Open AccessReanalysis of relativistic radiation belt electron fluxes using CRRES satellite data, a radial diffusion model, and a Kalman filter
Author(s) -
Shprits Yuri,
Kondrashov Dmitri,
Chen Yue,
Thorne Richard,
Ghil Michael,
Friedel Reiner,
Reeves Geoff
Publication year - 2007
Publication title -
journal of geophysical research: space physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/2007ja012579
Subject(s) - physics , van allen radiation belt , computational physics , electron , satellite , radiation , kalman filter , data assimilation , acceleration , diffusion , meteorology , classical mechanics , computer science , optics , magnetosphere , plasma , quantum mechanics , astronomy , artificial intelligence , thermodynamics
In this study we perform a reanalysis of the sparse MEA CRRES relativistic electron data using a relatively simple one‐dimensional radial diffusion model and a Kalman filtering approach. By combining observations with the model in an optimal way we produce a high time and space resolution reanalysis of the radiation belt electron fluxes over a 50‐d period starting on 18 August 1990. The results of the reanalysis clearly show pronounced peaks in the electron phase space density (PSD), which can not be explained by the variations in the outer boundary, and can only be produced by a local acceleration processes. The location of the innovation vector shows that local acceleration is most efficient at L * = 5.5 for electrons at K = 0.11 G 0.5 R E and μ = 700 MeV/G. Sensitivity numerical experiments for various values of μ and K indicate that peaks in PSD become stronger with increasing K and μ . To verify that our results are not affected by the limitations of the satellite orbit and coverage, we performed an “identical twin” experiments with synthetic data specified only at the locations for which CRRES observations are available. Our results indicate that the model with data assimilation can accurately reproduce the underlying structure of the PSD even when data is sparse. The identical twin experiments also indicate that PSD at a particular L ‐shell is determined by the local processes and cannot be accurately estimated unless local measurements are available.