage the IGS network GNSS data.
This work provides insight into the research activity performed in order to upgrade the IPS nowcast to a next generation TEC estimation system, integrating the NTCM/NTCM-G (Neustrelitz TEC Model) ionospheric model [7] and an advanced data assimilation solution based on Localized Ensemble Kalman Filtering (LEKF) [8]. NTCM is a fundamental enabler of the developed approach considering its huge advantage in terms of computational burden. Even different from a 3D model as NeQuick G, the NTCM version adapted for Galileo, named NTCM G (https://www.gsc-europa.eu/), has demonstrated comparable performance, while providing TEC propagation and ionospheric delay hundreds time faster than NeQuick.
In the framework of a high performance LEKF, given the large-scale system and costly propagation of the ensemble members, such a solution becomes more and more competitive considering the user increasing demand of near real-time information. Therefore, NTCM has been selected as the natural candidate to provide background and forecast TEC distributions requested by LEKF sequential data processing of IPS GNSS observables. In addition to the baseline single layer approach, the extension to the dual (multiple) layer solution proposed in [9] has been considered in order to better represent the lower and upper components of the ionosphere. This enhancement is achieved by introducing a set of scaling factor allowing to perform the NTCM update within the sequential process. The work provides also a complete description of such an implementation and presents all the results relying on the preliminary testing activity performed in view of its deployment on the target IPS platform.
The work is organized as follow. Section 1 describes the relevant design drivers allowing to define the novel nowcasting architecture based on NTCM and LEKF integration. General aspects relying on IPS GNSS network measurement acquisition, pre-processing and calibration using high accuracy corrections (ephemeris, clock, DCB) as well as TEC estimation product requirements (grid resolution, global to regional downscaling) are preliminary introduced. Advantages of using such an ensemble solution and relevant processing steps are described in section 2, which also includes hyperparameter definition necessary to optimize the estimation performances (covariance inflation, localization regions, etc.). Section 3 is devoted to the mathematical definition of the NTCM model background and forecast implementation together with observation model equations necessary to process the retrieved GNSS Slant TEC. Different criteria for ensemble generation (i.e. NTCM input parameter perturbation, measurement error budget) are then discussed for tuning purposes. The possibility to optionally move the architecture from single to double layer ionospheric model is here addressed clarifying the set of mapping function, interpolation techniques and scaling, which can be used to improve ionosphere representation. Section 4 provides all relevant results about the algorithm performances which are assessed in different ionospheric activity conditions by using as reference both external provider products (i.e. CODE) as well as cross-validation methods against a subset of reference station. The compatibility with IPS next generation needs are discussed in Section-5, deriving the next steps and further development to be carried out for the deployment in the prototype platform.
[1] Rovira-Garcia A, Juan JM, Sanz J, Gonzalez-Casado G. (2015). A Worldwide ionospheric model for fast precise point positioning. Geoscience and Remote Sensing, IEEE Transactions on, 53, 4596-4604.
[2] Rovira-Garcia A, Juan JM, Sanz J, González-Casado G, Bertran E (2016b) Fast precise point positioning: a system to provide corrections for single and multi-frequency navigation. NAVIGATION J Inst Navig 63(3):231–247. https://doi.org/10.1002/navi.148
[3] European GNSS (Galileo) Open Service—Ionospheric correction algorithm for Galileo single frequency users, Issue 1.2, September 2016, European Commission (EC)
[4] Rodríguez-Bilbao I, Moreno Monge B, Rodríguez-Caderot G, Herraiz M, Radicella SM (2015) Evaluation of precise point positioning accuracy under large total electron content variations in equatorial latitudes. Adv Sp Res 55(2):605–616
[5] Filippo Rodriguez, Roberto Ronchini, Stefano Di Rollo, Eric Guyader, Angela Aragon-Angel, et al..
The Ionosphere Prediction Service For GNSS Users. ITSNT 2018, International Technical Symposium
on Navigation and Timing, Oct 2018, Toulouse, France.
[6] Cesaroni, C., Spogli, L., Aragon-Angel, A., Fiocca, M., Dear, V., De Franceschi, G., & Romano, V. (2020). Neural network based model for global Total Electron Content forecasting. Journal of Space Weather and Space Climate, 10, 11.
[7] European GNSS (Galileo) Open Service—NTCM-G Ionospheric Model Description, Issue 1.0, May 2022, European Commission (EC)
[8] Durazo, J. A., E. J. Kostelich and A. Mahalov, \Local ensemble transform kalman Filter for ionospheric data assimilation: Observation influence analysis during a geomagnetic storm event", Journal of Geophysical Research: Space Physics 122, 9, 9652{9669, URL http://dx.doi.org/10.1002/2017JA024274, 2017JA024274(2017).
[9] Rovira-Garcia A, Juan JM, Sanz J, González-Casado G, Bertran E (2016b) Fast precise point positioning: a system to provide corrections for single and multi-frequency navigation. NAVIGATION J Inst Navig 63(3):231–247. https://doi.org/10.1002/navi.148



Previous Abstract Return to Session B1a Next Abstract