Sources and Losses of Ring Current Ions: an Update

SOURCES AND LOSSES OF RING CURRENT IONS: AN UPDATE

J. U. Kozyra Space Physics Research Laboratory, Departmentof Atmospheric, Oceanic and Space Sciences, The University of Michigan, Ann Arbor, Ml 48109—2143, U.S.A.

INTRODUCTION Majorprogress has been made in thelast threeanda halfyears.inourunderstandingof the roleof ionospheric plasma as asource for theplasma sheet and ring current. Observations by the Dynamics Explorer 1 satelliteover thepolar ionosphere and concurrent modelling efforts havetraced thepath of heated ionospheric ions from a localized source associatedwith the polar cusp along trajectories that energize theions anddeposit them in thenear-earth plasma sheet AMPTE ion composition measurements in the ring current havebegunto clarify the relativecontributions of theionospheric and solarwind ions in the plasma sheet and ring current andcharacterize differences inthe energization processes whicheach population undergoes. The role of charge exchange as amajorloss process for the ring current has beenconfirmed by all sky observationsonboard the ISEE Isatellite of energeticneutral atoms resulting from this interaction and associatedmodeling efforts. The oxygen ion component of the ring current has beenfound to be asource ofheating via Coulomb collisions for the thermal plasmas in theouter plasmaspherewith associatedloss of energy from thering current. The roleof wave-particle interactions in ring current decay, however, is still unclear. In thepresent paperno attempt is made to comprehensively review thelargebody of literature describing ouraccumulated knowledge of ring current ion sources andlosses. A number of excellent reviews haveappeared in thelast few years discussing thedevelopment of the ring current and its dynamics /1,2,3,4/ The presentreview will concentrate on new developments in our understandingof sources andlosses of ring current ions that havetaken placeduring theperiod 1985 - 1988, overlappingslightly some ofthelatter reviewscited above with reference to earlierstudies as required. The populations comprising theearth’s ring current are determined by the strength andcomposition of ion sources, transportprocesses, and loss mechanisms. The intent of this review is to focus on processes which supplyions to the ring current and plasma sheet and the decayof the ring current. The transportof ions from theplasma sheet earthward and the subsequent formation of the ring current will not be discussed.

RING CURRENT SOURCES Ionospheric Sources of Plasma Sheet Ions

It is generally considered that thering current formation is atwo-stepprocess (cf. Williams /4,5/ andreferencestherein). A series of injection eventsmove plasma inward from the plasma sheet to theouterradiation belts (25 RE). In response to enhanced storm-time, dawn-dusk electric fields, this outer trappedpopulation is convected inward (—2.5 -5 RE) to form the ring current. Therefore processes whichpopulate either the plasma sheet or these near earthregions with plasma are considered, for the purposesof this review, sources ofthe ring current plasma. Plasma in these regions is likely amixture of solar wind and ionospheric populations.

Observations fromthe Dynamics Explorer 1 satellite of upflowing ionospheric ions (H~,He, O~’N~’and~ alongwith theoretical models investigating the transportof these ions into the polar magnetosphere, haveestablished thedayside cusp region as acopious and persistent source ofupflowing heated ionosphericplasma f7,8,9,10.ll,l2,13.l4,15,l6,17.l8,19/. The characteristics of this ionosphericsource axe listed belowso that these may later be compared with measurements of the ion composition of the plasma sheet and ring current and the changes in this composition with solarcycle and magneticactivity. These ion upflowshave beenlabelled ‘upwelling ion events’because all ion speciesparticipating in theevent display theeffects of equal ion heating to parallel andperpendicular temperatures of about 2 and 10ev, respectively i9/. Upwelling ion eventshave been statistically correlated with field aligned currents/4/. For oneparticular event studied in derail, the ionospheric heating was coincident with magnetic field signaturesof strong field-alignedcurrentson the boundaryof thedayside polar cap /13/. Also present on these field lines was astrong, localized convection channel orjet. Observations and theoretical models point Out the importance ofion heating (which aids the heavy ions in overcoming the gravitational barrier) in producing heavy ion outflows in the polar magnetosphere. Studies of transient polar wind outflow in thepresence of low altitude ion heating i20/ indicate that significant heavy ion outflows will result 120/. Lockwood et a!. /21/have pointed out that ion outflows from this region couldbe enhancedby theionosphericeffectsof flux transfer events (FTE) on thedayside magnetopause whichmap to the polar cap boundary. Theroughlyequal hearingof thevarious ion species implies that field aligned flow velocities will be inversely proportional to thesquare root of the ion mass. This mass dependent velocity, in combination with theExB convection andthe localized nature of the upwelling ion source, results in adispersionof these ions by mass andvelocityas they convectacross the polar cap /12/. This effect has beentermed the “geomagneticmassspectrometer” and results in the lowest-energyand highest

(12)171 (12)172 J. U. Kozyra

mass ions extending turthest toward tile mgtltside trom thecleft ionosphere. This upwelling ion source regionhas also been termedthedeft ion fountain/221 because as convection becomes antisunward andstrong, ions are dispersedfurther onto the nightside, analogous to a fountainblowing in the wind. 0~outflows increase both with increasingsolar radio flux at 10.7cm (FlO.7) and with increasingmagneticactivity; H~outflows exhibit no markedvariation with solarcycle anda muchsmaller variation with magnetic activity than that shown by the0~outflows 123/. Figure 1 is a schematic illustrationof the ion ouflows from the polar cleftandpolar cap regions taken from Horwitz 124/.

The question of how these ionosphericoutflows from the cleftion fountain reach theplasma sheet was addressed by Qadis125 / using akinetic, guiding-centertrajectory-based code with ion motion computed only near the noon-midnight meridional plane. Theions areaccelerated alongtrajectoriesdue to the centrifugalforce arising from theion convectionvelocity andtheenergy supplied by thecomponent of thecurvaturedrift of the ion along theconvection electricfield. Figure 2a displays ion guiding center trajectoriesprojected on the noon-midnight plane from Cladis 125/for fourdifferent conditions of themagnetic andelectric field. 0+ ions can reach thecenterplane of theplasma sheet even at times of moderate convection.Total energyof 0+ ions along these trajectoriesare shown in Figure 2b as afunctionof time. During magnetically disturbed conditions the O~’trajectories intersect the centerplane ofthemagnetosphere more earthward than during quiet times. The 0+ ions can be energized to typical plasma sheet energies alongthese trajectories with travel times betweenthe ionosphere and thecenterplane of 1-2 hours. Figure 3a from Cladis /25/illustrates thedependence of the ion motion on its initial locationin thepolar ionosphere andFigure 3b demonstrates this dependence on initial ion velocity. Ions gain higherenergiesandcross the centerplane atlarger radialdistances as their initial positions approachthecusp field lines. A similartrend results from increasing the ion parallelvelocities. Thenet result is an increase in heavy ion content of theplasma sheet with decreasing radial distance. The combination ofthesatellite observations detailed above andthetrajectory calculationsof Cladis /25/ indicates alargeproportionof the ionospheric ions observed in theplasma sheet during magnetically disturbed periods originate in thecleft ion fountain. In addition aadis /25/has shown that these ions reach appropriateenergiesprior to injection in theplasma sheet and require no additional energization. The origin of theH+ injected from this source region, however, is still ambiguous, since theH+ population in thedayside cusp region at theenergiesand altitudes of interest, is likely a mixtureof solarwind andionospheric ions. On appropriate trajectories Lockwood etal. fl/have suggested that heavy ion outflows will act as a source for thenighrside aurora! acceleration region. Figure 4 illustratesschematically how this might occur. N, MEASUREMENT OF POLAR CAP MASSGEOMAGNETIc DENSITYELECTRON SPECTROMETER R PROFILES H.+ PROPOSED NIGHT, ACCELERATING ~ \\~( ~, O+POLAR WIN7

4 HIGH—ALTITUDESUPERSONIC ~ “NCLEFT~I~.V/ION ~ /~‘~~ PARA8OLICN. O~.FLOWN OF LIGHT ION IN4. H.1 FOUNT’~IN CLEFT LOW—ENERGY POLAR WINO IN ~ ~ HEAVY IONS

Fig. I. Schematic illustration of ion outflows in thepolar ionosphere taken from 124/.

ioll

_~5 IS 0 -S -10 ;S -20 10 2.000 0,000 8.000 8,000 10.000 X.1R 0) TIME (sJ Fig. 2. (a) Guiding-center trajectories projected on theX-Zplane of ions flowing out of the polar cleft region during different levels of magnetic disturbance. In the simulation, the ions were started at 1.2 RE, e = 22°,p = 0°with V = 12 km/s and Cr = 30°.Disturbed conditions in curves 1-4, respectively, arerepresented by parameters (Ei, Kp) of (23.3 mV/rn, 2-), (40mV/rn, 2+),(56.7 mV/rn, 3-to3+),and (80 mV/rn, >3+). Arrival times in stepsof 1000 seconds are denoted by dots along each trajectory. This figure wastaken from i25i) (b) Energyachieved by the0~ions as they travel along trajectories 1 -4 shown in panel a 125/. Sources and Losses of Ring Current Ions (12)173

‘~15 10 5 0S~15 20 ~15 10 5 0~0~20

)( (RE) X (R~J

Fig. 3. (a) Guiding-center trajectories of ions flowing out of theionosphere from different locations acrossthepolar cap. All of the ions were started with initial velocity of 12 kmjs andinitial pitch angle of30°under the influence ofa polarcap potential ofEi =4OmV/mandKp=2+. (fmm/25/). (b) Guiding-center trajectories ofions flowingout ofthe polarcleft region with different initial velocities. Curves 1-4 correspond to ions with initial velocities of 10.95, 14.70, 17.66, and 20.20km/s. respectively and pitch anglesof 0°.The cross polar cap potential washeld constantat Ei = 56.7 mV~n,and at 3-to 3+. Arrivaltimes in steps of 1000 seconds are denoted by dots along each trajectory. (from/25i).

1.1 LOW Op. WEAK ANTISIJNWAEO CONVECTION (El HIGH kp•STRONG ANTISUNWAPO CONVECTION

//~~*

~/— ~j_~ ~~v—~ ~ rS

TRANSVERSE PARALLEL ~ REAMS HEATING HEATING ECONICS c\, \ /

\ NH.’

Fig. 4. Schematic diagram oflow-energy ion outflows from theCleft Ion Fountain during solstice conditions. Solid lines represent ion trajectories, dashed lines represent auroral field lines, S denotes theaurora!oval in the summer hemisphere and W that in thewinterhemisphere. For conditions of weak, anti-sunward, convectiondepicted in (a), upflowing ions from this region act as asource ofheavy ions (particularly 0+) for thepolar cap. During strong, anti-sunwani,convectionillustrated in (b), ion outflows from thecleft region may act as asource of heavy ions for the auroral acceleration region. (from /7/).

Ionospheric Source - Direct Injection into the Ring Current

Evidence forthe direct injection of ionosphericions into thering current hasbeenaccumulating for many years(cf., Horwitz /26f, Williams127/; Young ,28 /). In a surveyof two years(November1981 to December 1983)ofobservations by theAUREOL 3 spacecraft, the characteristics of downwardflowing ion (l)F1) beamshave beenrecentlyinvestigated129/. H~ions are always dominant in these flows. This is veryinterestingsince upward flowing ion (UFI) beams havecompositions whichrangefrom pureH~topure 0~130,31/ and aremuch more frequently observed 132/. Bosqued eta!. 129/ interpret these results to mean that UF! beamsare scattered in pitch angle near theequatorialplane. Kaye etaL 133/ had previously suggested this as amechanism for direct injection of key ions into the ring current. 0+ ions have longer bouncetimes than H+ ions of comparable energy. The effects of scattering andtransportwill be largerfor longer bouncetimes. Therefore, UFI beams containing 0+ will be scattered in pitch angle more than their H~counterparts andwifi be less likely to reach theconjugate hemisphere andbe observed as DFI beams.

Upflowing ion beams andconicsgenerally occur at L shells poleward of theplasmasphere. Direct injection of upflowirig (‘It 100 CV) ions, predominantly 0+, inside or at theplasmapause, have now been observed during anumber ofmagnetic storms in the midnightand earlymorning sectorsduring themain phase of magnetic stomss /34,35/and in thedayside during therecovery phase/36/. Ions at energies less than 1 KeV participatein corotation and thus cannotbe strictly considered as ring current ions but these ions area partof theenergetic ion environment in theouter plasmasphere. Forcompleteness, they areincluded in this discussion. (12)174 J U. Kozyra

Newell andMeng /37/report on observations by the DMSP satellite of up to I KeV ion precipitation within theplasmasphere as a common featureduring and after prolonged magnetic substorm activity. They inferthat injections occurfrom postmidnight to — 8:30 MLTandthe ions subsequently co-rotate and precipitatewith a lifetimeof approximately I day. Thereis indirectevidence that these ions arepredominantly 0~.Though Newelland Ming site magnetospheric injection processesas a potential source, these observations areremarkably similar to those of Bosqued /34/, Jorjio /35/, and Menietri eta!. /36/ which indicate an ionospheric source.

Plasmasi,herjc Source- Direct Injection

Theoreticalstudies /3839,40,41 andreferences thereiril and observations (Korth etal. /42/, Inhestor etal../43/. and Fraser etal. /44/) indicate that plasmaspheric He+ and0+ ions can gyroresonate with ion cyclotronwaves in thedusk bulge region, and increase theirenergy and pitchangle. Afterthis local energization, the ions essentially become part of the ring current populations. Figure 5/45/shows therelationship betweentheobservation of ion cyclotron wavesand the appearance ofenergetic He+ ions in the inner magnetosphere.

Estimated Strength of the Ionospheric Source

Prompted by the discovery ofa newand verysignificant source ofupflowing ionospheric plasma from the cleft ion fountain described above, in addition to otherpreviously identified sourcesof upflowing plasma in theaurora! zone /46 ,23/. Chappell et al. /47/ have attemptedto estimate theionospheric contribution to magnetosphericion populations throughout the terrestrial magnetosphere. Using recentdata on the ionospheric source strength, along with estimated volumes and ion residencetimes for each magnetospheric region (plasmasphere, plasma trough, plasma sheet andmagnetotail lobes), the authorsdemonstrated that the ionosphere waslikely a dominant source of magnetosphericplasma during quiet and active magnetic conditions in all of these regions. Ofinterest for thepresent review is their estimationof the importance of the ionspheric sourceforplasma sheet ions. Sources ofionospheric plasma for the plasma sheet were considered to be: (I) upflowing ions from the aurora! zone through direct entry into theplasma sheet boundary layer (2) the polar wind (H~and He~),(3) thecleft ion fountain and(4) thepolar cap ionosphere. Becauseof the flow trajectoriesof the ions from each of these source regions, the cleft ion fountain sourceis more significant in the first 10-20RE ofthe plasma sheet, whereas the auroralsource is more important tailward. Composition will also vary as aresult ofthese trajectories. 0+ fluxes from the cleft ion fountain intersect the plasma sheet near 7-15 RE: whereas, H+ fluxes from this same region will contributeto theplasma sheet at distances of 15-50RE. Figure 6 from Chappel et a!. /47/ is a schematic diagram of themagnetosphere showing the plasma sheet during quiet andactive magnetic conditions; and displaying sample ion trajectories from each of themajor ionospheric source regions.

Table 1 summarizes the flux contributions to the plasma sheet during solarmaximum and solarminimum from each ofthe ionospheric sources. The largestoutflows of H+ and He+ occur from the polar wind source during quiet magnetic conditions at solar minimum. In contrast, thelargest 0+ outflows occur during active conditions at solarmaximum. Calculated plasma sheet densities due solely to an ionospheric sourcewere 0.73-0.42 cm-3 for quiet conditions and 6.3 - 6.0 cm-3 for active magnetic

13 JULY 1977 60 I~II!I~lII!IllI!!!!Ir: ~.... WAVES EL (0.32-l.5H ) I

2.. 100 ~ .01

H.’I (25eV— 16h.V)

r~\A~A ~. PROTONS ~ * —~43—l6I~I. •

,- 44 0L._

ELECTRONS c. T7T1 I (2.00 (3.00 14.00 (5.00 (6.00 UT 6.0 6.7 7.3 7.6 6.9 Ld 1.4 6.7 11.9 16.3 9.1 A~ 13.35 14.08 4.39 15.08 15.44 LT

Fig. 5. Simultaneous wave and particlemeasurements on board GEOS I on July 13 1977 indicative of He~heating by waves. The panels, from top to bottom, are (1) wave power in theL-mode relative to l0-~y~/Hz,(2) relativeabundance of HeW in the total ion population below 110ev, (3) He+ energy density above 25ev whichincreasesdramaticallyafter wave onset(indicative of wave heating), (4) energetic proton anisotmpy (ratio of90°to 150°fluxes) in the lowestthree energychannels, and (5) local coldelectron density determinedby the wave propagation experiment onGEOS I. (after/451). Sources and Losses of Ring Current Ions (12)175

conditions. These results demonstratethat the ionosphere is alarge source of magnetospheric plasma and, in fact, maybethe dominant source in the near-earth plasmasheet.

a bORE

—.--—l1

3OR~

OUIET PLASMA SHEET CONDITIONS b -.. WOE 6 IN

~ ~

ACTIVE PLASMA SHEET CONDITIONS

Fig. 6. Schematic diagram ofthe magnetosphere showing assumed sizes of theplasma sheet for (a) quite and (b) active magnetic condit3ons and depicting theouter flow boundary for polar wind (dotted line) andcleft ion fountain ions (dashed line).

TABLE I Detailed Fluxes into Plasma Sheet

Solar Maximum Solar Minimum

H- He 0 H~ He 0

Quiet plasma sheer Polar wind 4.7X l0~’ 3.3 x lO~ l.4x io~’ l.SX 1024 Cleft ion fountain 3.3 X i0~ 1.6 x 1Q25 6.3 x 1024 7.3 x lO~ Auroral zone 2.1 X 10” l.6X 10” l.7x 10” lOx 10” Polar cap 2.4 X 1024 2.4 X 1024 ~.3 x l0’~ 3.9 x l0~’ 25 3.3X1024 3.4X10” l.7x1026 l.8Xl024 2.lxlO” Total 7.4xl0 Total tons 1.1 X 1026 1.9 x 1026

Active plasma sheer Polar wind 3.2x 10” l.2x 1014 9.4x 1023 8.6x 1023 Cleft ion fountain 4.3X 1024 4.8x 1025 4.3x l0~ l.9x 1025 Auroral zone 2.6X 10” 7.7X 10” 3.3x ID” 3.2x 10” Polarcap 6.1X1024 2.5Xl0” 1.OxlO” l.5xl0”

Total 6.7X 1025 l.2X 1024 l.5X 1026 l.4X 1026 8.6X 10” 6.6x 10”

Total tons 2.2X 1026 2.1 x 1026

fluxes given in ions S~ After /47/. (12)176 J. U. Kozyra

SolarWind VersusIonospheric Source - Rine Current and PlasmaSheet Composition Measurements The AMPTE spacecraft, launched in 1984, are capable of making in-situ measurements of ring current and plasmasheet composition in theenergy range0.3-315 keV/e (AMPTE/CCE) and 5-270key/c(AMFrE/IRM). Initial composition measurements by AMP~Eand their implications aresummarized in Gloeckler and Hamilton /48/. Key results ofrelevance to clarifying theion sourcesare: 1) 0~,N~,He~’O~and N0~and 02~areobserved in thestorm-time ring current at energiesup to — 300 keY/c on a timescaleof a few hours proving the ability ofmagnetospheric processes to energize ionospheric ions to ring current energies /49/. The time scales forthe appearance of energeticheavy ions in the ring currentfollowing storm commencementareconsistent with transport times for ionospheric ions from thecleft ion fountainto the centerplane of the plasma sheet calculated by Cladis /25/. 2) Questions regarding theorigin of ring current ions can be addressed by considering observations ofoxygen ions with various charge states. Highercharge states of oxygen and carbon (C 6~O6~). in general are found in thesolarwind andlower charge statesof oxygen (0k, O2~)in the ionosphere. Anunambiguous determination of thesourceis not possible due to charge exchange processes /50/which may increase or decreasethechargestate of a particular ion. In many cases, however, thedominant source can still be identified. Observations indicate that ionospheric origin ions (O~,O2~)falloffwith increasing L while solarwind origin ion (C6~’06~intensities•) increase with increasing L (see Figure 7) indicating twodifferent sources, one in the near-earth regionand the other in the tail /51/. This is consistentwith theentry ofsolarwind ions into themagnetosphere and theentry of ionospheric ions through the cleft ion fountain andaurora!acceleration regions. The entry of solarwind ions is thought to occurnear thedayside cusp and thelow latitude flank. These solarwind ions then convect intothe tail and move toward thetail plasma sheet. The observations presented in Kremser et a!. /51/indicate that the ionosphere is indeed themain source of 0’ ions in the magnetosphere. Important contributions from thesolarwind are encountered in the outer magnetosphere.

3) Gloeckler and Hamilton/48/estimatedthe fractions of ionosphere to solarwind ions in the plasma sheet and ring current during quiet anddisturbed times taking into account charge exchange losses from higherto lower chargestates in each region and approximating solar wind and ionospheric contributions to H+ by looking at the amount of solarwind versusionospheric origin heavy ions in theregion. (SeeTable II after Gloeckler andHamilton /48/). These results indicate that as much as 40-80% of the ring current and near earthplasma sheet ion population is ionospheric in origin.

Lennartssonand Shelley /52/, with a largedata base from ISEE1 ofplasma sheet ion composition between0.1 and 16 keV/e between 10 and 23 Re during 1978 and 1979, were able to contribute information on the variation of the heavy ion content of the plasma sheet with magneticactivity andsolar cycle. This study built on the results ofprevious statistical studies of smaller subsets of the ISEE1 data base/53,54,55/. The majorfindings ofrelevance to this review are

I. The 0~density in the plasma sheet increased by a factorof 3 betweenearly 1978 and early 1979, possibly in response to increasing solar EUV, indicating asignificantsolarcycle variationin the heavy ion content of theplasma sheet. 2. Solar wind origin ions andionospheric ions populate theplasma sheet underdifferent geomagneticconditions. ISEE observations clearly show that O~and theratio of 0~to H~ increases with increasingAEonthe averagebut the average energy ofthe 0+ ions does not change. On theother hand, ions of solarwind origin are most numerous and least energetic during periods of extreme quiet. H~andHe~during quiet times appear to approach energies that are comparable to kinetic energies in thesolarwind. This behavior is displayed in Figure 8. Interpretation of Composition Variations

Manyof thevariations in thecomposition ofthe ring cunentplasma statistically observed by high altitude spacecraft (ISEEI. AMPTE) can be explained by thevariations of theionospheric source with solar cycleand magnetic activity /25.47 and others/. Ionospheric ions,at low enough energiesto be influenced by gravity and magnetospheric convection, are dispersed by mass and energy as they convect acrossthepolar cap /22,15,11,25/. This dispersion will affect theobserved spatial distribution ofplasma sheet ion populations, even in the absence of a solar wind source /47/, producing a ratio of O+/H+ in the plasma sheet that increases with decreasing radial distance. Cladis /25/demonstrated that, even during times ofmoderate convection, upwardflowing 0+ ions (with velocities exceeding escape velocity) could be transported to theinner plasma sheet (Figure2a). Observations ofthe ionospheric outflowof 0+ ions by the Dynamics Explorersatellite /7,10,22.56/ indicate that the 0÷outflowincreases asmagnetic activity increases; and Young eta!. /57/ have demonstratedthat thenumber densities of He+, 0~,and O~in theplasma sheet arid ring current have a strong solar cycle variation increasingwith increasing solar radio flux (F10.7) upon which is superposed magnetic activity variations. These findings indicate that increased0~density in theplasma sheet during high magneticactivity (shown in Figure 8) is more likely due to an increase in the source of escaping 0~rather than to a modification of thetransport path i25/. Sources and Losses of Ring Current Ions (12)177

AMPTE1CCE-CHE1I LJMOIMPAE 1 SER 198t - 16 NOV. 1985 Kp~O1 Kp:i-2 Kp:2-3 Kp:3-4 Kp~4~5Kp~5~6

357935793579357935793579 I.

— ALL DATA —— — SELECTED DATA (LOW BACKGR.I

6~and, C6+ versusL shell for different levelsof magnetic activity as indicated by K~. SolidFig. 7.linesRelativerepresentfluxesallofdataO~,while~2+,dashedO lines represent data selected for low background counts. (from/51/).

Table 2 Estimates ofionospheric origin ion ratios andfractions of ionospheric origin ions in various magnetospheric regions

R,gion

Plasma shoot RJ 58 current

Ratio Condition 15 R5’ 8—9R,~’ 5—7 R1’ 3—5R1’

Quiet II ±I 9 ±I 6.5 ±1.2 II ±0.5 Disturbed 6 ±0.5 i.3 ±0.3 t.3 ±0.03 O’/Hct Quiet 24 ±22 22 ±I 16 ±4 27 ± Disturbed 20 ±I 70 ±10 00 ±20 Quiet 0.36 ±0.06 0.45 ±0.05 0.35 ±0.10 0.75 ±0.05 Disturbed 0.62 ±0.05 0.30 ±0.05 0.67 ±0.05 Quiet 0.37 ±0.07 0.47 ±0.06 0.38 ±0.11 0.76 ±0.06 4 Disturbed 1.760.65 ±0.06 0.461.21 ±0.07 1.00.75 ±0,06

‘uI’Q28—2266eV/c. ~ I.5—3l5keVfe. 20—3156eV/c. After /48/. ‘a (,, 1/e,,)~,,_;r~—

RING CURRENT LOSS PROCESSES

Charee-Exchange Decay

The dominant loss process for the ring current is chargeexchange of ring current ions with thehydrogen geocorona although pitch angle diffusion dueto waveparticleinteractions and Coulomb collisional losses areassumed to play a role (ci. Williams /1,3,4,6,27/). The role ofthese secondary processes hasnot been clearlydefined. In the last several years, interesting results have appeared verifying the dominant roleof charge exchange and clarifyingsomewhat the role of wave particleinteractions and coulomb collisions with the cold thermal ions ofthe plasmasphere. Recently significantprogress has beenmade towardobtaining a global view of ring current development and evolution /58/. Roelofet a!. have developed a method for identifying and analyzingenergeticneutralatoms (ENA) emanating from the ring current region as a result ofcharge exchange reactions of ring current ions with the neutral H geocorona. The modeling process provides additional conflmiation of the dominant role charge exchange plays in ring current decay. (12)178 J. U. Kozyra

The ISEE1 energeticneutralatom (ENA) emission function cannotbe inverted uniquely from asingle vantagepoint to obtain information on the ring current ion distribution. Analysis of theENA image proceeds by the following method: (I)a simple model of thering current ion population is constnzcted, (2) theENAflux due to charge exchange with the H~geocorona is obtainedusing this simple modelof the ring current ion distributions, (3) the inferred neutralparticle flux is convolved with the ISEEI telescope response function, (4) thesimulated instrument response is compared to observations made by theISEE1 spacecraft, (5) this process is iterateduntil reasonable agreement is achieved. The remarkable agreement in general morphology, obtained by the authors, between these twoimages, using this procedure with reasonable ring current ion distributions, provides strong support for thedominant role of chargeexchange in ring currentdecay. Pitch Angle Diffusion Itis generally agreedthat pitch angle diffusion as a resultof wave particleinteractions constitutes a minor but important ring current loss mechanism (see reviews /13,4,6/), but theplasma wavescausing theobserved pitch angle diffusion have not been clearlyidentified within theplasmasphere. Recent observational evidence from the VIKING spacecraft for theimportanceof pitch angle diffusion in ring current decay was offered by Studeman eta!. /59/. During the May 2-3, 1986 storm onset, while increased fluxes of tens of key protons were observedon theeveningside of theinner ring current region, the momingssde continued to exhibit thequiet-time inner ring current proton spectra for several hours. Drift periods due to the gradients in the magnetic field for 30 and 100 key protons on an L shell of 4 were calculated by the authors to be 7.1 and 1.6 hours, respectively compared to chargeexchange c-folding decay times of9.4 and 101 hours, respectively. They conclude that chargeexchange was partly responsible for theloss of low-energy ring current protons as they travelled along drift paths prior to reachingthe morningside but could not accountforlosses at high energies. Observations of isotropic proton pitch angle distributions were taken as an indication that strong pitch angle scatteringwas removing protons from the ring current before these ions were able to reach the momingside. A searchfor thewaves responsible for the loss of ring current protons via pitchangle diffusion continues. As early as 1970, Cornwall et a!. /60/ predicted that ion cyclotron waves(ICW’s) should be present in the plasmapause region and theresulting pitch angle diffusion would represent a loss process for the ring current protons. To date, ICW’s have not been observed with sufficient regularityor duration within thehigh-density, plasmasphere to support their importance in ring current decay in this region(see discussion in/61/; although some indirect evidenceof their presence exists. The proton precipitation, which presumably results from pitch angle diffusion of ring current ions resonating with these waves,has been observed. A region of moderately anisotropicproton precipitationis usually seen betweenL shells of2 and4 at all local times and during varying amounts ofmagnetic activity; and, well poleward of theplasmapause, a regionof intense isotropic pitch angle distributions exists /62,63,64,65/. This outer zone of strong pitch angle diffusion was also predicted to be dueto cyclotronresonance with protons at higherL values (L>5) wherethemagnetic field intensity hasdecreased sufficiently to again produce instability /66/.

ICW’s, which may be associatedwith this outer zoneof proton precipitation, have beenobserved. Statistical studies of the occurrence ofICW’s at geosynchronous orbit have beenperformedon GEOS 2data from a four month period (August- November1978)and on ATS-6 data overa four month interval(June - October 1974) with surprisingly similar results /67,68,69/. In these studies, the maximumoccurrence of left-hand polarized waves at geosynchronous orbit occurred near dusk (Figure 9a), in association with the comparatively higher plasma densities in thedusk bulge region of theplasmasphere. The highest densities in the dusk bulge region170/ were shown statistically to occur at thesame magnetic local rime (Figure 9b) as the maximum occurrence of ICW’s in the above studies. These ICW occurrenceswere associatedwith increases in thecold He~ abundance andanisotropic pitch angle distributions of ions having energies >20 KeV (cf. Korlh eta!. /67/). These results directly associate anisotropic ring current distributions with thepresence of ICW’s in the dusk bulge region. The relationship between magnetic activity and thecharacteristics of ICW’s observed by ATS-6 has been explored Till. ICW’s associated with substorms appearin the frequency range below fj.je-I- in theafternoon sector and are accompanied by a decrease in themagnetic field strength. This decrease is assumed to be due to theinjection ofenergetic, anisotropic ring current particles. ICW’s on the dayside occurin the frequency rangeabove fHe+ in association with a magnetic field compression which is assumedto enhance the anisotropyofenergetic protons as a resultof betatron acceleration Thus, it appears that ICW’s generatedin the duskbulge region provide a mechanism for ring current proton loss and maybe associated with the region of strong proton pitch angle diffusion statistically observed to occur on these L shells. Although observations indicate that ICW’s are present in the dusk bulge region, theconfirmation of thetheoretically-predicted importance of ICW’s in thehigher-density plasmaspherehas not beenforthcoming. Severalpossibilities (amongothers) existto explain the discrepancybetweentheory andobservations in this region. Cornwall eta!. /60/ pointed out that if theCoulomb collision frequency and ICWgrowth rate become comparable, wave growth will be3prohibited.and 10-2 54Theforauthorsplasmasphericestimatedensitiesthe of Coulombi02 cm3. collisionRevised ICWfrequencygrowthto beratesi0~in thes~presencefor plasmasphericof reasonabledensitiesamountsof l0~ofcrxrheavy ions ate 5 x 10~to 7 x 10-2 si /61/. Therefore, for densities higher than theseveral hundred per cctypical of thedusk bulge region, Coulomb collisions may effectively prevent ICW growth. The possibility also existsthat these wavesare present in the high-density plasmasphere but at such low frequenciesthat they are difficult to measure by spacecraft that have traversed this region. The theoretical results of Kozyra eta!. /61/ indicate that wave frequenciesmaybe shifted below 1 Hzin the presence of observed amounts ofheavy ions in the thermal and energeticplasmas in theouter plasmasphere.

Anotherdevelopmentin thearea of wave particle interactionshas to do with the conceptoflimiting integral flux developed by Kennel andPetschek/72/ which assumes an equilibrium existsbetween wave growth and particle fluxes. Increased ring current and radiation belt particle fluxes resultin increased wave growth and ultimately in increased ion precipitation due to pitch angle diffusion into theloss cone. This process maintains a limiting particle flux distribution. Schulz and Davidson[13/ have generalizedtheconcept oflimiting integral flux above a certain energythreshold developed by Kennel and Petschek /72/ to provide a limit of the differentialflux with energyabove this threshold. Davidson eta!. /74/ have so farapplied this formulation only to ring current electronspectra with interestingresults. Sources and Losses o Rine Current Ions (12)179

______to to .~...‘‘. ~ — I

SUM IONS S

I- H~~4 \...•_

- - ~H. — ~—‘.

~ H~ ~ LZ3~~~ ;~o

- 0’ ~

I ~ ~~

50 tOO lOGO MAX HOURLY AR FYI

After/5V. MAX HOURLY AE Yl

Figure 8. Results of a statistical study of ISEE-l ion mass spectrometer data in 1978 and 1979 in the energy range0.1 kev/e - 16 kev/e. Panel (a) gives average absolute ion densities in the plasmasheet andpanel (b)average plasmasheet ion energies as a function of AE. The dashed line in panel (b) is 4 times the He~energy.

0t ______U k.fl2’ 0

I LH SO 000 06 12 18 00 N., Local T~rtte

AH*~ 041IOtS2024MLT(HSS)

06.00 12.00 18.00 LOCAL TIME

Fig. 9. (a) Local time distnbution ofleft-hand polarized waves from four months (June 19- October 17, 1974)of ATS 6 observations (upperpanel, after /79/) and from four months (August-November 1978) ofGEOS 2 data (lower panel, after /8(V). Also in the lower panel are displayed theenergetic (45-59 key) proton anisotropy andthe occurrence of He+ flux events (He~fluxes exceeding acertain threshhold). In both casesthe occurrence frequency of left-hand, polarized waves peaks near 16 hours localtime associatedwith the dusk bulgeregion. (b) Median Ne values versus magnetic local time for L=5-6 (open squares), L=6-7 (filled squares andL=7-8 (filled circles), clearly showing the dusk bulge region which peaks near 16 hours magneticlocal time. (after 170/).

Coulomb Collisions

Coulomb collisionsbetweenring current protons and thermal electrons within theouterplasmaspherewere first proposedby Cole /75/ as theenergy source to powerstable auroral red (SAR) and conversely, act as a sink for ring current energy. The roleof Coulomb collisionsbetweenring current ions and thermal elections in theformation of SAR arcs was evaluated /76/ using observations of thermal andenergetic ion populations by DE-1 in theouter plasmasphere and nearly simultaneous measurements ofenhanced electrontemperatures by DE-2within theSAR arc at F region heights during four SARarcs in 1981. Figure 10 displays observations of energetic and thermal plasmas, simultaneously observed F-regionelectron temperatures, and calculated column-integratedelectronheating ratesfor one modelled SAR arc occurrence. The authors found that (1) sufficient energy was transferred in this manner to support theenhanced (SARarc) F region election temperatures measured on these field lines, (2) the latitudinal variation of the heatingrate was consistentwith theobserved variationof the election temperature acrossthe SAR arc, and (3) in all cases, ring current 0+ was themajor sourceof energyfor the SAR arcs. All observations were madeduring storm recovery phases. Thus, Coulomb collisions play a rolein thedecay of 10-100 key RC 0~during the recovery phase. (12)180 J. U. Kozyra

23 OCTO6ER 1981 SAR ARC 108

i0~ .__—~~ 127 to6 (I/CM2 ~ MEASURED /5/KEy) ENERGETIC ION FLUX io~ 5YEICSONDE—l

THERMAL ION CHARACTERISTICS MEASURED 10 8YRIIISONDE—l

I 0~ ~ ‘ H TEMPERATURE (K) IO~

ION DENSITY to2 / / (CM-3)

iot ~ I

~ - P4. - 6000

~ 4000 N. (crn~) 1. (1(1

‘I - 2000

0 ALT 927 920 911 900 889 KM

ELECTRONCOLUMN 0.0160.O12~I HEATING RATE (ERGS/ 0.004+ Ct12/S) ~ 3.95 3.28 2.77 2.40 2.11 L SHELL Fig. 10. Summary of energeticand thermal plasma characteristics andcalculated column electron healing rates for theSAR arc observed on October 23, 1981. The panelsin order from top to bottom display: (I) energetic ion flux at selectedenergies measured by DE-l in thenear-equatorial plasmasphere, (2) thermal ion density andH~temperaturemeasured by DE-1 in the near-equatorial plasmasphere, (3) electrondensity and temperature measured nearly simultaneously by DE-2 in the F region and (4) calculated column electron heating rate due to Coulomb collisions between theenergetic ions and thermal electrons. (after 176).

Coulomb Decay Time Coulomb Decay Time RIng Current H+ and Thermal Electrons Ring Current 0+ and Thermal Electrons

10’ Te~9330K Te=9330K 100cc 100cc ~ 10~ 500CC 500 cc E 940cc 101 ______~ 940 cc

>.IS ~_. 2820cc ~ 2820cc UC, :: ~ i04 Coulomb ______CE uI.um.

10~ i04~ 1 10 100 E(KEV) 1 10 100 E(KEV) Fig. 11. Coulomb decay times (solid lines) versus energetic ion energyfor an energetic H~and 0~test particlecolliding with thermal electrondistributions at 9330 K for selected thermal electrondensities calculated using the theoretical formulation of178/. Plotted on thesame axes are charge exchange lifetimes of H~and 0~ions atL = 3.5 (From /77/).

Finally, Coulomb collisions may play a major rolein the decay of the low energy portion of thering current within the plasmasphere. Figure 11 displays ring current charge-exchange lifetimes at L = 3.5 from Smith eta!. /77/ and Coulombdecay times /78/fora range ofreasonable plasmaspheric densities. Forprotons with energies below a few keV and 0+ at energies below 20 keV, Coulomb decay lifetimes arecomparable to orshorter than charge exchange lifetimes. Sources and Losses of Ring Current Ions (12)181

SUMMARY

The cleft ion fountainhas been identified as a prodigious source ofupfiowingsuprathenna! ionospheric plasma. Modeling efforts have traced thepath of these ions from the polar ionospherealong trajectories where theions are energizedto keV energies and deposited in thenear earthplasma sheet. Mass and energy dispersionof these ions accounts in a naturalway for theobserved variation in heavy ion contentof theplasma sheei Observations of ion composition in the plasma sheet by the AMPTE and ISEE spacecraft establish that ionospheric ions dominate inthe near earth plasma sheet but solarwind ions becomesignificant tailwaixl. The heavy ion content of theplasma sheet increases with both solarcycleand magnetic activity. Directinjection ofionospheric ions into thering current has beenobserved in theouter plasrnasptiere. Severalmechanisms forthe direct injection of ions from the plasmasphereand ionosphere into thering current haveappeared. Estimation of ionospheric source strengths and residence times have led to an estimate ofthe magnetosphericdensities that would resultsolely from an ionospheric outflow populating the magnetosphere. Estimated densities were quite reasonable even without inclusion of a solar wind source of ions. Ring current ions decay primarily via chargeexchangewith thehydrogen geocorona, however, the roles of pitchangle diffusion and Coulomb collisions in this decay process are being clarified.

Modeling and observations ofENA by the ISEEI spacecrafthas led to are-affirmation of thedominant roleofcharge exchange in ring current decay. Ion cyclotronwaves contribute to ring current decayin the dusk bulgeregion. The roleof low frequency. (<1 Hz)ion cyclotronwaves in theplasmasphereis still unclear. Other wave modes may be responsible for the pitch angle diffusion and subsequent loss of ring current ions. Coulomb collisional energy losses from ringcurrent 0~to thermal electrons are sufficient to powerSAR arcs and representan energy sink for ring current 0~within theplasmasphere. Coulomb collisions may beimportant for decay of low energy (< 10 Key) ring current ions in the plasmasphere.

REFERENCES

1. DJ. Williams, The earth’s ring current: causes,generation and decay, Space Science Reviews, 34,223 (1983) 2. DJ. Williams, Snace Science Reviews, 42, 375 (1985) 3. DJ. Williams, The earth’s ring current: Present situation and future thrusts, Phvsica Scriota, T18, 140 (1987a) 4. DJ. Williams, Ring current and radiation belts, Rev, of Geonhys., 25, 570 (l987b) 5. L. R. Lyons and D. J. Williams, Ouantitative Aspects of Maanetospheric Physics, D. Reidel , Dordrecht, 1984. 6. Spjetdvik, W. N. and P. L. Rothwell, The radiation belts, in: Handbook ofGeophysics and the Space Environment, ChapterS, Scientific Editor, A. S. Jursa, AirForce Geophysics Laboratory, USAF, 1985. 7. M. Lockwood, J. H. Waite, Jr., T. E. Moore, J. F. E. Johnson, and C. R. Chappell, J Geoohvs Res., 90,4099 (l985a) 8. M. Lockwood eta!., JGeoohvs. Res.. 90, 9736 (1985b) 9. M.Lockwood,T. E. Moore,J.H. Waite,Jr., C. R.Chappell,J. L. Horwitz, arrdR A. Hecls,~atur~,316, 612 (1985c) 10. J.H. Waite, Jr., eta!.,3. Geoohvs Res., 90, 1619 (1985) 61, AGU, Washington, D.C., 1986. 12.11. T.E.J.H. Waite,Moore,Jr.,C. R.etal.,Chappell,GeophysM.Monogr.Lockwood,Ser.,andVolI.38,H. Waite,edited Jr.,by T.S. Chang,Geophvs.p. Res., 90, 1611(1985) 13. T.E. Moore, M. Lockwood, M.0. Chandler, J. H. Waite, Jr., A. Persoon and M. Suguira, J. Geonhys. Res. 91, 7019 (1986) 14. JI. Horwitz, Features of ion trajectories in the polarmagnetosphere, Geophys. Res. LetI, 11, 1111(1984) 15. J.L. Horwitz andM. Lockwood, The cleft ion fountain: A two-dimensional kinetic model, 3. Geophys. Res., 90,9749 (1985) 16. J.L. Horwitz,I. H. Waite, Jr., and T. E. Moore, Geophys Res. Lett., 12, 757 (1985) 17. J.L Horwitz eta!, Geophys. Monogr. Ser., voL 38, edited by T. Chang, p.56, AGU, Washington, D.C.. 1986. 18. C. Gurgiolo and J. L. Burch, Composition of the polar wind - Notjust H~and He~,Geonhys. Res. Let:., 12,69 (1985) 19. J.L. Green and S. H. Waite, Jr., Geophys. Res. Let:.. 12, 149 (1985) 20. T.I. Gombosi andT. L. Killeen, 3. Geophys. Res., 92.4725 (1987) 21. M. Lockwood, M. F. Smith, C. 3. Farrugia, and G. L. Siscoe, S. Geophys. Res., 93, 5641 (1988) 22. M. Lockwood eta!.,J.Geoohvs. Res.. 90. 9736 (l985b) 23. A.W. ‘(au, E. G.Shdlley, W. K. Peterson, and L. Lenchyshyn, 3. Geoohvs. Res.. 90, 8417 (1985) 24. J.L. Horwitz, Core Plasma in the Magnetosphere, Rev, ofGeonhys., 25, 579 (1987) 25. J.S. Cladis, Geoohvs. Res. Let:., fl, 893-896 (1986) 26. J.L. Horwitz, The ionosphere as a source of magnetosphericions, Rev. Geonhvs. and Soace Phys., 20,929 (1982) 27. DJ. Williams, Ring currentcomposition and sources: An update, Planet. Space Sci., 29, Il, 1195 (1981) 28. D. T. Young, Near-equatorialmagnetospheric particles from —1ev to —1Mev, Rev, of Geoohvs and Space Phvs, 21,402 (1983) 29. J.M. Bosqued, I. A. Sauvaud, D. Delcourt, andR. A. Kovrazhkin, J. Geophvs. Res., 91, 7006 (1986) 30. ItO. Johnson, Energetic ion composition in the earth’s magnetosphere, Rev. Geoohys., 17, 696 (1979) 31. P. F. Mizera, era!., Theaurora inferred from S3-3 particles and fields, 1. Geophvs. Res., 86, 2329 (1981) 32. A.G. Ghielmetti, R. D. Sharp, E. 0. Shelley, and R. G. Johnson, J. Geoohys. Res., 84, 5781 (1979) 33. S.M. Kaye, E. 0. Shelley, R. D. Sharp, and It. 0. Johnson, Ion composition of zipper events, J. Ge ohys.Res~,86, 3383 (1981) 34. f.M. Bosqued, Ion precipitation into the ionosphere during geomagnetic storms, Adv. Soace Res., 5, 179 (1985) 35. N.Y. Jorjio eta!., Adv. Soace Res., 5, 141 (1985) (12)182 J. U. Kozyra

36. 5. D. Menietti, et al.. 3. Geoohvs. Res., 90, 1653 (1985) 37. P.T. Newell and C.-I. Meng, 3 Geophy. Res., 91, 11133 (1986) 38. L. Gomberoff and P. Vega, J. Geoohys. Res., 92, 7728 (1987) 39. L. Gomberoff and M. Molina. J. Geonhys. Res., 90, 1776 (1985) 40. R.Gendrin, Soace Sci. Reviews. 34, 271 (1983a) 41. R. Gendrin, in Hieh-Latitude Soace Plasma Physics, edited by B. Hultqvist and T. Hagfors, 415. Plenum, New York, 1983b. 42. A. Korth, 0. Kremser, S. Perraut, and A. Roux, Planet. Soace Sci., 32, 1393 (1984) 43. B. Inhester, U. Wedeken, A. Korth, S. Perraut,and M. Stokholm, 3. Geonhys. Res., 55, 134 (1984) 44. 8. J. Fraser, S. C. Samson, R. L. McPherron, and C. T. Russell, Adv. Soace Res., 6, 223 (1986) 45. D.T. Young, et al., J. Geonhys. Res., 86, 6755 (1981) 46. H.L. Collin, R. D. Sharp, and E. G. Shelley, J. Geonhys. Res., 89, 2185 (1984) 47. C. R. Chappell, T. E. Moore, andJ. H. Waite, Jr., 3. Geonhys. Res., 92, 5896 (1987) 48. 0. Gloeckler and D. C. Hamilton, AMPTE ion composition results. Physica Scrinta, T18, 73 (1987) 49. B. Klecker, E. Mobius, D. Hovestadt, M. Scholer, 0. Gloeckler, and F. M. Ipavich, Gcoohvs, Res. Let:, 13. 632 (1986) 50. W.N. Spjeldvik andT. A. Fritz, S. Geonhys. Res., 83, 1583 (1978) 51. 0. Kremser, W. Studemann, B. Wilken, 0. Gloeckler, D.C. Hamilton, F.M. Ipavich, J. Geonhvs Res., 92, 4459 (1987) 52. W. Lennartsson and E. G. Shelley, Survey of0.1 to 16 kev/e plasma sheet ion composition, J. Geoohvs. Res., 91,3061(1986) 53. W. K. Peterson, R. D. Sharp, E. 0. Shelley, and R. 0. Johnson, J. Geonhys. Res., 86, 761 (1981) 54. R. D. Sharp, W. Lennartsson, W. K. Peterson, and E. 0. Shelley, S. Geophys. Res.. 87, 10420 (1982) 55. W. Lennartsson and R. D. Sharp. Adv. Snace Res., 4,411(1985) 56. A. W. Yau, B. A. Whalen, W. K. Peterson andE. 0. Shelley, Distribution of upflowing ionospheric ions at the high-altitude polar cap and aurora! ionosphere, S. Geophys. Res., 89, 5507(1984) 57. D. 1. Young, H. Balsiger, and 3. Geiss, S. Geonhys. Res., 87, 9077 (1982) 58. E. C. Roelof. D. G. Mitchell, and D. 3. Williams, 3. Geoohvs. Res., 90, 10991 (1985) 59. W. Studemann et al., Geoohvs. Res. Let:., .1.4,455-458, 1986. 60. J. M. Comwall, F. V. Coroniti, and R. M. Thome, Turbulent loss ofring currentprotons, 3. Geonhys. Res., 75, 4699, 1970 61. J. U. Kozyra, T. E. Cravens, A. F. Nagy, E.G. Fontheim, R. S. B. Ong, J. Geophys. Res., 89, 2217 (1984) 62. D. S. Williams and L. R. Lyons, S. Geonhys. Res., 79, 4195 (1974a) 63. D. J. Williams and L. R. Lyons, S. Geonhys. Res., 79, 4791 (l974b) 64. 3. A. Lundblad, and F. Soraas, Planet. SoaceSd., 26,245 (1978) 65. F. Soraas, 3. A. Lundblad, and B. Hultqvist, Planet. Soace Sci., 25, 757 (1977) 66. 3, M. Comwall, F. V. Coronici, and R. M. Thome, 3. Geophys. Res., 76, 4428 (1971) 67. A. Korth, G. Kremser, S. Perraut, and A. Roux, Planet. Soace Sci., 32, 1393 (1984) 68. BJ. Fraser, J. C. Samson, R. L. McPherron, and C. T. Russell, Adv. Soace Res., 6, 223 (1986) 69. BJ. Fraser, Pc 1-2observation of heavy ion effects by synchronous satellite ATS-6, Planet. Soace Sci., 30, 1229 (1982) 70. P.M.E. Decreau, C. Beghin and M. Parrot, 3. Geonhys. Res,, 87, 695, (1982) 71. 5. Ishida, S. Kokubun, and R. L. McPherron, 3. Geonhys, Res., 92, 143 (1987) 72. C. F. Kennel and H. E. Petschek, Limit on stably trapped particle fluxes, LGeoohys. Res., 71. 1 (1966) 73. M. Schulz and 0. T. Davidson, Limiting energy spectrum ofa saturated radiation belt. J. Geonhvs. Res.. 93,59(1988) 74. 0. T. Davidson, P. C. Filbert, R. W. Nightingale, W. L. lmhof, 3. B. Reagan, and E. C. Whipple, S. Geonhys. Res., 93, 77 (1988) 75. K. D. Cole, Stable auroral red arcs, sinks for energy ofDst main phase, S. Geoohvs. Res., 70, 1689 (1965) 76. 5. U. Kozyra, E. 0. Shelley, R. H. Comfort, L. H. Brace, T. E. Cravens, and A. F. Nagy, S. Geoohys. Res., 92, 7487 (1987) 77. R. H. Smith, N. K. Bewtra, and R. A. Hoffman, S. Geonhys. Res., 86, 3470 (1981) 78. L. Spitzer, Jr., Physics ofFully Ionized Gases, Interscience, New York, 1962. 79. B.l. Fraser, and R. L. McPherron, S. Geonhys. Res., ~, 4560, 1982. 80. A. Roux, S. Perraut, 31. Rauch, C. deVilledary, 0. Kremser, A. Korth, D.T. Young, S. Geophvs. Res., &7., 8174, 1982.