Information

Global Ionospheric Specification

The Global Ionosphere Specification (GIS) data assimilationmodel is developed to provide 3-D electron density profiles as a near real-time data product of FORMOSAT-7/COSMIC2 (F7/C2) mission.
The GIS assimilates slant total electron content (TEC) observed from ground-based Global Navigation Satellite System (GNSS) receiving stations and space-based radio occultation (RO)measurements by F7/C2,by using Gauss-Markov Kalman filter, into international reference ionosphere (IRI) model to providecontinuous three-dimensional ionospheric electron density distribution (Lin et al., doi:10.1002/2017JA024185, 2017). Sunspot number and IG index are used as ensemble parameters in calculating the background error covariance values. In the first time-step run, background IRI density and error covariance are used as initial the forecast values, which are then updated with the measurements. The Kalman filter is implemented in the forecast step by allowing the assimilation system to evolve from current time step to next time step by longitudinally shifting ionospheric electrondensity at the rate of the Earth’s rotation. To take care of climatological variation, a 10% relaxation to the time advanced density is implemented by using background electron density predicted by IRI.
The forecast and measurement update cycles are repeated every 1 hour, providing hourly GIS density profiles with a latitude/longitude resolution of 5 x 2.5°, and vertical resolution of 20 km. The horizontal domain covers 90°S to 90°N latitudes and 180°W to 175°E longitudes, and the altitude region is from 100 km to the F7/C2 orbit.

Figure 1 illustrates the operational GIS data assimilation system. In the first step, the accumulated RO and GNSS TEC data for 1 the hour period, as well the background error covariance values are loaded. The system also runs the IRI model to provide the initial background. The system accumulates RO and GNSS TEC for each 1-hour period. The GNSS TEC is corrected for plasmaspheric contribution by subtracting the plasmaspheric electron content calculated using an empirical model based on day of year, latitude, longitude, hour and F10.7 solar flux. As the F7/C2 satellites are not yet completely separated into their final mission configuration, the number of available RO TEC points in one-hour period varies from 15,000 – 25,000. The assimilation system currently uses about 1300 GPS receivers, and in one-hour there are about 25,000 ground TEC values available for assimilation.
As it takes more than 24 hours for the GNSS TEC from all the stations to become available, the GIS is currently run with about 2-day delay. Step 2 includes the implementation of Gauss-Markov Kalman filter model, where the measurements are used to update forecasted values of density and error covariance. The updated density values will be available to step 3, where it is stored as 3D electron density values in netCDF format for the current time step. Meanwhile, the updated density and error covariance are used to forecast the values for the next time step.
The update and forecast cycles are repeated every hour. The assimilation system is built based on Fortran using MPI libraries, and once initiated, it will complete the run for 24-hour period. The locations of input data, as well as output path could be passed to the program using a name-list file.

 

Figure 1 Operational GIS data assimilation system. The system accumulates RO TEC for every 1 hour

With horizontal and vertical ionospheric information obtained from ground- and space-based measurements this data product can study large-scale ionosphericelectron density variations. Figure 2 shows the wave-4 modulations revealed in GIS density on 19 July 2019 at 1300 LT. Note that by using traditional Abel inversion density profiles, it requires to combine several days of data to generate such a global electron density map. However, GIS can provide such global density variations every day at each hour. The GIS offers the advantage of studying day-to-day ionospheric variability, particularly over low latitudes where the F7/C2 measurements are focused.

Figure 2 Wave-4 peaks detected revealed by GIS density profiles. The map is generated by using 1 hour of F7/C2 and GNSS TEC on 19 July 2019, at 1300 LT.