DISCUSSION Simulation of Radiation Belt Simulation

ULF OSCILLATIONS

In both observations and simulations of the January 1997 event, very limited particle acceleration at MeV energies was seen during the initial phases of the storm. The magnetic cloud was embedded in a solar wind of moderate velocity $\sim400$ km/s, compared to estimates as great as 1400 km/s for the March 24, 1991 event, where significant radiation belt flux enhancement occurred on the time scale of an MeV electron drift period following the SSC. Instead of a high speed interplanetary shock, the organizing feature of the event studied here was an extended period of southward IMF characterized by substorm activity preceeding the rise in relativistic electron fluxes after 09:00 UT on January 10. In the period between about 9 and 12 UT, ground magnetometers located at College and Gakona, Alaska, recorded large amplitude (several hundred nT) oscillations in the magnetic field coincident with the rise in electron flux observed by GPS spacecraft between 4.2 and 4.5 R_E. These waves had periods of around 10 minutes, also seen in riometer and scanning meridian photometer data, corresponding to the drift frequency of, e.g. 1.6 MeV electrons at L=4.2.
  

Figure 1: Auroral absorbtion at 38.6 MHz recorded at Gakona, Alaska by one of the 16 beams of a phased-arrayriometer antenna. The antenna array is phased inly in the meridional direction and covers a range of magnetic latitudes from 61 to 65 degrees. The trace shown is for beam 9 corresponding to the nearly vertical (63 deg) direction. A sequence of long-period (10-15 min) absorbed oscillations, reflecting similar modulation of energetic election precipitation, is evident from 10:30 - 12:30 UT.

The riometer data shown in Figure 1 indicates further enhancement in activity around 11 UT when a moderate solar wind pressure pulse impacted the magnetosphere [Li et al., 1998]. However, this enhancement was clearly embedded in ULF wave activity in the same frequency range over a three hour period at Gakona, during which significant electron flux increase was seen in the simulations at L = 4.

To analyze the effect of ULF oscillations in the simulations, a power spectrum of the model MHD fields has been produced in the various field components.
  

Plate 4 : Power spectrum of the radial component of the MHD simulation electric field in the period between 10:00 UT and 12:16 UT on 10 Jan 98, at the local time of Gakona, AK. The ascending black lines indicate dipole drift frequency as a function of L for particles with energies shown in Figure 1. The descending lines between L = 4.2 and L = 6.6 give the dipole drift frequency of a particle moving radially through this range while conserving the relativistic adiabatic invariant.

A frequency analysis of the simulated radial electric field at the azimuthal location of Gakona (214.85 deg E) is shown in Plate 4, taken over 10:00 - 12:16 UT on January 10. Assuming the oscillations seen are Alvenic in nature, this component of the electric field corresponds to a toroidal-mode field line resonance [Southwood, 1974]. The ascending black lines in the figure indicate the drift frequency of a particle at constant energy, while the descending lines between L=4.2 and L=6.6 indicate the drift frequency seen by a particle moving through this range of L-shells at constant M, where M is the relativistic first adiabatic invariant [Schulz and Lanzerotti, 1974]. Clearly, a particle moving from just inside geosynchronous orbit to the radial distances covered by the GPS spacecraft would encounter significant power in the spectral range matching its drift frequency. This suggests that a drift resonance between the particles and fields might be responsible for the energization observed in both in-situ measurements and in the simulation.

DISCUSSION Simulation of Radiation Belt Simulation