Numerical estimates of drift loss and Dst effect for outer radiation belt relativistic electrons with arbitrary pitch angle

Satellite observations often show that relativistic electron fluxes that decrease during a geomagnetic storm main phase do not recover their prestorm level even when the storm has substantially recovered. A possible explanation for such sustained flux dropout is that the electrons that move to larger shells (L shells) aided by the disturbance storm time (Dst) effect associated with the main phase geomagnetic field depression may be suffering drift loss to the magnetopause, resulting in irreversible (nonadiabatic) flux decreases during a geomagnetic storm. In this study, we have numerically evaluated the drift loss effect by combining it with the Dst effect and including off-equatorially mirroring electrons for three different storm conditions obtained by averaging 95 geomagnetic storms that occurred from 1997 to 2002. Using the Tsyganenko T02 model and our own simplified method, we estimated the storm time flux changes based on the guiding center orbit dynamics. Assuming that there is no competing source mechanism taking place at the same time, our calculations of the electron fluxes at equatorial midnight suggest that the drift loss when combined with the Dst effect can be responsible for flux dropouts, which can be seen even inside the geosynchronous orbit during storm periods. Specifically, by evaluating omnidirectional flux values at three specific times that correspond to the storm onset time, the time of minimum Dst value, and the end of the Dst recovery, we have obtained the following numerical results. First, for the strong storm with -150 nT < Dst(min) <= -100 nT, the combined drift loss and Dst effect can cause a complete dropout of the electron flux for r >=similar to 5R(E) at the end of the storm recovery. A nearly full recovery of the particle flux by the adiabatic Dst effect is seen only for r <similar to 5R(E). For the moderate storm with -100 nT < Dst(min) <= -50 nT, the overall flux decrease level at the end of the storm recovery is less significant compared to that of the strong storm. However, the combined loss effect can still penetrate into r similar to 5 R-E, leading to some partial dropout of the flux. For the severe storm with Dst(min) <= -150 nT, the flux dropout is far more significant than for the other two storms, indicating that the combined drift loss and Dst effect can even reduce the flux level at an inner region of r similar to 4 R-E. But in this case, the solar wind dynamic pressure is so high that the dayside magnetopause can cross the geosynchronous orbit. Consequently, the flux dropouts seen in actual observations can be primarily attributed to a fast and direct loss to the magnetopause at times when the magnetopause crosses the geosynchronous orbit. It is possible that our numerical results may quantitatively change to some extent with different magnetospheric models and assumptions and may also change depending on the validity of the fully adiabatic invariants assumption.