Investigation on the internal acceleration process of the outer radiation belt using the particle filter

It is known that high energy electrons in the radiation belts often cause satellite anomalies and malfunctions. Thus, a forecast of the time variation of the energetic electrons is necessary to protect satellites in the radiation belts. Time variations of the radiation belt electrons have been modeled with the Fokker-Plank equation. Performance of the forecast using the Fokker-Planck equation depends on the parameters used in the model, so that improvement of the parameters is important for the space weather forecast. We performed data assimilation using the particle filter by a code which was developed by Miyoshi et al.[2006]. We prepare 1000 particles used for the calculation. In this study, phase space density, the diffusion coefficient, and wave amplitude, and the source amplitude of the internal acceleration compose the state vector. The observation vector consists of the differential flux measured by the Tsubasa satellite. We also apply the particle smoother to estimate the smoothed distribution. While there were several discrepancies between the simulation without the data assimilation and the observations, the data assimilation improves the simulation result, and captures the typical flux variations of the outer belt during magnetic storms. We also discuss the internal acceleration process on the basis of the source amplitude estimated through the data assimilation.