Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.izmiran.rssi.ru/magnetism/Feldstein_2012/PDF/396.pdf Äàòà èçìåíåíèÿ: Wed Mar 11 16:44:16 2009 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:21:18 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: http astrokuban.info astrokuban

GEOPHYSICAL RESEARCH LETTERS, VOL. 27, NO. 17,PAGES 2813-2816, SEPTEMBER 2000 1,

On the two-phase decay of the Dst-variation
Yo I. Feldstein and L. A. Dremukhina Institute of TerrestrialMagnetism, Ionosphere Radio Wave Propagation, and Troitsk, Moscow
Region, Russia

U. Mall

and J. Woch

Max-Planck Inst, itut ruer Aeronomie, Katlenburg-Lindau,Germany

parameterof to+ = 17.3 hours[Hamiltonet al., 1988]. Daglis [1997]has supportedthis explanation,noticing shows two-stage a decaypattern, decaying first quickly that, the O+ fraction of the ring current energydrops and then decayingwith a larger decayparameter. This simultaneously with Dst. Unfortunately, one has to finding discussed the context the magnetic is in of storm admit that no experimental verification for the explawhich was observedon November 25-27, 1986. Con- nation of Akasofu et al. has been reported and that trary to frequentlyusedinterpretationsas two spatially Hamilton et al.'s two-ion speciesexplanation faces some
that during the storm recoveryphasethe Dst-variation
separated ion populations or two different, atomic ion components. we propose an alternative explanation for
difficulties as well for moderate storms.

Abstract.

We discuss the well known observation

Takah, iet al. [1990] have made a simulationof ash,

the storm-time ring current,. They examined the behavior of charged particles in a dipole magnetic field upon which they imposed a time-dependent electric field. Tl·ey argued that, the Dst recoveryis a combination of a "flow out" effect,(or decrease injection) and of a lossprocess trapped particles by chargeexchange. of The con·bination of these two processescan generate Introduction various patterns of ring current decay. Protons with It is generallyobserved that during the storm recov- energies of 20 keV and more which are once trapped ery phasethe Dst- variationshows two-stage a decay can escapefrom the model magnetospherethrough the pattern. decayingfirst quicklyand then decaying with dayside magnetopause. This can also be a causefor the a larger decayparameter. To date, two suggestions try fast,initial ring current decay. Undoubtedly, to examine to explain this peculiarfeature. The first,proposed by the D st-variation it, is necessaryto take into account the Akasofuet al. [1963]assumes that the two-phase re- lossprocesses trapped particles different from charge of coveryin the Dst-variationis causedby two spatially exchange processes. Such processesare, in particular,

this feature. We arguethat duringthe recovery phaseof the magneticstorm the D st decayis controlledby the decayof a two currentsystem'the ring current(DR) and the magnetospheric current (DT). tail

separated ring current populations.The outer compo- precipitationof ionsinto the upperatmosphere [Kozyra nent was assumed to exist in storms of all sizes and to et al.. 1998]. Losses ionsdueto drift,throughthe dayof produce the slowpart of the recovery largestorms. side magnetopauseas well as due to precipitation into in The inner populationwas suggested exist only in the atmosphereexist generallyin the day-afternoonside to very large storms. Its decay was identified xvith the of the magnetosphere duringthe main phaseof the magrapid initial recovery. The difference decaytimesfor netic storm. They causean abrupt asymmetry (AS}") in the two components the ring currentwasattributed of the longitudinal distribution of the magnetic disturof to the radial dependence the neutralhydrogen of den- bance A H on the Earth's surface and a fast decay of sity causing difference chargeexchange a in lifetimes. the .4S}' near the transition from the main phase to The secondexplanationproposed Hamilton et al. the recovery by phaseof the magneticstorm [Feldsteinet [1988] not add another did ringcurrent the picture al., 1994]. to

but suggested the two-phase that recovery a large in
storm resultsfrom a ring current,made up of two dif-

ferentatomiccomponents + and O+) whichdecay The storm of November 25-27, 1986 (H with two differentdecaytimes. Usingthe ChamberTo exemplify' the difficulties which the above two
lain model,givinga neutraldensity about310 atoms of

cm theycalculated 75keVhydrogen -3, for a decay pa- November 25-27, 1986. The measurementspresented
rameterof rH+ = 73.0 hoursand for 75 keV oxygen a
in this paper were made with the charge-energy-mass spectrometer (CHEM) on the AMPTE/CCE whichdetects ions in the energyper chargerange 1-310 keV/e. The data, collected during Pass8 (1549 UT - 1851 UT on November25, 1986) and Pass10 (0733 UT - 1029UT on November26, 1986), were used to calculate the en-

explanations encounter we have chosen the storm of

Copyright 2000 by theAmerican Geophysical Union.
Papernumber2000GL003783. 0094-8276/00/2000GL003783505.00
2813


2814

FELDSTEIN

ET AL.- ON TWO-PHASE

DECAY

OF THE DST

.20
ú '

recover,' phase. UsingAMPTE/CCE data the decay parameter oxygen 10.5 hoursand for hydrogen for is 71.· hours. However,analyzingthe data more closely, we find that during the storm'smain phasemaximum and the recovery phasethe oxygen ionsenergydensity is onlyabout, density 10 (seeFigure3 for L-shells 4./5-

5.0,thesame values onall L-shells 2./5 6./5). are from to
i:lil 80
-loo

-80
-lOO

This storm is not a special case. The oxygen ions con-

tributionto the total energyof the ring currentduring the mainphase about27-29the moderate was stormof September 1984[Gloeckler al., 1985;Krimigiset 4-7, et

-120
ú

,.,.,...·
....................

ergycan't explainthe first rapid decav the intensity of 23 24 25 26 27 28 of the Dst-variation during the beginning the recovof ' Day eryphase evenif theywould havedisappeared the from Figure 1. Dst geomagnetic index (left,axis) and DT ring current regioncompletely. magnetic fieldvariation(rightaxis)duringthe magnetic storm on November23-27, 1986. AMPTE/CCE pass
-140

ú

al.. 198.5]. is evident It that the measured oxygen en-

numbers and durations at 2.5 < L < 6.5 are shown on

the panel.

Discussion

and conclusions

We propose herean alternative explanation the for two-stage decay.We suggest Ds! that duringthe rephaseof the magnetic stormthe Dst decayis ergy densitiesfor dillbrent ions during the above given coycry controlled the decayof two different by currents:the time intervals. Figure 1 presentsthe Dst index between November 23 and 27. From 1800 UT on Novemring current (DR) and the magnetospheric curtail ber 25 until 0600 UT on November 26 one observes a rent (DT). The magnetic fieldvariation (DCF) controlledby the solarwind pressure (Psw) is not taken very rapid recoveryof the Dst values. AMPTE/CCE
passes shownon the panel. Figure 2 showsthe hyare drogenion energydensity(markedby dots), the oxygen energydensity(markedby the dashed line), and the total energydensity for five ions species (marked by the solidline). Figure 2 revealsthat during the main phase maximu·n (Pass 8) and during the next available pass (Pass10) the ionswhich are the carriersof the ring current create a singlelarge-scaleregion. The local time of the passes was 17 - 22 hr, the satellite crossed L-shells 2.5 - 6.· during 3 hours. Obviously,no breakdownin
two separate zones is seen.

into account because during the recovery phaseof this stormPsw changes than 2 nPa. The values the less of magnetotail currentsystem (DT) duringthe magnetic storm on November23-27, 1986 is being publishedin Drem?tkh, et al. [1999]and are presented Figure i',,a, in
1. We have used the DT values which were calculated

using pa. the raboloid model the magnetospheric of magnetic field by Ale.zeev al. [1996]. Comparing et the
calculated values of DT with the measured Dst values

Addressing Hamiltons'explanationwe note that there are indeed two ion populationswhich decayduring the
Pass 8

we see that DT contributessignificantlyto the Dstvaria. tio·{. Since we observethat DT recoversfirst,very fast and then remains in a narrow range xvedivide the recovery phaseof the magneticstorminto two phases.
Pass 10

'·
......

'·'--'Q(H+)
0(0+· _

ú

lOO·ú · /!
ú
40-

ú

ú
,
L

!
O'

,!
ú · . ........

,.
. ....
· ú ú . , ú i . ,

..."..
2 3 4 5

-·o

6

7

L

L

shows total O of ionsH+, O+, He+, He++, and N+. the

Figure 2. Energy density [keV/cm of ions function L forthemainphase O 3] as of maximum (pass andthe 8) recovery (pass Thethick withpoints phase 10). line shows+ data, dashed shows+ data, thinline H the line O the


FELDSTEIN
05- _
ú ...........................

ET AL.-

ON TWO-PHASE

DECAY

OF THE

DST

2815

5

at, the beginning of the recovery phase. In Figure 4a we

't

displaythe valuesln(DT·/DTi) ibr the first interval. In Figure 4b which shows the valuesln(DT·/DTi) for the
second interval, DT· is in this case the value of DT at

'/
6x,x,,, 8
.
40 50 60 70

9

10 11
80 90 100

13

-e TM
110 120

the beginningof pass9 (this point also marks the end of the fast,recoveryof DT). We use the data in Figure 4 to compute rOT during the beginning stage of the recoveryphase (on the left, panel) and during the final stageof the recoveryphase(on the right panel). It is seen that at, the beginning stage of the recovery
phase TDT is equal to 7.3 !·ours. This means that the

Time since 23. Nov 96 UT 0:00 [hours]

Figure

3.

Variation of the energy density ratio

Dst intensity decreases 2.7 times during the 7.3 hour interval. Since the DT intensity variation is the substantial part of the D st, we argue that it is possible to interpret the observedchangeof the D st magnetic
during the first stage of the recovery phase. After the

Q(O+)/Q(H +) in the ring currentregionduringthe field by the decay of the magnetotail current system
main and recovery phase of the magnetic storm at 4.5 < L < 5.0. The numbers in the figure denote the
satellite passes.

AnalogouslyD st can be divided into two phases,a fast and then a slowerrecoveryphase. The first,phaselasts ring current decay, mainly its proton constituent, with
until UT 04:00 of November 26 and then the second

first rapid decay of DT, DT decreases only gradually with row · 183 hours(on the right panel of Figure 4) and its contribution to the Dst-variation is negligible. Instead, the D st variations are now controlled by the
rH+ equal to 71.5 hours.

phasebegins.We noticethat the decayparameter rDT is of the order of a few hoursduring the first stageof the recovery phase.During the second stageof the recoveryphase, whenthe contribution DT is negligible, of the magnetic field variations are/nainly determin6d by the DR decay. From plasma,measurementswe know that the DR, decay parameter rDR is equal to several te]·sof l·oursand thereforethe magneticfield intensity changes slower are duringthe final stageof the recovery phasethan during the beginningstage. In orderto determinethe magnetotailcurrentdecay parameter rDT we assume that, the intensity DT decays exponentially.Thereforewe usedthe normal procedure to calculaterDT by formingthe ratios DT·/DTi in one
hour intervals where DT· is the minimum value of DT

Kozyr'a al. [1998] et havemodeled magnetic the storm
of February 1986 including ion precipitations into the
atmosphere as an additional loss mechanism. Liemohn

et al. [1999]alsotook into account gainsof particles the
and energythrough the nightsideboundary and the loss through the dayside boundary due to convectivedrift ibr modeling the two storms of July 1991 and September 1998. Both examinations show a good agreement with observed Dst values during the main phase and

early phase the Dst recovery. of Liemohnet al. [1999]
inferred that the convectivedrift, lossout of the dayside magnetopause the dominant,processin removing ring is current particles. The ratio of the dayside outflow energy loss rate to the charge exchangeenergy loss rate exceeded10 during the main phase for September 1998
In(DT/DTi)=O.OO5dT+0.085

In(DT/DTi)=0.1
1.5

2dT+0.1

2

·.s- I . ,n(·mz·m,) I
·1.0
1.0

0.5"

ú

0.5

0.0··

ú

.,

.

,

ú
0.0

b)
25 30 35 40

-0.5 .... ·............
0 2 4 6 8 10 12 14
0 5 10 15

...................

20

dT[hour]

dT[hour]

Figure 4. \:ariations the tail current of magnetic fieldintensities ratio (]n(DT·/DTi) duringthe earIvrecovery

phase andin thelaterecovery (a), phase Corresponding of thedecay (b). values parameter r are
r · 183 hours respectively.


2816

FELDSTEIN

ET AL.: ON TWO-PHASE

DECAY

OF THE DST

storm and always exceedsunity during the early recovery phasesof both storms. Possibly,this ratio changes from one storm to another storm. This is indirectly shownby the resultsof a.successful model by Kozyra et

Acknowledgments. The authors thank M. Greenspan fox' AMPTE/CCE ion data andL. Gromova discusthe for sions and help in the preparation the manuscript.The of xvork was supported grant RFBR 99-05-65611 by by and
ISSI, Bern.

al. [1998]which doesnot take into account, losses the
out of the dayside magnetopause. Results from calculations of particle and energy lossesdue to different loss

References
Akasofi·. S.-I.. S. Chapman, and D. Venkatesan, The main phase of great, magnetic storms. ,1. Geophys. Res., 68,
3345-3350, 1963.

mechanisms obtainedby Lieraohn et al. [1999]allow
one to interpret some known experimental features in the behavior of the ring current decay parameter ,- dur-

Alexeev. I. I.. E. S. Belenkaya, V. V. Kalegaev, Y. I. Feldstein and A. Grafe, Magnetic storms and magnetotail curing the main (r·p) and recovery(rRp) phases the of rents, .1. Geophys. Res., 101, 7737-7747, 1996. magneticstorm [Feldstein,1992]: 1) r·e valuesare Daglis, I. A., The role of magnetosphere-ionosphere cousmaller than is inferred from the chargeexchangemechpling in magnetic storm dynamics, in Magnetic Storms, Geophys. Monogr. Set., vol.98, edited by B.T. Tsurutani, anism. This meansthat dissipationis faster; 2) raze is W.D. Gonzalez, Y. Kamide, and J.K. Arballo, p. 107-116, controlled by the rate of energy injection into the inner AGU. XVashingt, DC, 1997. on, magnetosphere (F). r·,·e decreases while F increases Dessler. A. J.. and E. N. Parker, Hydromagnetic theory of (h'om r^te --- 8 hr for injection F --- -10 nT/hr to geomagnetic storms, J. Geophys. Res., 64, 2239-2248,
1959.

r·,·e --· 2 hr ibr F -.- -100 nT/hr; 3) parameter r has different values during the main phase and the recovery phase of a storm. V-Mp < rRp, that means faster dissipationof the ring current energy occursduring the main phasethan during the recoveryphase;4) we increases the ring current,decaysfrom as
DR...0 -100 nT to r·e -'- 14 hr for DR ,., -10 nT.

Dremukhina. L. A.. Y. I. Feldstein, I. I. Alexeev, V. V.

Kalegaev and M. E. Greenspan,Structure of the magnetospheric magnetic field during magnetic storms, J. Geophys. Res., 10·, 28351-28360, 1999.

Feldstein, Y.I., Modeling of the magnetic field of magnetospheric ring current as a function of interplanetary

medium, Space Sci. Rev., 59, 83-165, 1992. Feldstein, Y. I., A. E. Levitin, S. A. Golyshev, L. A. Dremukhi The above presentedexperimental features of the deU. B. Ves[chezerova, E. Valtchuk, A. Grafe, Ring curT. cay parameter's behavior is explained naturally on the rent and au,'oral electrojets in connection with interplanbasisof the conclusion Liemohn et al. [1999]that by etary medium parameters during magnetic storms, Ann. there is an essential influence of convective drift losses Geoph:qs.. 602-611. 1994. 12, Gloeckler, G., B. ¾Vilken,W. Studemann, F. M. Ipavich, on the magnetosphericenergy budget. Drift losses D. Hovestadt, D.C. Hamilton, and G. Kremser, First

through the dayside magnetosphere are more intensive during the main phaseof a storm than during the recovery phase, and they are the largest during the intensive injections. Although injection intensities decreaseas the recoveryphase begins,somelevel of injection exists throughout the whole recovery phase especially during its early stage. Presumably, a fast recovery of Dst after the main phase maximum of a storm can be caused by two reasons: particle lossesdue to drift, through the dayside magnetopause and rapid decay of the tail current system. The energy of ions drifting through daysidemagnetopause is few tens of keV while the tail current closed through daysidemagnetopauseis generatedin the magnetospheric plasma sheet by particles with characteristic energiesof about few keV. Thus these two sourcesfor rapid D st recovery are caused by two different plasma

composition measurement of the balk of the storm-time

ring current (1-300 keV/e) with AMPTE/CCE, Geophys.
Res. Lett., 12, 325-328, 1985. Hamilton D.C., G. Gloeckler, F. M. Ipavich, W. Studemann. B. ¾Vilken and G. Kremser, Ring current development during the great geomagnetic storm of February 1986. ,1. Geoph'ys.Res., 93, 14343-14355, 1988. Kozvra..l.U.. M.-C. Fock, E. R. Sanchez,D. S. Evans, D. C. Hamilton. and A. F. Nag5,, The role of precipitation losses producingthe rapid early recoveryphase of the in

great magneticstormof February1986,J. Geophys. Res.,
103, 6801-6814, 1998.

Krimigis,S., G. Gloeckler, W. McEntire, T. A. Potemra, R.
F. L. Scarf, and E.G. Shelly, Magnetic storm of September 4, 1984: a synthesisof ring current spectra and energy densities measuredwith AMPTE/CCE, Geophys. Res.
Left., 12, 329-332, 1985.

Liemohn, Ix4. W., J. U. Kozyra, V. K. Jordanova, G. V. Khazanov, M. F. Thomsen, and T. E. Cayton, Analysis of early phase ring current recoverymechanisms during

species and can exist simultaneously. The relative contribution of each speciesto the rapid Dst recovery can changefrom one storm to another. To determineit
elaborate calculation of each contribution due to every

geomagnetic storms,Geophys. Res. Lett., 26, 2845-2848,
1999.

Takahashi,S., T. Iemori, and M. Takeda, A simulationof the storm-timering current, Planet. SpaceSci., 38, 11331141,1990.

sourceto Dst-varia.tion is required. The Dessler-Parker

relationship [Dessler and Parker, 1959]holdsonly for t, rapped particlesin the Earth's magneticfield. One
should take into account that the relation of Dessler

Y. I. Feldstein and L. A. Dremukhina, IZMIRAN,Troitsk,

and Parker is used by Lieraohnet al. [1999]to determine the contribution to D st by particles drifting

142090, Troitsk, Moscow Region, Russia. (e-mail: lgrom @izmir troit sk.ru) ova an. U. ·[all (e-mail: mall@linmpi.mpg.de) and J. ¾Voch,
.·Iax-Planck-Institut fimr Aeronomie. D-37189 KatlenburgLindau. Germany

throughthe daysidemagnetopause. Theseparticlesare not trapped and the magneticfield generatedby them
should be determined by another procedure.

(ReceivedMarch 16,2000; revisedMay 31,2000;
acceptedJune 7,2000.)