3The F3/C and F7/C2 RO uniformly/globally observe ionospheric electron density profiles, while worldwide ground-based GPS receivers derive the TEC. This team employs both electron density profiles and TEC data advance and develop models to study structures/dynamics and observe novel features of the ionospheric space weather.

Monitoring Model

Global ionospheric maps (GIM) have been daily and routinely published by Center for Orbit Determination in Europe (CODE), European Space Operations Centre (ESOC), NASA’s Jet Propulsion Laboratory (JPL), International GNSS Service (IGS), and Technical University of Catalonia (UPC) to estimate ionospheric effects on radio propagation and the provide correction for single frequency GNSS navigation (Fig. 3).  Global vertical TEC observations from one to two hundreds ground-based receivers and from satellites, like TOPEX/Jason-1, are represented as single layer with spherical harmonic expansion [Schaer et al., 1996].  On the other hand, RO observational systems, such as F3/C and F7/C, can provide global TEC observations as well. Unfortunately, the massive RO dataset has not been merged into GIM systems yet.

Figure 3. An example of CODE GIM, one of the five major organizations published GIM.

(ftp://cddis.gsfc.nasa.gov/gps/products/ionex/).

F3/C mission conducts global scintillation observation with constantly operation period covering solar minimum and maximum.  The F3/C constellation continuously recording the 50-Hz sampling rate GPS L band signal to noise ratio, and calculated them into 1-Hz sampling rate S4-index.  The TACC (Taiwan Analysis Center for COSMIC) provides over 3000 S4-index profiles daily with uniformly spatial distribution and high vertical resolution (about 1km).  Due to the difference of observation elevation angle, the scintillation from RO and ground-based observation are unlikely the same.  To solve this crucial problem, Liu et al. [2015] develop a conversion method that simulates ground-based scintillation observation from RO one.  With the aid of this conversion geometry, we can finally initiate the study of global L-band scintillation by the converted S4-index, and conduct the F3/C global S4 model construction.  With F3/C’s high spatial and temporal resolution, especially upon the oceanic region that cannot be provided by ground-based observation, the F3/C S4-index can practically update existing scintillation model.  The updated empirical scintillation model called F3/C S4 scintillation model, which involved the F3/C RO S4-index data, can provide better quality and accuracy of worldwide satellite signal scintillation distribution.

Nowcast Model

Global Electron Data Assimilation model (GEDA model) have conducted a preliminary model to assimilate two different types of total electron content (TEC) observations from ground-based TEC and space-based TEC measured by F3/C based on the Gauss-Markov Kalman filter measurement update step with the IRI model as the background model [Lin et al., 2015].  The preliminary ionospheric data assimilation model can produce the 3D global electron density structures (Fig. 4).  GEDA model will be conducted the Kalman filter forecast step to nowcast reliable 3D electron density structures.  In the Kalman filter, the assimilated electron density structure is expected to become more realistic over time though a recursive application of update and forecast steps in theory.

Figure 4. Global 3D electron density structure, NmF2, and hmF2 of ionospheric data assimilation model.

Hirooka et al. [2012] apply the Residual Minimization Training Neural Network (RMTNN) to investigate the three-dimensional structure of electron density in the ionosphere prior to the 2007 Southern Sumatra earthquake (M8.5).  Yue et al. [2012] report a global 3D ionospheric electron density reanalysis based on multisource data assimilation.  The data include global navigation satellite system (GNSS) observations of ionospheric total electron content (TEC) from ground-based stations, ionospheric radio occultations by CHAMP, GRACE, COSMIC, SAC-C, Metop-A, and the TerraSAR-X satellites, and Jason-1 and 2 altimeter TEC measurements.

Forecast Model

The ionospheric drivers that are crucial to the specification of the ionospheric weather involve not only solar radiation and high-latitude electric fields and particle precipitation but also the thermospheric composition, temperature, and winds.  To carry out a long-term forecast, the neutral drivers in the atmosphere should be taken into consideration.  To develop the Electron-Neutral Data Assimilation model (GENDA model), observations of F3/C and F7/C2 as well as IGS TEC are assimilated into a neutral-ion coupled model by using the Ensemble Kalman Filter (EnKF).  The EnKF is proposed as a Monte-Carlo approximation of a KF [Evensen, 1994], which allows the use of a fully nonlinear dynamical model as a forecast model.  In EnKF, the background error covariance is represented by the ensemble covariance calculated from an ensemble of model forecasts, and therefore the space- and time-dependent description of the background error covariance is attained with a reasonable computational cost.  Once observations are ingested into the forecast model, adjusted state variables are estimated by optimally combining the observational and a model. These state variables are consistent with underlying physics and chemistry of the real ionosphere as much as possible given uncertainty of the model. Validating to independent observation, this procedure can improve the ionospheric electron density structure (Fig. 5) as well as unobserved model state variables, such as neutral temperature and density [Lee et al., 2012, Matsuo et al., 2013].  Moreover, the adjusted state variables should include electron density, atomic oxygen ions, neutral temperature and winds are further carried back into the model to forecast for next time steps based on laws of time evolution and physical properties.  This forecast results can provide more reliable status of ionosphere and thermosphere and much longer forecasted period since the ionospheric and neutral atmospheric drivers are taken into account in this procedure (Fig. 15) [Hsu et al., 2014].

Figure 5. Global NmF2 maps before and after assimilating the F3/C observations.  The black dots indicate the observation locations [Lee et al., 2012].

References

Evensen, G. (1994), Sequential data assimilation with a nonlinear quasigeostrophic model using Monte Carlo methods to forecast error statistics, J. Geophys. Res., 99, 10,143–10,162.

Fuller-Rowell, T., E. Araujo-Pradere, C. Minter, M. Codrescu, P. Spencer, D. Robertson, and A. R. Jacobson (2006), US‐TEC: A new data assimilation product from the Space Environment Center characterizing the ionospheric total electron content using real-time GPS data, Radio Sci., 41, RS6003, doi:10.1029/2005RS003393.

Hirooka, S., Hattori, K., Nishihashi, M., Kon, S. and Takeda, T. (2012), Development of ionospheric tomography using neural network and its application to the 2007 Southern Sumatra earthquake. Elect. Eng. Jpn., 181, 9–18. doi:10.1002/eej.22298.

Hsu, C.-T., T. Matsuo, W. Wang, and J.-Y. Liu (2014), Effects of inferring unobserved thermospheric and ionospheric state variables by using an Ensemble Kalman Filter on global ionospheric specification and forecasting, J. Geophys. Res. Space Physics, 119, 9256–9267, doi:10.1002/2014JA020390.

Lee, I. T., T. Matsuo, A. D. Richmond, J. Y. Liu, W. Wang, C. H. Lin, J. L. Anderson, and M. Q. Chen (2012), Assimilation of FORMOSAT-3/COSMIC electron density profiles into a coupled thermosphere/ionosphere model using ensemble Kalman filtering, J. Geophys. Res., 117, A10318, doi:10.1029/2012JA017700.

Lee, I. T., H. F. Tsai, J. Y. Liu, C. H. Lin, T. Matsuo, and L. C. Chang (2013), Modeling impact of FORMOSAT-7/COSMIC-2 mission on ionospheric space weather monitoring, J. Geophys. Res. Space Physics, 118, 6518–6523, doi:10.1002/jgra.50538.

Lin, C. Y., T. Matsuo, J. Y. Liu, C. H. Lin, H. F. Tsai, and E. A. Araujo-Pradere (2015), Ionospheric assimilation of radio occultation and ground-based GPS data using non-stationary background model error covariance, Atmos. Meas. Tech., 8, 171–182, doi:10.5194/amt-8-171-2015.

Liu, J.Y., S.P. Chen, W.H. Yeh, H.F. Tsai, P.K. Rajesh, (2015a). Worst-case GPS scintillations on the ground estimated from radio occultation observations of FORMOSAT-3/COSMIC during 2007-2014, Surveys in Geophysics, doi: 10.1007/s10712-015-9355-x.

Liu, J.Y., C.Y. Lin, H.F. Tsai, (2015b). Electron density profiles probed by radio occultation of FORMOSAT-7/COSMIC-2 at 520 and 800 km altitude, Atmospheric Measurement Techniques, doi:10.5194/amt-8-3069-2015.

Matsuo, T., I.-T. Lee, and J. L. Anderson (2013), Thermospheric mass density specification using an ensemble Kalman filter, J. Geophys. Res. Space Physics, 118, doi:10.1002/jgra.50162.

Schaer, S., G. Beutler, M. Rothacher, and T. Soringer, Daily global ionosphere maps based on GPS carrier phase data routinely produced by the CODE Analysis Center, in Proc. IGS Analysis Center Workshop, Silver Spring, Maryland, March 19 – 21, 19.

Yue, X., et al. (2012), Global 3-D ionospheric electron density reanalysis based on multisource data assimilation, J. Geophys. Res., 117, A09325, doi:10.1029/2012JA017968.