Observations and models of current disruption in the Earth's magnetosphere are briefly reviewed. At the approach of current disruption onset, the cross‐tail current sheet shows a rapid growth in the current density, a large upsurge in the duskward ion bulk speed to nearly the ion thermal speed, an increase in the plasma pressure and its isotropy, a rise in the plasma beta, and a decrease in the current sheet thickness to a length scale comparable to the thermal ion gyroradius. During current disruption, there are (1) large changes in the local magnetic and electric fields, (2) significant magnetic and electric fluctuations over a broad frequency range, (3) magnetic field‐aligned counterstreaming electron beams, (4) ion energization perpendicular to the magnetic field, and (5) reduction in the cross‐tail current by an amount similar to that built up during the growth phase. Observations further indicate that regions of local reversal of the north‐south magnetic field component are not necessarily sites of intense particle energization. Remote sensing of disruption activities shows that at least some current disruptions are not caused by a disturbance propagating earthward from the tail beyond 10 RE downstream. The timescale involved is comparable to or shorter than the ion gyroperiod. Current disruption thus has spatial and temporal scales outside the MHD regime. Several models for current disruption are briefly discussed. Two roles are considered for the cross‐field current instability proposed for current disruption. It can provide anomalous resistivity for magnetic reconnection as advocated by the traditional viewpoint or act singly to instigate global changes of the magnetosphere during the initial substorrn expansion phase. The latter role is elaborated by showing that the instability may modify significantly the local current density and any such process will alter the force equilibrium in the current sheet and give rise to an efficient plasma and energy transport on a global scale. Furthermore, such a process can generate field‐aligned current with intensity comparable to those associated with an auroral breakup arc at substorrn expansion onset. This scenario leads to a new emphasis that in addition to magnetic reconnection, rapid conversion of magnetic energy into particle energy in magnetotail systems may take place without a magnetic X line or separatrix playing the key role in energy conversion.

For many years, researchers have utilized definitions of the substorm phenomenon that are not consistent among one another, and this has created great difficulties in comparing the results reported in the literature by the various researchers. In August 1978, nine magnetospheric physicists active in the field of substorm research met in Victoria, British Columbia, Canada, to attempt to reach a consensus on an acceptable definition for a magnetospheric substorm. This paper reports the agreements reached at the Victoria workshop and presents an operational definition of the magnetospheric substorm and a critique of the various signatures by which researchers can identify the time sequence and spatial extent of the substorm.

Three decades of research in magnetospheric substorms has not led to a general consensus view of the substorm process. Several substorm models, mostly phenomenological, are presently under consideration. These competing models, each being justifiable on the basis of certain features of a substorm, have major differences as well as similarities alllOng them. A synthesis substorm model is desirable, as first suggested by Siscoe (1986). In this paper we construct a coherent description of substorm development by extracting some important components from these existing models. The scenario of the synthesis model includes the ionospheric influence on substorm expansion onset, current disruptions leading to convection surges and tailward propagating rarefaction waves, wave-induced precipitation, local time expansion of the disturbance region via velocity-shear-related instabilities, plasma sheet heating by resonant absorption of hydromagnetic waves, and the formation of magnetic reconnection domnin$. This synthesis represents one possible way to integrate the different existing models coherently. INTRODUC•ONThere is no doubt that a major task in magnetospheric research is to understand magnetospheric substorms since most dynamical changes in the magnetosphere occur during substorm intervals. Starting from the early days of substorm research, observed substorm phenomena have been cast in the framework of the near-Earth neutral line model [e.g., Hones, 1979]. As more satellites are flown, equipped with increasingly sophisticated instruments at faster sampling rates, the opinions on the correct description of the substorm process have become more diverging than converging.Over the last two decades, several alternative substorm models have appeared in the literature. These include wave-induced model gives only a phenomenological description, it serves the important role of providing a sequential evolution which can then be tackled quantitatively in a piecewise fashion. We emphasize that this is only an initial step toward integrating the principal substorm features from different models. Modifications are anticipated when additional features recognized to be essential parts of a substorm are assimilated to give a more complete picture of this dynamic episode. PRINCIPAL SUBSTORM Pttl•OMENA AND SUBSTORM MODELS Substorm phenomena have been identified in various regions of precipitation model [Parks et al., 1972; Kropotkin, 1972], the the near-Earth space during each of these substorm phases. Table 1 current disruption model [Chao et al., 1977], the configurational lists some of these key phenomena often discussed in connection (or ballooning) in_stability model [Roux, 1985; Roux et al., 1990], with substorm activities in the regions of the solar wind, the ionothe boundary layer model [Rostoker and Eastman, 1987], the mag-sphere, the near-Earth tail, the midtail, and the far tail [Huang, nerosphere-ionosphere coupling model [Kan et al., 1988, 1990; 1987]. The lack ofdramatic substorm features in the midtail during Rothwell eta...

Observations from the Charge Composition Explorerin 1985 and 1986 revealed fifteen current disruption events in which the magnetic field fluctuations were large and their onsets coincided well with ground onsets of substorm expansion or intensification. These events are of short durations locally (∼1–5 min). They are mostly confined to within ∼0.5 RE of the neutral sheet and 1 hour local time from the magnetic midnight. Over the disruption interval, the local magnetic field can change by as much as a factor of ∼7. In general, the stronger the current buildup and the closer to the neutral sheet, the larger the resultant field change. There is also a tendency for a larger subsequent enhancement in the AE index with a stronger current buildup prior to current disruption. For events with good pitch angle coverage and extended observation in the neutral sheet region we find that the particle pressure increases toward the disruption onset and decreases afterward. Just prior to disruption, either the total particle pressure is isotropic, or the perpendicular component (P⊥) dominates the parallel comment (P∥), the plasma beta is seen to be as high as ∼70, and the observed plasma pressure gradient at the neutral sheet is large along the tail axis. The deduced local current density associated with pressure gradient is ∼27–80 nA/m² and is ∼85–105 mA/m when integrated over the sheet thickness. We infer from these results that just prior to the onset of current disruption, (1) an extremely thin current sheet requiring P∥ > P⊥ for stress balance does not develop at these distances, (2) the thermal ion orbits are in the chaotic or Speiser regime while the thermal electrons are in the adiabatic regime and, in one case, exhibit peaked fluxes perpendicular to the magnetic field, thus implying no electron orbit chaotization to possibly initiate ion tearing instability, and (3) the neutral sheet is in the unstable regime specified by the cross‐field current instability. Subsequent to the disruption onset, enhancement of magnetic noise over a broad frequency range, magnetic field aligned counterstreaming electron beams, ion energization perpendicular to the magnetic field, and current reduction in the amount similar to that of current buildup during the growth phase are observed. These features seem to be compatible with the predicted development of the cross‐field current instability.

We investigate a cross‐field current instability (CFCI) as a candidate for current disruption during substorm expansions. The numerical solution of the linear dispersion equation indicates that (1) the proposed instability can occur at the inner edge or the midsection of the neutral sheet just prior to the substorm expansion onset although the former environment is found more favorable at the same drift speed scaled to the ion thermal speed, (2) the computed growth time is comparable to the substorm onset time, and (3) the excited waves have a mixed polarization with frequencies near the ion gyrofrequency at the inner edge and near the lower hybrid frequency in the midtail region. On the basis of this analysis we propose a substorm development scenario in which plasma sheet thinning during the substorm growth phase leads to an enhancement in the relative drift between ions and electrons. This results in the neutral sheet being susceptible to the CFCI and initiates the diversion of the cross‐tail current through the ionosphere. Whether or not a substorm current wedge is ultimately formed is regulated by the ionospheric condition. A large number of substorm features can be readily understood with the proposed scheme. These include (1) precursory activities (pseudobreakups) prior to substorm onset, (2) substorm initiation region to be spatially localized, (3) three different solar wind conditions for substorm occurrence, (4) skew towards evening local times for substorm onset locations, (5) different acceleration characteristics between ions and electrons, (6) tailward spreading of current disruption region after substorm onset, and (7) local time expansion of substorm current wedge with possible discrete westward jump for the evening expansion.

The progressive developments in the radial profiles of the particle pressure, plasma beta, and electric currents of the storm time ring current are investigated with data from the medium energy particle analyzer on the AMPTE Charged Particle Explorer spacecraft. Measurements of ions from 25 keV to 1 MeV, which carry 70–85% of the energy density of the entire ring current population, are used in this work. Two geomagnetic storms in September of 1984 are selected and four traversals of the equatorial ring current region during the course of each storm are studied. It is shown that enhancements in the particle pressure occur initially in the outer region and reach the inner region in the late phase of the storm. Structures suggestive of multiple particle injections are seen in the pressure profile. The leading and trailing edges of the particle injection structures are associated, respectively, with the depressions and enhancements of the westward current densities of the ring current. Plasma beta occasionally increases to values of the order of 1 in some regions of the ring current from prestorm values of the order of 0.1 or less. It is also found that the location of the maximum ring current particle pressure can be several earth radii from where the most intense westward ring current flows. This is a consequence of the dominance of pressure gradient current over the current associated with the magnetic field line curvature and particle anisotropy.

Magnetospheric plasma parameters during geomagnetically quiet conditions are studied with ion measurements covering energies from ~1 keV to ~4 MeV from the Charge Composition Explorer. It is found that the observed quiet time plasma pressure in the midnight sector is comparable (within a factor of 2) to that deduced by Spence et al. (1989) and Kan et al. (1992) from inverting the magnetic field models of Tsyganenko and Usmanov (1982) and Tsyganenko (1987). The radial profile of the total plasma pressure shows a peak generally at L = 3 to 4 and decreases from L = 4 to L = 9 rather monotonically. No largescale earthward decrease in plasma pressure occurs in these outer L shells, contrary to some theoretical expectation for the magnetospheric closure of the region 2 field-aligned current system. The plasma pressure within L ~ 9 is generally anisotropic with a larger pressure component perpendicular to the magnetic field than parallel to the field. This anisotropy tends to increase with smaller L. While the anisotropy is always below the fire hose instability threshold during all quiet time passes studied, it is near the mirror instability threshold in the midnight sector at large L shells due mostly to the large value of plasma beta (>1) there. The radial profile of the azimuthal volume current density is also computed, with the result suggesting that the current density associated with the quiet time ring current population is distributed broadly over the L shells between L = 3.5 and L = 7 with average values in the range of ~1 to 4 nA/m 2 and peak values in the range of ~4 to 8 nA/m 2. Another current region situated beyond L = 7, which may be related to the inner portion of the quiet time cross-tail current, is also apparent and has average and peak current density values similar to those of the ring current.

The AMPTE CCE spacecraft observed a transverse Pc 5 magnetic pulsation (period --• 200 s) at 2155-2310 UT on November 20 (day 324), 1985, at a radial distance of 5.7-7.0 Re, at a magnetic latitude of 1.2ø-1.9 ø , and near 1300 magnetic local time. The magnetic field perturbation was observed primarily in the radial component with an amplitude of 15 nT peak to peak. Ion fluxes (energy > 50 keV) measured by the medium energy particle analyzer (MEPA) on board CCE were also observed to oscillate at the frequency of the magnetic pulsation. The wide range of energy and pitch angle of ions covered by the MEPA allowed us to study the ion flux oscillations in great detail. It is found that (1) regardless of energy the oscillation amplitude tends to maximize near the field-aligned directions while it is essentially zero at 90 ø pitch angle, (2) for a given energy and the given location (east or west) of ion guiding centers, flux oscillations at pitch angle a and at its conjugate, 180 ø-a, are 180 ø out of phase,for a given look direction, the oscillation phase changes with energy, and (4) for a given pitch angle and energy, the eastside flux oscillation leads the westside flux oscillation. These observations can be explained by the adiabatic theory of ion flux pulsations with finite Larmor radius effects included (Southwood and Kivelson, 1981; Kivelson and Southwood, 1983), if we assume an antisymmetric standing wave on the field line, westward propagation of the wave, and a large azimuthal wave number Iml • 110. These properties of the wave are consistent with a second-harmonic standing Alfv6n wave excited in the region where the ring current ions have an inward density gradient. ]. Hughes et al. [ 1979] studied ion flux oscillations observed with the University of California, San Diego (UCSD), plasma analyzer on board ATS 6 in association with a meridionally polarized Pc 4 magnetic pulsation.They found that the low-energy ions (energy <2 keV) oscillated in quadrature with the radial magnetic field oscillation, whereas the high-energy ions (energy -7-30 keV) oscillated in antiphase with the magnetic oscillation. They explained these observations with electric field acceleration for the low-energy ions and with pressure balance between magnetic field and particles for the high-energy ions.