During substorms the (1-100) keV ions are accelerated in the tail plasma sheet and injected by short pulses into the auroral flux tubes carrying the information on the source distance, scale-size and temporal history of plasma acceleration.

We present observations by CLUSTER crossing the auroral zone flux tubes at ~4 Re distance during the substorm on February 14, 2001, when signatures of near-Earth reconnection was observed by Geotail at 23 Re in the plasma sheet.

We found :

* Multiple energy-dispersed ion beams are confirmed to be the transient effects, with other effects (spatial velocity filtering) contributing to the observed dispersion. The beam multiplicity with the ~1-3 min quasi-periodicity is not yet explained by any substorm theory.

* The ion beams cover the same region (poleward half) of the auroral oval where the low energy ions are extracted from the ionosphere, and where Alfven waves are observed carrying net downward parallel Poynting flux into the ionosphere. The beam generation region is not limited to the newly reconnected flux tubes, but evidently includes a portion of the closed flux tubes.

* Surprising variety of apparent (TOF) injection distances (between ~10 Re and >60 Re) have been inferred using the full energy and temporal resolution. Substructure (injection-lets) of the main ion beams (CIS) and apparent differences of dispersion slopes in the energy ranges covered by CIS and RAPID instruments indicate the complicated spatio-temporal history of particle acceleration in Earthward-convecting flux tubes.

* We interprete the (most frequent) distance ~20-30 Re as the distance to the active magnetic reconnection region (in agreement with Geotail observations). The largest apparent distance (60-90 Re) could be an artefact of spatial dispersion (convection filter) or of the real injection with the flight time " 3/4 of bounce period.

* Two observations at the transition between the dipolelike and current sheet regions at ~10Re , both include substantial downward energy flow to the ionosphere, both are believed to be associated with breaking of Bursty Bulk Flows :
** Sharp quasi-stationary outer boundary of energetic electrons (event EEW, width less than ion gyroradius, lifetime ( 5 min), with associated narrow westward plasma flow channel at its outer edge; and
** Strong transient and localised ion injection (observed only on the innermost SC#1) with flight distance of only 5Re (event T).

Event:

* Continuous ~300-400 nT AE activity following disturbed period on previous day;
* Substorm onset at ~0026 UT (injections to 6.6 Re, auroral zone magnetic bays); activations at 0047, 0052, 0056?, 0108 UT (see magnetograms in Figure 1 );
* Auroral activity (IMAGE FUV WIC camera, Figure 2): double oval with many local activations in the poleward oval;
* Midtail plasma sheet (GEOTAIL at 23 Re premidnight, Figure 1): Between 0030 and 0120 UT - thin current/plasma sheet and signatures of active reconnection at r <~20 Re (southward BZ and sporadic tailward flows)

CLUSTER orbit, coordinates and mapping ( Figure 3)

* Inward (equatorward) crossing of auroral zone at midnight at r~4Re, 0040-0110UT;
* Roughly meridional spacecraft separation (80-->120 km in the ionosphere between SC#1 and #4);
* Expected orbital time delay about 3-5 min between spacecraft #1 and #4;
* CLUSTER magnetic field fits best the Kp=4 T89 model, used for mapping.

Survey of CLUSTER observations : ( Figure 4, survey for SC#1):

* Poleward boundary (0043 UT,PCB): flux increase of energetic particles (up to >100 keV), then onset of variations of convection and magnetic field (0045UT);
* Basically transverse magnetic variations (dB1 much less than dB3 which is less than dB2) due to the Alfven waves, complicated E&B variations (interference of incident and reflected AW) with short bursts of parallel Poynting flux (S(() predominantly directed toward the ionosphere;
* Eastward convection ( V = [ExB], +V2 to the east) dominates in the poleward oval;
* Transient event T (0055UT): following the energetic particle flux increase --> a large variation of E&B resulting in the largest downward Poynting flux pulse;
* Energetic Electron Wall (EEW) at 0105UT, sharp (<10s, (Je /Je (10-100) energetic electron flux increase (double), accompanied by the ion pressure increase and impulsive convection (pulsation at SC#1, only one period at the other spacecraft)

Energetic Electron Wall (EEW) - a sharp spatial boundary at the outer edge of the Radiation Belt:

* Observed at all spacecraft (with time delays roughly consistent with spacecraft separation) --> sharp spatial boundary, lifetime >5 minutes;
* Thickness (EE flux increase) less than 2 spins (8 sec, w~3.6km mapped to the ionosphere), rce << w < rcp ;
* Imbedded into the broad downward Field-Aligned Current (rising dB2);
* Westward convection pulse and downward Poynting flux pulse just preceeding the energetic electron flux increase --> narrow westward flow channel at the outer boundary of the EEW;
* Associated with (more broad) ion pressure increase (3-40 keV protons);
* Mapped to ~10 Re equator (at the transition between the current sheet and dipolelike magnetic field) and to the slit between poleward and equatorward ovals.

CIS view of ion population (Figs.5, 7).

Low resolution (CODIF, Figure 5):

* Two overlapping ion beams B1,B2 at auroral energies ( ~0045-0100 UT, mostly protons), observed nearly simultaneously by all spacecraft (transient effect), although with systematic energy difference being consistent with spatial convection filtering;
* Systematic change of Low Energy Cutoff versus Latitude at auroral energies, basically spatial effect (convection filtering);
* Low energy (ionospheric) ions extracted from the ionosphere (beam along B) in the poleward oval where the beams B1,B2 are observed (~0045-~0100 UT);
* Spatial boundaries crossed by CLUSTER include the poleward oval boundary (~0045-0047 UT), equatorward boundary of low energy ionospheric outflow (0056-0101 UT) and ion pressure boundary (0105-0110UT). The latter boundary corresponds to Energetic Electron Wall (EEW, Fig.4)

High energy+time resolution and Time-of-Flight estimates (HIA, Figure 7):

* Low energy cutoff (LEC) and dispersion slope of the main beams B1 and B2 correspond to the large apparent flight distance of ~60..90 Re ;
* Main beams B1 and B2 (especially at SC#1) are split into the beam-lets which display the steeper traces (with flight distances roughly ~20 Re),
* Beam-lets start at higher energy than expected from main beam center energy (produced by localized impulsive acceleration of beam particles?);
* Steep dispersion (transient and localized, seen only at one SC#1) at 0055UT with flight distance being only 5 Re (!) for event T.

RAPID view of energetic ions and electrons ( Fig.6),
comparison to CIS results ( Fig.7)

* Energetic ion beam B1 (up to 100 keV) including He and O ions is seen near the poleward boundary at all spacecraft with species-dependent time delay , however energetic heavy ionospheric ions appear earlier than protons at the same velocity as predicted from average LEC or B1 beam in CIS data ( Figure 7);
* Observations show less similarity between different spacecraft (SC#1 and #3) in the [poleward oval than those in the auroral energy range (Figs.5), suggesting more spatial structuring and temporal variability for energetic particle fluxes;
* There is no simple relationship between the features in auroral (<32keV, CIS) and energetic (>30 keV, RAPID) energy ranges. Traces of main beams B1-B2 in CIS data have no continuation at high energies, they may have associated energetic acceleration events at proper time but with different dispersion slope (suggesting common origin and possibility of multiple bounces);
* Most of well-tracked dispersion features (especially at SC#3) give the flight distances about 20-30 Re, in reasonable agreement with near-Earth reconnection location;
* Good agreement with CIS for transient acceleration event T, produced energetic ions and electrons at the flight distance ~5Re (roughly, at r~10 Re in the tail);

Interpretation of ion beam structure and TOF results Fig.8)

Evidence of near-Earth reconnection (Geotail);
Multiple beams + beam substructures with steeper dispersion slopes;
Conflicting TOF results from CIS and RAPID instruments, including two flight-time scales, ~20Re and 60-90Re.
Suggesting the equatorial location of ion acceleration region and impulsive appearance of magnetic reconnection we discuss two opportunities:

Version 1: Two active reconnection regions; two flight-time scales naturally appear,
Problems:
* conflict between RAPID and CIS dispersions and early appearance of O+beam for the beam B1 is unresolved;

Version 2: One active near-Earth reconnection region; longer flight-time scale could be formed due to multiple bounce ((t=3/4(B) of particles launched by near-Earth reconnection.
Problems:
* don't see low energy part of the direct proton beam (t=1/4(B on the expected (could be a confinement of low-energy ions due to E(( or cross-B localization effects?);
* don't see high energy (RAPID) part of reflected beam at (t=3/4(B (that part may be scattered in pitch-angles and therefore destroyed during CS crossing if Bn is less than 10-20nT);

Coexistence of Beam1 and Beam2 particles in the same flux tube could be explained by impulsive acceleration in the closed inward-convecting plasma sheet flux tubes (inward propagation of fast wave from impulsive reconnection source?) or by convection-related spatial velocity filtering (version 1).

--------------------------------------END