Simulation of the March 24, 1991, storm sudden commencement (SSC) has illuminated the rapid formation of new radiation belts on the particle drift time scale. While this event was the most dramatic in terms of radiation belt effects of the last solar maximum, comparable signatures of such events were seen in 1962 by Explorer 15 and in 1986 by DMSP. Several smaller MeV proton events with comparable particle morphology, but less radial transport and energization, were observed during the lifetime of the CRRES satellite (July 1990 - October 1991), which was well instrumented for both particle and field measurements inside geosynchronous orbit. Typically a solar proton event is accompanied by an SSC in the CRRES data set, while the converse is not always true. An SSC accompanied by solar protons produces a trapped population which remains on closed drift orbits until ring current buildup disrupts trapping, either by violation of the adiabatic trapping criterion or generation of waves whose frequency is comparable to periodic particle motion. Thus, new radiation belts formed around L=4 by SSC injection are short-lived compared to the March 24, 1991, storm, wherein solar protons were transported radially inward to L=2.5, with greater energization corresponding to first adiabatic invariant conservation than for the weaker events.

INTRODUCTION

Storm sudden commencement (SSC) compressions of the dayside magnetopause have been shown to produce radiation belt electron and proton flux enhancements which can be very long lived inside of L=3, as for the March 24, 1991, geomagnetic storm near the last solar maximum (Vampola and Korth, 1993; Blake et al., 1992a, b; Looper et al., 1994), also for the February, 1986, storm near solar minimum (Gussenhoven et al., 1989). Several smaller events with similar morphology, producing a trapped proton population around L=4 following an SSC, were observed during the lifetime of the CRRES satellite from 25 July 1990 to 12 October 1991 (Gussenhoven et al., 1994). These events have been interpreted in terms of radial acceleration of solar protons inward by the induction electric field associated with the SSC (Hudson et al., 1997). This mechanism was first applied to electron acceleration during the March 24, 1991, event (Li et al., 1993a), next to proton acceleration for this event (Hudson et al., 1995), and then to electron acceleration for a weaker event on August 18, 1991 (Brautigam, 1995). The SSC produces a magnetosonic pulse which propagates ahead of the magnetopause compression, modeled by Li et al. (1993a) as a constant velocity (across L) impulse in B_z. The pulse reflects off of the ionosphere producing a bipolar pulse in azimuthal E, spreading around the flanks of the magnetosphere from the longitude of interplanetary shock impact. This model has been generalized by use of MHD fields from global 3D simulations of the interplanetary shock impact on the magnetosphere for varied solar wind velocity and IMF (Hudson et al., 1997).

This paper addresses the shorter lifetime of weaker proton injection events seen in the CRRES data set, relative to the March 24 storm.

OBSERVATIONS

Figure 1 shows an example of solar energetic proton (SEP) injection at 10.7 MeV on August 27, 1991 (orbit 961), as measured by the Protel instrument on CRRES. This instrument makes a 24-point spectral measurement in 1-2 MeV channel widths covering the energy range 1-100 MeV each second (Violet et al., 1993). The injection inside L=4 followed a moderate SSC with about a 25 nT increase in hourly averaged Dst shown in Figure 2, as compared with the March event, which produced the largest spike in H component of low latitude ground magnetograms ever reported (234 nT at Memambetsu, geomagnetic latitude 34.8^o N (Araki et al., 1996)). The newly trapped solar protons at L ~ 4 remain after the SEP event has subsided, then disappear when Dst drops abruptly on August 30 (CRRES orbit 967) to -110 nT.

Fig. 1. Proton fluence in 10.7 MeV channel of Protel instrument on CRRES (Violet et al., 1993), averaged over 0.05 R_E L-bins for each 10 hour orbit from 1 August to 13 September 1991. Encompasses solar energetic proton (SEP) event and SSC at 1515 UT on August 27 (orbit 961), with formation of short-lived trapped population at L=4, which is distinct from the diminishing SEP population.

Fig. 2. Plot of hourly averaged Dst for 25 August to 1 September 1991.

Fig. 4. Plot of hourly averaged Dst for 24 July to 16 August 1990.

Figure 3 shows the 10.7 MeV proton fluence for another example, with SEP/SSC injection on August 1, 1990 (orbit 17), and depletion of the newly trapped belt on August 14 (orbits 48-49) (Gussenhoven et al., 1994; Hudson et al., 1997). The August 1 injection followed a major storm on July 27 prior to turn on of the Protel instrument on CRRES (July 29, orbit 7). Trapping of solar protons was not evident for this earlier event, which had a softer proton spectrum at geosynchronous as seen in NOAA summary plots of GOES 7 data than the August 1 event, e. g. flux of > 10 MeV protons was 20 vs. 200 (cm²-s-sr)^-1. However, Protel was not turned on until after the very large Dst drop apparent in Figure 4, and only the decay of the solar proton event is evident in Figure 3. The depletion of trapped proton flux following the August 1 injection was delayed with respect to a minor solar proton event and coincided with an oscillatory buildup in Dst seen in Figure 4, rather than a major storm.

Fig. 3. Proton fluence in 10.7 MeV channel of Protel instrument on CRRES, same format as Figure 1, for 29 July through 25 August 1990, encompassing two complete SEP event, SSC at 0741 UT (around orbit 15 perigee) and formation of a new proton belt on August 1 and depletion on August 14, 1990 (Gussenhoven et al., 1994).

June and July of 1991 were extremely active periods at the sun, with multiple SEP/SSC events as seen by the GOES geosynchronous spacecraft and CRRES (Hudson et al., 1997), and by LANL instruments on USAF geosynchronous spacecraft. The latter observed multiple magnetopause crossings, indicating that the magnetopause was compressed inside geosynchronous orbit for many hours over six time intervals in June and on July 8 (McComas et al., 1994, Table 1). Figure 5 summarizes the 10.7 MeV proton fluence measured by CRRES from May 1 to July 23, 1991. In addition to SEP events (and data dropout around orbit 720), a depletion in flux on June 5 (orbit 767) followed the SSC at 0337 UT on June 4 (orbit 763) and coincided with buildup in Dst seen in Figure 6a (orbit 767). The subsequent increase of solar proton flux leading up to the SSC at 1012 UT on June 12 (orbit 783) resulted in an injection of trapped 10.7 MeV protons apparent in Figure 5, which filled in the preceding depletion. Another SEP/SSC injection event occurred on July 8 at 1636 UT (orbit 845), followed by a depletion associated with the Dst decrease on July 9 (orbit 847) with minimum around -195 nT. A further depletion down to L=3 occurred as Dst approached another deep minimum (-185 nT) on July 13 (orbit 856) (Gussenhoven et al., 1994; Figure 6b). The absence of a significant solar proton source population for the SSC at 1012 UT on July 12, as compared with the July 8 SSC (Gussenhoven et al., 1994), and the weaker SSC compression suggested by both Dst (Figure 6b) and LANL geosynchronous data which indicated no magnetopause crossing (McComas et al., 1994), resulted in no significant injection for the July 12 event, in contrast with the July 8 event.

Fig. 5. Proton fluence in 10.7 MeV channel of Protel instrument on CRRES, same format as Figure 1, for orbits 680 to 880 (1 May through 23 July 1991). There is a decrease in maximum L of 10.7 MeV protons on orbit 767 (June 5), and increase on orbit 783 (June 12), coincident with SSC's noted in text. A similar decrease in maximum L was seen on orbit 846 (9 July) and orbit 856 (13 July) (Gussenhoven et al., 1994).

Fig. 6. (a) Plot of hourly averaged Dst for 31 May to 7 June 1991; (b) same for 6 July to 14 July 1991.

By contrast to the active June - July 1991 time period, Figure 7 shows a relatively long-lived new belt which formed on February 1, 1991, at L=3.5-4 due to an SSC at 1842 UT (Hudson et al., 1997), and survived the ensuing Dst=-80 nT storm, followed by a relatively long quiet period leading up to the March 24 event. Compared with the two trapping events shown in Figures 1 and 3, the 10.7 MeV protons penetrated deeper into L=3.5 for the February 1 event, indicating a stronger injection. The fluence dropped by half an order of magnitude on March 5 (orbit 542), coincident with the buildup of Dst to a maximum of -55 nT, diminishing further with a second buildup to -67 nT on March 10 (orbit 553). Again, the flux decrease is better correlated with Dst than the timing of the preceding SSC's (1619 UT on March 4, orbit 540, and 2245 UT on March 9, early in orbit 553).

Fig. 7. Proton fluence in 10.7 MeV channel of Protel instrument on CRRES, same format as Figure 1, for orbits 455 to 555 (28 January through 11 March 1991) (Hudson et al., 1997).

Examining IMP 8 data where available for the events described here, we find that both the August 27 and August 30, 1991, storms (Figures 1 and 2) were characterized by interplanetary shocks, but with a larger solar wind speed (550 km/s vs. 450 km/s) and slightly stronger (-20 nT vs. -15 nT) and more prolonged southward IMF in the latter case. These differences are reflected in the greater Dst response of the magnetosphere seen in Figure 2, and loss of the newly trapped proton belt seen in Figure 1, coincident with the second storm. Both the August 1, 1990, trapping event and loss on August 14 (Figures 3 and 4) were characterized by interplanetary shocks of comparable solar wind speed ~ 500 km/s, however IMF B_z was positive for the injection event on August 1 and oscillatory, around +/-10 nT, for the depletion on August 14. No IMP 8 solar wind data were available for the large pre-CRRES storm on July 28, 1990, likewise they were unavailable for the February 1 and March 24, 1991, events. One cannot, therefore, compare solar wind conditions for February 1 with March 4 and March 9, 1991, where IMP 8 data were available, in order to understand additional factors which made the Dst=-80 nT storm of February 1 tolerable, while depletion occurred on March 5 and 10 when Dst dropped to the -60 nT range (not shown).

In June and July, 1991, the IMP 8 solar wind data were complex and intermittent. June 4 IMP 8 data showed an interplanetary shock around 1600 UT followed by southward B_z=-20 nT, with a data gap at the time of the 0337 UT SSC. More notable was the high solar wind speed of 850 km/s associated with the June 12 injection event, while the July 8 SSC was preceded by an interplanetary shock of nominal solar wind speed (400 km/s), followed by southward B_z > -20 nT, decreasing in magnitude over 9 hours. This latter event produced the first storm evident in Figure 6b, and associated proton depletion in Figure 5. No IMP 8 data were available for the July 12 SSC and storm.

Returning to the great storm of March 24, 1991, Figure 8 shows the proton differential number flux perpendicular to the local magnetic field direction vs. energy interpolated between discrete energy channels of Protel for (a) the descending leg of the orbit in which the SSC occurred, and (b) the ascending leg of the subsequent orbit. Solar protons penetrate into L=4 in (a), with trapped inner zone protons apparent in the lower left corner. The inner edge of the newly formed radiation belt is visible at L=2.5 immediately following the SSC, on the time scale of 25 MeV proton drift from the shock impact longitude (Blake et al., 1992a, b; Hudson et al., 1995). The first outbound pass of CRRES subsequent to the SSC injection (b) showed a broad peak in proton flux at energies of 20-40 MeV around L=2.5, and a secondary peak near L=3.2. The > 15 MeV electron count rate in the region from L ~ 2-3 showed strong oscillations with a period of around one minute. These oscillations occurred too long after the injection event to be interpreted as drift echoes, and have been interpreted as due to ULF waves associated with sudden impulses subsequent to the SSC (Hudson et al., 1996b). The secondary proton peak had vanished when CRRES next crossed the L=3.2 region some five hours later inbound. It is possible that this secondary peak in flux evident in Figure 8b was due to an injection that did not survive the ensuing Dst=-300 nT storm, although temporal effects involving the accompanying ULF waves are also a possible explanation for its disappearance on the time scale of a few hours.

Fig. 8. Proton differential number flux perpendicular to local magnetic field direction vs. energy interpolated between discrete energy channels of Protel for (a) descending leg of orbit 587 in which March 24, 1991, SSC occurred, and (b) ascending leg of orbit 588.

DISCUSSION

We first briefly describe the mechanism which has been proposed for radial transport and acceleration of solar protons to form a new trapped population (Hudson et al., 1995; 1997). Next we consider mechanisms by which the newly trapped protons abruptly disappear at a fixed energy and L, as seen in Figures 1, 3, 5, and 7.

An SSC accompanied by solar protons produces a trapped population from those protons which experience significant inward radial transport and perpendicular acceleration due to the induction electric field associated with the SSC magnetic field compression. The March 24, 1991, SSC demonstrated that new radiation belts can form on the MeV particle drift time scale, which coincides with the SSC time scale, or tens to a few hundred seconds, in contrast to the hours to days of the geomagnetic storm. The induction electric field which accompanies magnetopause compression transports particles radially inward, increasing energy with conservation of the first adiabatic invariant. The mechanism is resonant over a range of initial energies, or equivalent magnetic moments, such that protons (or electrons) gradient-drifting either too slowly or too fast relative to the azimuthal spreading of the pulse do not spend an optimum amount of time in the region of maximum pulse amplitude (Hudson et al., 1997). The pulse simply pushes cold plasma in and then out again, while very energetic protons (> 25 MeV) are hardly affected by the pulse as they pass through it, with relativistic electrons which drift more slowly for the same energy affected up to about 50 MeV (Li et al., 1993a; Ginet et al., 1994). The particle drift includes both gradient B and E B contributions in the equatorial plane. The azimuthal electric field component E produces inward (and outward) radial motion, while the generally smaller radial E_r, neglected in the analytic pulse model (Li et al., 1993a) and included in MHD-field driven simulations (Hudson et al., 1997), produces azimuthal acceleration. The gradient B drift actually reverses direction where the pulse magnetic field gradient exceeds that of the primarily dipole background field, allowing more energetic particles whose drift motion is dominated by this energy-dependent drift component to stay longer in the region where the pulse amplitude is large (Hudson et al., 1995).

The analytic model of the SSC compression developed by Li et al. (1993a) demonstrates the basic mechanism of induction electric field acceleration, while Hudson et al. (1997) have shown that a substantially smaller pulse field amplitude than used to model the March 24, 1991, event (40 vs. 240 mV/m) can produce radial transport by 1 to 2 R_E inside the SEP penetration boundary. A high speed interplanetary shock first preloads the magnetosphere with energetic solar protons, then provides a mechanism for further acceleration and trapping via inward radial transport, which increases the perpendicular to parallel energy ratio. The March event had both a more energetic, higher flux source population, penetrating into lower L, and a larger induction electric field for further acceleration within the magnetosphere than the weaker events. Once trapped at L=2.5, the new proton (and electron) belt was long lived compared to trapping events around L=3-4 shown in Figures 1, 3, 5, and 7.

In order to understand the loss of newly injected protons as Dst builds up, we have first evaluated the trapping parameter for equatorially mirroring protons in a dipole field (cf. Northrop, 1963; Schulz, 1991):

Fig. 9. Plot of outer boundary in L vs. energy for onset of stochastic instability and loss of adiabatic trapping, with =0.187 as an upper limit. =3/4 maximum value for access to trapping with sufficiently small rigidity (Fermi, 1949) is also shown. Outer boundary for trapped protons in Figures 3 and 8 (dots), Figures 1, 5 (squares), and 7 (triangle) is indicated, with both flux peaks shown for Figure 8 (30.9 MeV).

More generally, for a non-dipole magnetic field the trapping parameter is (Schulz, 1991)

evaluated at the magnetic equator. The two magnetic gradient scale lengths are both equal to 3/LR_E for a dipole field. We have calculated the change in the perpendicular gradient scale length as a function of Dst using the Schulz (1996) model for ring current - Dst radial dependence in the magnetic equatorial plane, and find a 15% increase for Dst=-50, 32% for Dst=-100, and 55% for Dst=-200 nT. This scale length enters Eq. (2) linearly as a decrease in . Other factors such as change in solar wind dynamic pressure and IMF B_z also affect the magnetic field gradient scale length at L=4, as incorporated by Tsyganenko (1987, 1989). While the radial magnetic field gradient scale length increases for a proton at fixed drift invariant, the parallel scale length decreases, becoming infinitesimal for a ring current confined to the equatorial plane. Since the parallel and radial scale lengths are equal in a dipole field, and the perpendicular scale length increases while the parallel decreases due to ring current perturbation, the parallel scale length is smaller in the vicinity of maximum ring current strength. The net effect is an increase in epsilon which may violate the adiabatic trapping criterion, leading to stochastic motion, pitch angle diffusion and loss to the atmosphere. Using a model for the ring current magnetic field perturbation at the equator (Schulz, 1996) and symmetry arguments yields a 6.8, 14, and 24% decrease in parallel scale length for Dst=-50, -100, and -200 nT, respectively, with corresponding linear increase in .

CONCLUSIONS

While the SSC injection mechanism has been discussed and modelled in related papers (Li et al., 1993a; Hudson et al., 1995; 1996a, b; 1997), this paper presents first results on the loss mechanism of newly trapped solar protons. Trapping occurs on the time scale of an SSC, given a pre-existing solar energetic proton source population in the outer magnetosphere. Both the SSC and the SEP event are ultimately caused by an interplanetary shock which may be traceable to a coronal mass ejection. Higher solar wind speed events such as February, 1986, where the IMP 8 magnetometer went off-scale (> 1300 km/s), June 12, 1991 (850 km/s, measured by IMP 8), and the solar wind speed inferred for the March 24, 1991, event (1400 km/s, Shea and Smart, 1993), produce injections with deeper penetrations into the magnetosphere. All of this suggests that solar wind dynamic pressure determines the effectiveness of penetration in L and corresponding energization, as has been shown in global MHD simulations of such events (Hudson et al., 1997).

Regardless of the strength of injection in L and energization, the outer boundary of newly trapped protons in energy-L space has been shown to fall just below the =0.187 boundary for a dipole field in the CRRES cases examined. Subsequent to injection, perturbations of the magnetic field due to ring current buildup can be of sufficient magnitude to cause loss of adiabaticity. This has been quantified by calculating the decrease in radius of curvature which yields a linear increase in (2) of 14% when Dst drops to -100 nT. The new proton belt trapped at L=2.5 for the March 24 injection was robust enough to survive a Dst=-300 nT storm. All other examples found in the CRRES data set are closer to the peak in ring current magnetic field perturbation, i. e. in the L=3.5-4.5 range, and subject to loss by the mechanism proposed here.

This survey has provided a general explanation for the complex behavior of the proton trapping boundary evident in Figure 5, previously noted by Gussenhoven et al. (1994) for the July 1991 time period. A drop in flux inside L=3 occurred on June 5 (orbit 767) when the magnetic field was disturbed by buildup of the ring current reflected in Dst (Figure 6a). This depletion in flux was filled in by an SSC injection event on June 12, associated with the highest solar wind speed shock of injection events found during the lifetime of CRRES (850 km/s), excluding March 24, 1991. Subsequent depletions evident in Figure 5 were associated with the ring current injections indicated by Dst buildup on July 9 and 13 (Figure 6b). As previously noted by Gussenhoven et al. (1994), there was no significant solar proton source population for the July 8 SSC, which might otherwise have led to injection.

Extended to other events during the lifetime of the CRRES satellite, we find several cases where a trapping event associated with one SSC is disrupted by a subsequent storm (August 1, 1990; February 1, 1991 and August 27, 1991, injections). In other cases, one finds an impressive storm such as July 27-28, 1990, with no evidence for solar proton trapping. It appears likely that even with injection, a very strong magnetic storm may so perturb the magnetic field environment through ring current buildup that significant long term trapping does not occur. However, the late July 1990 storm had a lower flux of energetic (> 10 MeV) solar protons as seen by GOES 7 than injection events discussed here, supporting the conclusion that an adequate solar proton source population was absent for this event. The most impressive storm of the last solar maximum occurred in March 1989, where Dst reached -600 nT (Allen et al., 1989), however the energetic solar proton flux was lower than for the March 1991 storm, as seen in GOES summary plots (Hudson et al., 1997). With the absence of interplanetary shock initiation as well, no evidence was reported of a new radiation belt associated with the March 1989 storm.

The weaker injection events are relatively short-lived compared with the March 1991 storm because they are close to the adiabatic trapping limit described by the parameter (2). must be sufficiently small (e. g. < 0.187) to avoid stochastic instability (Chirikov, 1987) and migration across the loss cone boundary in phase space. Examination of Eq. (1) shows that stability is improved for protons transported to lower L, since varies as L², depending on energy only as E^0.5. The induction electric field acceleration preserves the first adiabatic invariant, which for nonrelativistic protons means that varies as L^0.5 as protons are transported inward. In contrast to the March 1991 storm which transported protons into L=2.5, weaker injection events trap solar protons where the ring current magnetic field perturbation peaks, around L=4 (Schulz, 1991). Solar protons appear to be only rarely transported into low values of L ~ 2.5, as for the March 1991 storm, with two previous candidates found in the literature (McIlwain et al., 1963; Gussenhoven et al., 1989). The latter could have been a redistribution of inner zone protons, again due to deep penetration of SSC induced electric fields.

Besides the adiabatic trapping criterion considered here, there exist other possible loss mechanisms associated with the main phase of geomagnetic storms which have not been ruled out. These include electromagnetic waves which violate any of the three adiabatic invariants, notably ion cyclotron waves which violate the first (Kennel and Petschek, 1966) and ULF waves excited by drift-bounce resonance with ring current ions, whose growth is enhanced by buildup of the stormtime ring current (Chen and Hasegawa, 1988, 1991; Kivelson and Southwood, 1985; Li et al., 1993b). A drift resonance (third invariant violation) with ULF waves in the 1-2 minute period range is possible for 5-10 MeV protons at L ~ 4, for example. ULF waves were observed within the plasmasphere in this frequency range following the March 24, 1991, SSC (Yumoto et al., 1992; Wygant et al., 1994; Hudson et al., 1996b]. An examination of CRRES electric and magnetic field data for the other SSC and storm events is underway.

ACKNOWLEDGMENTS

Work at Dartmouth was supported by AFOSR grant F49620-93-1-0101, and at Dartmouth and Berkeley by NASA grant NAG 5-1098 and NAGW-4728. Work at the Aerospace Corporation was supported by the Air Force under contract Fo4701-88-C-0089. Computations were performed on the SDSC and PSC Crays. The NOAA National Geophysical Data Center is acknowledged for providing GOES 7 data, and NSSDC for online access to IMP 8 data. MKH would like to thank Michael Schulz for helpful discussions and MPE, Garching for hospitality while this research was performed.

REFERENCES

Allen, J., H. Sauer, L. Frank, and P. Reiff, Effects of the March 1989 solar activity, Eos, Trans. Am. Geophys. Union, 70, 1479 (1989).

Araki, T., S. Fujitani, M. Emoto, K. Yumoto, K. Shiokawa, et al., The Anomalous Sudden Commencement on March 24, 1991, J. Geophys. Res., 102, 14,075 1997.

Blake, J. B., M. S. Gussenhoven, E. G. Mullen, and R. W. Fillius, Identification of an unexpected space radiation hazard, IEEE Trans. Nuc. Sci., 39, 1761 (1992a).

Blake, J. B., W. A. Kolasinski, R. W. Fillius, and E. G. Mullen, Injection of Electrons and Protons with Energies of Tens of MeV into L < 3 on March 24, 1991, Geophys. Res. Lett., 19, 821 (1992b).

Brautigam, D. H., Radiation Belt Electron Dynamics During a Magnetospheric Compression Event, Eos, Trans. Am. Geophys. Union, 76, F945 (1995).

Chen, L., and A. Hasegawa, On Magnetospheric Hydromagnetic Waves Excited by Energetic Ring Current Particles, J. Geophys. Res., 93, 8763 (1988).

Chen, L., and A. Hasegawa, Kinetic Theory of Geomagnetic Pulsations, 1, Internal Excitations by Energetic Particles, J. Geophys. Res., 96, 1503 (1991).

Chirikov, B. V., Particle Dynamics in Magnetic Traps, in Reviews of Plasma Physics, vol. 13, edited by B. B. Kadomtsev, pp. 1-91, Consultants Bureau, New York, NY (1987).

Fermi, E., On the Origin of the Cosmic Radiation, Phys. Rev., 75, 1169 (1949).

Ginet, G. P., W. J. Burke, and J. Albert, An Analysis of Electron Energization Seen in Simulations of the March 24, 1991, SSC, Eos, Trans. Am. Geophys. Union, 75, 305 (1994).

Gussenhoven, M. S., E. G. Mullen, and E. Holeman, Radiation Belt Dynamics During Solar Minimum, IEEE Trans. Nucl. Sci., 36, 2008 (1989).

Gussenhoven, M. S., E. G. Mullen, and M. D. Violet, Solar Particle Events as Seen on CRRES, Adv. Space Res., 14, (10)619 (1994).

Hudson, M. K., A. D. Kotelnikov, X. Li, I. Roth, M. Temerin, et al., Simulation of Proton Radiation Belt Formation During the March 24, 1991, SSC, Geophys. Res. Lett., 22, 291 (1995).

Hudson, M. K., S. R. Elkington, J. G. Lyon, V. A. Marchenko, I. Roth, et al., MHD/Particle Simulations of Radiation Belt Formation During a Storm Sudden Commencement, in AGU Geophys. Monogr., Workshop on Trapped Radiation Belts, pp. 57-62 (1996a).

Hudson, M. K., A. D. Kotelnikov, X. Li, J. G. Lyon, I. Roth, et al., Modelling Formation of New Radiation Belts and Response to ULF Oscillations Following March 24, 1991, SSC, in Workshop on the Earth's Trapped Particle Environment, AIP Conference Proceedings 383,} edited by G. D. Reeves, pp. 119-127 (1996b).

Hudson, M. K., S. R. Elkington, J. G. Lyon, V. A. Marchenko, I. Roth, et al., Simulation of Radiation Belt Formation During Storm Sudden Commencements, J. Geophys. Res., 102, 14,087 (1997).

Il'in, V. D., I. V. Il'in, and S. N. Kuznetsov, Accumulation of Oxygen Ions in a Geomagnetic Trap, Cosmic. Res., 31, 687 (1993).

Kennel, C. F., and H. E. Petschek, Limit on Stably Trapped Particle Fluxes, J. Geophys. Res., 71, 1 (1966).

Kivelson, M. G., and D. J. Southwood, Charged Particle Behavior in Low Frequency Geomagnetic Pulsations, 4, Compressional Waves, J. Geophys. Res., 90, 1486 (1985).

Li, X., I. Roth, I., M. Temerin, J. R. Wygant, M. K. Hudson, et al., Simulation of the Prompt Energization and Transport of Radiation Belt Particles During the March 24, 1991, SSC, Geophys. Res. Lett., 20, 2423 (1993a).

Li, X., M. K. Hudson, A. Chan, and I. Roth, Loss of Ring Current O⁺ Ions due to Interaction with Pc5 Waves, J. Geophys. Res., 98, 215 (1993b).

Looper, M. D., J. Blake, R. Mewaldt, J. Cummings, and D. Baker, Observations of the Remnants of the Ultrarelativistic Electrons Injected by the Strong SSC of 24 March 1991, Geophys. Res. Lett., 21, 2079 (1994).

McIlwain, C. E., The Radiation Belts, Natural and Artificial, Science, 142, 355 (1963).

McComas, D. J., R. C. Elphic, M. B. Moldwin, and M. F. Thomsen, Plasma Observations of Magnetopause Crossings at Geosynchronous Orbit, J. Geophys. Res., 99, 21249 (1994).

Morfill, G., and M. Scholer, Study of the Magnetosphere Using Energetic Solar Particles, Space Sci. Rev., 15, 267 (1973).

Schulz, M., The Magnetosphere, in Geomagnetism, vol. 4, edited by J. A. Jacobs, pp. 87-293, Academic Press, London, England (1991).

Schulz, M., Direct Influence of Ring Current on Auroral Oval Diameter, J. Geophys. Res., 102, 14,149 (1997).

Shea, M. A., and D. F. Smart, March 1991 Solar-Terrestrial Phenomena and Related Technological Consequences, in Proceedings of 23rd International Cosmic Ray Conference, Calgary, Canada (1993).

Tsyganenko, N. A., Global Quantitative Models of the Geomagnetic Field in the Cislunar Magnetosphere for Different Disturbance Levels, Planet. Space Sci., 35, 1347 (1987).

Tsyganenko, N. A., A Magnetospheric Magnetic Field Model with a Warped Tail Current Sheet, Planet. Space Sci., 37, 5 (1989).

Vampola, A. K., and A. Korth, Electron Drift Echoes in the Inner Magnetosphere, Geophys. Res. Lett., 19, 625 (1993).

Violet, M. D., K. Lynch, R. Redus, K. Riehl, E. Boughan, et al., Proton Telescope (PROTEL on the CRRES Spacecraft), IEEE Trans. Nucl. Sci., 40, 242 (1993).

Wygant, J. R., F. Mozer, M. Temerin, J. B. Blake, N. Maynard, et al., Large Amplitude Electric and Magnetic Field Signatures in the Inner Magnetosphere During Injection of 15 Mev Electron Drift Echoes, Geophys. Res. Lett., 21, 1739 (1994).

Yumoto, K., Y. Tanaka, T. Oguti, K. Shiokawa, Y. Yoshimura, et al., Globally Coordinated Magnetic Observations along 210^o Magnetic Meridian During STEP Period: 1. Preliminary Results of Low-Latitude Pc 3's, J. Geomag. Geoelectr., 44, 261 (1992).