GNSS radio occultation (RO) derived electron density quality in high latitude and polar region: NCAR-TIEGCM simulation and real data evaluation
|Author||Yue, Xinan; Schreiner, William; Kuo, Ying-Hwa; Wu, Qian; Deng, Yue; Wang, Wenbin;|
|Keywords||Abel inversion; AURORA; COSMIC; Electron density; GNSS radio occultation; TIEGCM|
Global Navigation Satellite System (GNSS) based radio occultation (RO) technique has shown powerful ability in ionospheric electron density profiling in the past decade. The most frequently used Abel inversion method in electron density retrieval has some biases because of the used spherical symmetry assumption. Our previous series simulations and evaluations mainly concentrated in the middle and low latitude regions have shown some systematical bias especially in lower altitude of low latitude region. However, the RO derived electron density quality in the high latitude and polar region is rarely investigated and not quantitatively clear yet. In this study, the Abel inversion error over high latitude and polar regions are systematically investigated for the first time based on NCAR-TIEGCM simulations and real data evaluations. The TIMED data driven NCAR-TIEGCM modeled electron density during 2008 are used to simulate the COSMIC RO events. The Abel inversion error can then be estimated by comparing Abel retrievals from TIEGCM simulated occultation with the original TIEGCM simulations. The Abel inversion can reproduce the season, altitude, latitude, and local time variation patterns of electron density and auroral zone electron density nighttime enhancement well in high latitude and polar region. The Abel inversion tends to underestimate the electron density in the auroral zone and overestimate it on both the equatorward and poleward sides of the auroral zone. As simulated by the TIEGCM model, the significant relative error (\>25\%) mainly occurs in lower altitude (\<250\ km) inside and around auroral zone region. Above 250\ km, the relative error mostly is less than 25\%. Specifically, RMSE (root mean square error) of NmF2 error from simulation is \~8.5\%. The Abel error under real ionosphere situation would be worse because the ionosphere could be more complicated and noisier than the model simulation. The error distribution and its seasonal, local time and latitude variations can be explained by the spherical symmetry assumption used in the Abel inversion associated with the corresponding ionospheric electron density variations. The comparisons between PFISR and COSMIC RO electron density during 2007\textendash2011 and some previous validation studies agree well with our simulation results. We hope these results can stimulate more studies in high latitude ionospheric research using RO data.
|Year of Publication||2013|
|Journal||Journal of Atmospheric and Solar-Terrestrial Physics|
|Number of Pages||39-49|