ApJ, 594, 1068

17 Pages·2003·2.32 MB·English

TheAstrophysicalJournal,594:1068–1084,2003September10 E
#2003.TheAmericanAstronomicalSociety.Allrightsreserved.PrintedinU.S.A.
DYNAMICAL AND PHYSICAL PROPERTIES OF A POST–CORONAL MASS EJECTION CURRENT SHEET
Yuan-KuenKo,1 JohnC.Raymond,1 JunLin,1 GarethLawrence,2 JingLi,3 and AndrzejFludra4
Received2002October18;accepted2003May10
ABSTRACT
IntheeruptiveprocessoftheKopp-Pneumantype,theclosedmagneticfieldisstretchedbytheeruptionso
muchthatitisusuallybelievedtobe‘‘open’’toinfinity.Formationofthecurrentsheetinsuchaconfigura-
tionmakesitpossiblefortheenergyinthecoronalmagneticfieldtoquicklyconvertintothermalandkinetic
energies and cause significant observational consequences, such as growing postflare/CME loop system in
thecorona,separatingbrightflareribbonsinthechromosphere,andfastejectionsoftheplasmaandthemag-
neticflux.Aneruptionon2002January8providesusagoodopportunitytolookintotheseobservational
signatures of and place constraints on the theories of eruptions. The event started with the expansion of a
magnetic arcade over an active region, developed into a coronal mass ejection (CME), and left some thin
streamer-likestructureswithsuccessivelygrowingloopsystemsbeneaththem.Theplasmaoutflowandthe
highlyionizedstatesoftheplasmainsidethesestreamer-likestructures,aswellasthegrowingloopsbeneath
them,leadustoconcludethatthesestructuresareassociatedwithamagneticreconnectionsite,namely,the
currentsheet,ofthiseruptiveprocess.WecombinethedatafromtheUltravioletCoronagraphSpectrometer,
LargeAngleandSpectrometricCoronagraphExperiment,EUVImagingTelescope,andCoronalDiagnostic
Spectrometer on board the Solar and Heliospheric Observatory, as well as from the Mauna Loa Solar
Observatory Mark IV K-coronameter, to investigate the morphological and dynamical properties of this
event, as well as the physical properties of the current sheet. The velocity and acceleration of the CME
reachedupto1800kms(cid:1)1and1kms(cid:1)2,respectively.Theaccelerationisfoundtooccurmainlyatthelower
corona(<2.76R ).Thepost-CMEloopsystemsshowedbehaviorsofbothpostflareloops(upwardmotion
(cid:2)
with decreasing speed) and soft X-ray giant arches (upward motion with constant speed, or acceleration)
accordingtothedefinitionofSˇvestka.Inthecurrentsheet,thepresenceofhighlyionizedions,suchasFe+17
andCa+13,suggeststemperatureashighasð3 4Þ(cid:3)106K,andtheplasmaoutflowshavespeedsrangingfrom
300to650kms(cid:1)1.Absoluteelementalabundancesinthecurrentsheetshowastrongfirstionizationpotential
effectandhavevaluessimilartothosefoundintheactiveregionstreamers.Themagneticfieldstrengthinthe
vicinityofthecurrentsheetisfoundtobeoftheorderof1G.
Subjectheadings:Sun:corona — Sun:coronalmassejections(CMEs)
On-linematerial:colorfigures
1. INTRODUCTION flares,solarenergeticparticles,andradiobursts.Filaments
sometimes also constitute part of the CME from active
Coronal mass ejections (CMEs) are powerful, transient
regions. There have been many multiwavelength studies
expulsions of coronal material into interplanetary space.
of individual CME events and statistical studies of the
Since the launch ofthe Solar and Heliospheric Observatory
CME properties (e.g., Howard et al. 1985; Hundhausen,
(SOHO), a large number of CMEs have been observed,
Burkepile,&St.Cyr1994;Goldstein,Newgebauer,&Clay
fromtheirinitiationatthesolarsurfacetothepropagation
1998;Innesetal.1999;St.Cyretal.1999,2000;Ciaravella
through the corona into interplanetary space. Because of
et al. 2000; Akmal et al. 2001; Gopalswamy et al. 2001).
their intimate association with the solar cycle and their
ManyeffortshavebeenmadetomodelCMEeruptionproc-
potentially strong effect on the space environment at the
esses in an attempt to understand their initiation, explain
Earth, CME study has been a very vigorous field in both
theobservedstructureanddynamicalproperties,andmodel
observationalandtheoreticalresearch.
their propagation in the heliosphere (e.g., Mikic´ & Linker
ThereareingeneraltwotypesoforiginforCMEs.Oneis
1994; Chen 1996; Antiochos, DeVore, & Klimchuk 1999;
thefilamentorprominenceeruptionfromquiet-Sunregions
Amarietal.2000;Lin&Forbes2000;Low2001;Rileyetal.
that usually contains cool chromospheric material and
2002; Manchester et al. 2003; for reviews, see, e.g., Wu,
exhibits complex twisting structure. The other is a more
Andrews,&Plunkett2001).
powerfuleruptionfromactiveregionsoftenassociatedwith
Thegenerallyacceptedexplanationfortheenergysource
of eruptive phenomena in the solar atmosphere, such as
solarflares,eruptiveprominences,andCMEs,istheenergy
stored in the coronal magnetic field prior to the eruption.
1Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
Cambridge,MA02138. These eruptive phenomena commonly involve fast mass
2CenterforSolarPhysicsandSpaceWeather,CatholicUniversityof ejectionsandintenseheating,whichimplyrapiddissipation
America,Washington,DC20064. of the magnetic energy. In the coronal environment, the
3InstituteforAstronomy,UniversityofHawaii,2680WoodlawnDrive,
electricalconductivityoftheplasmaandthelengthscaleof
Honolulu,HI96822.
themagneticfieldarefairlylarge,sorapiddissipationofthe
4Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire
0X110QX,UK. magneticfieldisalmostimpossible(e.g.,Priest1982;Priest
1068
POST–CORONAL MASS EJECTION CURRENT SHEET 1069
are heated by shocks due to reconnection (Forbes,
Malherbe, & Priest 1989). The cool loops, which lie inside
the hot loops, are formed from the hot loops by radiative
cooling(Sˇvestkaetal.1987;Forbes&Acton1996).Obser-
vations supporting this scenario include rising X-ray loops
(e.g., Sˇvestka 1996), cool loops lying inside the hot loops
(e.g., McCabe 1973; Sˇvestka et al. 1987; Schmieder et al.
1996), and Dopper shift measurements (Schmieder et al.
1987) providing evidence of chromospheric evaporation.
The density of the postflare loops was found to be a few
times 1010 cm(cid:1)3 in the X-ray/EUV loops (e.g., Withbroe
Fig.1.—Postflare/CMEloopmodelbyKopp&Pneuman(1976).(a) 1978)andashighas1012cm(cid:1)3intheH(cid:1)loops(e.g.,Heinzel
Themagneticfieldispushedopenbyaneruption,andacurrentsheetsepa-
&Karlicky´ 1987).Botharemuchdenserthanwhatisfound
rates two antiparallel magnetic field lines. (b) The opened configuration
relaxesintoaclosed,nearlypotentialfieldviathemagneticreconnectionin abovenontransientsolarstructure,whichisgenerallybelow
thecurrent sheet.Thisprocessproducestwobrightandseparatingflare 109cm(cid:1)3justabovethesolarsurface(e.g.,Guhathakurta&
ribbonsonthesolardiskandacontinuallygrowingflareloopsysteminthe Fisher1995;Parentietal.2000).
corona(reproductionofFig.1ofKopp&Pneuman1976).
Recent research indicates that the closed magnetic field
does not necessarily become fully open as the Kopp-
&Forbes 2000) unless the magneticconfiguration includes Pneuman–typeeruptionoccurs.Instead,themagneticstruc-
someneutralregions,suchasneutralpoint,nullpoint,cur- tureishighlystretched by the eruption, and a current sheet
rent sheet, or quasi-separatrix layers (e.g., De´moulin et al. forms(seeFig.2).Bothnumericalsimulationsandanalytical
1996; Priest & Forbes 2000), in which the fast magnetic studies show that when the footpoints of a closed magnetic
reconnectioncanoccureasily. structure are sheared (Mikic´ & Linker 1994), or the whole
Kopp & Pneuman (1976) proposed a scenario for two- structure is stretched outward by the catastrophic loss of
ribbonflares(Fig.1;theideawasraisedbyCarmichael1964 equilibriuminamagneticfluxrope(Forbes&Isenberg1991;
forthefirsttime.)Inthismodel,energyisstoredinaforce- Isenberg,Forbes,&De´moulin1993;Forbes&Priest1995),a
free magnetic arcade or flux rope prior to the eruption. current sheet develops in the stretched configuration. With
Eventually, the field erupts to form a fully open magnetic dissipation inthe current sheet, the stretched magneticfield
configuration including a neutral current sheet that sepa- lines start to reconnect, producing new closed field lines
ratestwoantiparallelmagneticfieldlines(Fig.1a).Finally, below the current sheet, so the preexisting closed magnetic
the open configuration relaxes into a closed, nearly poten- field never becomes fully open. There are several ways to
tialfieldviamagneticreconnectioninthecurrentsheet,and trigger dissipation in the current sheet. The current sheet is
the magnetic energy is converted into kinetic and thermal subjecttothetearing mode instability,and magneticrecon-
energies (Fig. 1b). The observational consequences of this nection occurs as soon as the length of the current sheet
process are two bright and separating flare ribbons on the becomes longer than (cid:4)2(cid:2) times its width (Furth, Killen, &
solardiskandacontinuallygrowingflareloopsysteminthe Rosenbluth 1963). Alternatively,as the current sheet builds
corona.Themotionsoftheflareribbonsandloopsarenot up,itscurrentdensitymayexceedthethresholdofamicro-
duetomassmotionsoftheplasmabuttotheupwardpropa- instability, which creates anomalous resistivity (Galeev &
gationoftheenergysourceontonewmagneticfieldlinesat Zelenyi 1975; Heyvaerts & Priest 1976) and triggers rapid
higheraltitude(Schmiederetal.1987).AccordingtoFigure reconnection. The scenario sketched by Figure 2 suggests
1,theupwardpropagationoftheenergysourceimpliescon- thattheroleplayedbyreconnectioninaneruptiveprocessis
tinuous erosion of the current sheet due to reconnection. twofold. First, reconnection breaks the magnetic field lines
As pointed out by Forbes (2000), images obtained from thatpassoverthefluxropeandareanchoredintheboundary
YohkohSoftX-rayTelescope(SXT)providemanypiecesof surface at both ends. These field lines produce strong mag-
evidence that suggest magnetic reconnection in the corona netic tension force in a stretched configuration that would
during an eruption. These include a hard X-ray source otherwise prevent the flux rope from escaping to form a
located above the soft X-ray loops (Sakao et al. 1992; CME.Second,reconnectiondumpslargeamountsofenergy
Masudaetal.1994;Bentleyetal.1994),cuspstructuresug- intheloweratmosphereoftheSun,creatingintenseheating,
gesting either an X-type or a Y-type reconnection site which accounts for the traditional flare ribbons and loops.
(Acton et al. 1992; Tsuneta 1993; Doschek, Strong, & Figure2incorporatestheflaremodelssummarizedbyForbes
Tusneta 1995), bright features at the top of the soft X-ray &Acton(1996)andtheCMEmodelofLin&Forbes(2000,
loops (Tsuneta et al. 1992; McTiernan et al. 1993), and hereafterLF00).Itclearlyillustratestherelationshipamong
high-temperature plasma along the field lines mapping to solarflare,eruptiveprominence,andCMEandsuggeststhat
the tip of the cusp (Tsuneta 1996). Recently, Yokoyama thesethreeeruptivephenomenaareactuallydifferentmanifes-
et al. (2001) reported the observation of plasma inflow tationsofasinglephysicalprocessthatinvolvesadisruptionof
mergingtoalocalareaabovethecuspstructure. thecoronalmagneticfield.
Observationsshowthatbrightribbonsandloopscanlast OneofthemostsignificantpredictionsoftheLF00model
formorethan10hr(e.g.,Sˇvestka1976).Ribbonsandloops is that a current sheet develops following the onset of the
appeartomovethroughthechromosphereandthecorona, eruption.Becausethetimescaleofreconnectionislongcom-
andtheoutermostedges ofthehotloopsmaptotheouter pared with the timescale of the onset stage (Alfve´n time-
edgeoftheflareribbons(Schmiederetal.1996;Antiochos scale), dissipation of the current sheet by reconnection is
etal.2000),whilethecoolH(cid:1)loopsmaptotheinneredge slow,sothecurrentsheetisabletobecomefairlylong.The
oftheribbons(Rust&Bar1973).Accordingtothemodels, evolutionofthecurrentsheetissignificantlyconstrainedby
thehotloopsarethenewlyreconnectedclosedloops,which the local Alfve´n speed. In an isothermal corona, the local
1070 KO ET AL. Vol. 594
tion perpendicular to its vertical extent (refer to Fig. 2).
Therefore,itsthicknessissosmallthatweusuallytreatthe
currentsheetasaninfinitelythinlayer.Inaddition,thelow
plasmadensityinthecoronaimpliesthattheplasmainside
thecurrentsheetshouldbetenuous.Therefore,itisdifficult
toobserve thecurrent sheetinan eruptiveprocess because
both its size and emission are easily dominated by other
large-scale and bright structures nearby. Ciaravella et al.
(2002) analyzed the spectral data obtained from the Ultra-
violetCoronagraphSpectrometer(UVCS;Kohletal.1995)
on SOHO after an eruption on 1998 March 23. After the
bulkCMEmateriallefttheinnercorona,abrightpostflare/
CMEarcadewasseenbytheEUVImagingTelescope(EIT)
and Yohkoh/SXT. UVCS detected a very narrow and hot
featurethatwasmostprominentinthe[Fexviii]emissionin
the space between the arcade and the CME core. The
appearance of the [Fe xviii] emission means a temperature
as high as 6(cid:3)106 K in that area (Ciaravella et al. 2002).
Theirresultisquiteconsistentwiththetheoreticalexpecta-
tionsofthecurrentsheetinthefluxropemodelsketchedin
Figure2.Webbetal.(2003)analyzed26‘‘candidatediscon-
nectionevents’’(CDEs)associatedwithCMEsobservedby
theSolarMaximumMission(SMM).TheseCDEswerefol-
lowedbyraylikestructuresthatappearedafewhoursafter
the CDE and sometimes were found to connect to newly
formedstreamersclosetothelimb.Theyregardthiskindof
raylikestructuresasevidenceofthecurrentsheetdeveloped
followingaCME(cf.Fig.2).
Arecenteruptionprovidesuswithauniqueopportunity
forscrutinizingthecurrentsheetaswellastheglobalmor-
phologicalfeaturesoftheeruption.Thiseruptionoccurred
attheeastlimboftheSunon2002January8.TheCMEleft
behind some long and thin streamer-like structures with
continuous outflows as seen in white light. Lower in the
corona, UV/EUV observations showed formation of the
Fig. 2.—Upper part: Sketch of the flux rope/CME model of Lin & post-CME loops, and an electron temperature of
Forbes (2000) showing the eruption of the flux rope, the current sheet ð3 4Þ(cid:3)106Kwasfoundat1.53R .Temperaturesthishigh
(cid:2)
formed behind it, and the postflare/CME loops below, as well as the at this height are not commonly seen in the quiescent
inflows and outflows associated with the reconnection. Lower part:
corona, even above most active regions (Ciaravella et al.
Enlargedviewofthepostflare/CMEloops(adoptedfromForbes&Acton
2002).Webelievethatthesestreamer-likestructuresconsist
1996). The upper tip of the cusp rises as reconnection happens
continuously. of a current sheet in which the magnetic reconnection
occurs. The observed high temperature results from the
heating in the magnetic reconnection processes that also
Alfve´nspeedincreaseswithheightatlargealtitudes,sothe produce continuous outflows along the current sheet and
current sheet is quickly eroded by reconnection, and its the formation of the post-CME loops beneath it. In many
length consequently shrinks several hours after the onset respects,thiseruptionislikemanyobservedpreviously,but
(LF00; Forbes & Lin 2000). In a more realistic corona, on the location of the source region (AR 9782/9785; see the
theotherhand,thelocalAlfve´nspeeddecreaseswithheight nextsection)andtheorientationofthedisruptedmagnetic
at large altitudes; erosion of the current sheet is thus not configuration lead to a current sheet just along the line of
veryfast,andthecurrentsheetmayremainalongtime(Lin sight(i.e.,edge-on),asillustratedinFigure2.Thismakesit
2002, hereafter Lin02). The model predicts that the mag- possibleforustoanalyzetheplasmapropertiesandinvesti-
netic energy around the current sheet is converted into gate the evolution of the current sheet, as well as the post-
kineticandthermalenergyoftheplasmaduringtherecon- CME loops below it. In the next section, we describe this
nection process. The plasma then flows along the current event as observed in white light and UV/EUV and find
sheet both upward and downward at approximately the manyaspectsconsistentwiththepredictions ofthemodels
Alfve´n speed and is heated to high temperature inside the described in this section. In x 3, we analyze the dynamical
currentsheet(comparedtotheambientcorona).Themodel and physical properties of this current sheet and interpret
alsopredictscertaindynamicalpropertiesoftheCMEand thedatabasedontheworkbyLF00andLin02.Finally,we
the current sheet that we will discuss in detail in a later summarizeouranalysisinx4.
section.
Thereis,however,littledirectobservationalevidencethat
2. THE OBSERVATIONS
can confirm the occurrence of the current sheet. The high
electrical conductivity and nearly force free environment Wesuggestedintheprevioussectionthatsolarflares,erup-
confinethecurrentsheettoaverylocalregioninthedirec- tiveprominences,andCMEsaredifferentmanifestationsof
No. 2, 2003 POST–CORONAL MASS EJECTION CURRENT SHEET 1071
a single physical process that involves a disruption of the
coronalmagneticfieldfollowedbymagneticreconnection.In
thissection,wedescribeaCMEfollowedbytheformationof
thepost-CMEloopsandthecurrentsheetthatprovidesgood
illustrationofthisidea.ThiseventwasobservedbytheLarge
Angle and Spectrometric Coronagraph Experiment
(LASCO;Brueckneretal.1995),EIT(Delaboudinie`reetal.
1995),UVCS,andCoronalDiagnosticSpectrometer(CDS;
Harrisonetal.1995)onboardSOHOaspartofSOHOJoint
ObservingProgram151,aswellasbytheMaunaLoaSolar
Observatory (MLSO) Mark IV K-coronameter (MK4).
Because of the dynamical nature of all eruptions, it may be
helpful to refer to the movies5 of this event recorded by
SOHO/LASCOandSOHO/EIT.
2.1. TheCoronalMassEjectionandtheCurrentSheet
On2002January8,amajoreruptionoccurredattheeast
limboftheSun.Thecorrespondingsourceactiveregion(see
x 2.4) was located just behind the limb. Therefore, it is
unclear whether a filament preexisted in the relevant mag-
netic configuration and erupted or whether this eruption
started with a flare. The movie obtained by EIT 195 A˚
Fig.4.—Height-timeplot(filledcircles)oftheleadingedgeoftheCME
showsmovement/expansionofthecoronalloopsthattook
measuredinLASCOC2andC3withthefittingcurveofasecond-order
placeforafewhoursbeforetheeruptionwasclearlyseenat polynomial(solidline).Aconstantaccelerationof0.19kms(cid:1)2providesa
17:36UT.Intheimageoneframebefore(17:24UT),there goodfittothedata.AlsoplottedisthetimeframethatbracketstheCME
is some clearer signature of the expansion of the magnetic onsettime,indicatingalargeraccelerationof(cid:4)1kms(cid:1)2inthelowercorona
below2.76R .
arcade.Figure3showstheCMEonthecompositeimageof (cid:2)
EIT195A˚ runningdifferenceat18:24UTandLASCOC2
at 18:30 UT. The overall coronal disruption due to the
CME eventually spanned 180(cid:5) between the two poles by 17:54UT.Itthensubsequentlytraveledto3.34R(cid:2)at18:06
18:54 UT. Figure 4 is the height-time plot of the leading UT,5.25R(cid:2)at18:30UT,and14.12R(cid:2)(inC3)at19:42UT.
These data points fit well to a curve that corresponds to a
edge of the CME. The leading edge of the CME first
appearedintheLASCOC2fieldofview(FOV)at2.76R , constant acceleration rate of about 0.19 km s(cid:1)2. Further-
(cid:2)
more, from this curve, we can evaluate the tangents at
pointsandinferspeedsof,forexample,around700kms(cid:1)1
5Availableathttp://lasco-www.nrl.navy.mil/daily_mpg. at3R toover1800kms(cid:1)1by14R .ThisisafastCMEby
(cid:2) (cid:2)
comparisonwiththeoverallCMEpopulation(St.Cyretal.
2000).
Wenotethatwiththisconstantaccelerationrate,thetime
when the CME started would be much earlier than as
observedinEIT.Therefore,theremustbeadifferentaccel-
erationratebeneaththeC2occultingdisk.Alsoplottedon
Figure4aretwopointsbracketingthetimeframewhenthis
CMEprobablystarted(between17:24and17:36UTofEIT
images). Taking 17:30 UT as the CME initiation time, we
obtainanaverageaccelerationofabout1kms(cid:1)2between1
and2.76R .ThisimpliesthattheCMEunderwentamuch
(cid:2)
larger acceleration in the lower corona than that seen in
LASCO C2. Previous studies found low-coronal accelera-
tions of as much as 2–7 km s(cid:1)2 (e.g., Zhang et al. 2001;
Alexander,Metcalf,&Nitta2002).Usinglow-coronadata
of MLSO MK3 and MK4, along with the SMM and
LASCO data to the outer corona, Burkepile et al. (2002)
found that the largest CME acceleration occurs in the low
corona(<2.9R ).
(cid:2)
AfterthefrontofthisfastCMEhadpassed,thinthreads
of streamer formed behind it at a position angle (P.A.) of
(cid:4)101(cid:5),as indicated by the arrow in the upper left panel of
Figure 5. Blobs of material can be seen flowing along the
streamer continuously as it gradually moved northward
Fig.3.—CompositeimagefromEIT195A˚ runningdifferenceat18:24 (Figs.5,6,and7).Weinterpretthisthinstreamerasthecur-
UT,andLASCOC2at18:30UT,showingthe2002January8CMEerup- rent sheet left behind the CME. The blobs are the plasma
tion.ThenonradialexpansionofthefieldlinesinthisCMEextendedfrom
outflow that is produced by magnetic reconnection in the
thesolarsurfacewellintotheoutercorona.[Seetheelectroniceditionofthe
Journalforacolorversionofthisfigure.] current sheet (see x 1). They result from the magnetic
1072 KO ET AL. Vol. 594
Fig.5.—EITandLASCOC2compositeimagesshowingtheevolution Fig.6.—EITandLASCOC2compositeimagesshowingtheevolution
ofthecurrentsheetfeatureandthepost-CMEloops.Upperleft:EIT195A˚ of the current sheet feature(indicatedby the arrow) andthe post-CME
at19:48UT,C2at19:54UT,January8.Thearrowpointstothecurrent loops.Upperleft:EIT195A˚ 11:24UT,LASCOC2at11:30UT,January9.
sheet feature (CS) at its early appearance. Upper right: EIT 195 A˚ at Upperright:EIT195A˚ at16:48UT,C2at16:54UT,January9.Lowerleft:
21:17UT,C2at21:32UT,January8.Thearrowspointtothefirstappear- EIT195A˚ at06:48UT,C2at06:54UT,January10.Lowerright:EIT195
anceofthepost-CMEloopsintheEITimageandthecurrentsheetfeature. A˚ at23:18UT,C2at23:13UT,January10.NotethattheEITimageisnot
Lowerleft:EIT195A˚ at00:48UT,C2at00:54UT,January9.Lowerright: toscalewithC2inordertoshowthefeaturenearthelimbmoreclearly(see
EIT195A˚ at06:24UT,C2at06:30UT,January9.Thearrowspointtothe also Fig. 9 for enlarged EIT images). [See the electronic edition of the
northandsouthpost-CMEloopsystems.NotethattheEITimageisnotto Journalforacolorversionofthisfigure.]
scalewithC2inordertoshowthefeaturenearthelimbmoreclearly(see
also Fig. 9 for enlarged EIT images). [See the electronic edition of the
Journalforacolorversionofthisfigure.]
obvious that the thin streamer of interest (i.e., the current
sheetfeatureatP:A:(cid:4)83(cid:5))seeninC2isassociatedwiththe
reconnection in the nonuniform plasma and the magnetic southernpartoftheloopsystem.Wenotethattheappear-
field. anceofthetwoloopsystemsintheMK4image(seetheinset
Several other CMEs occurred after this January 8 CME in the left panel of Fig. 10) is similar to the giant X-ray
event.Particularly,aCMEfromtheeastlimbat12:06UT loops/cuspformedafteraflare/CME(e.g.,Tsuneta1996),
ofJanuary10seemstopushthiscurrentsheetfarthernorth- andthethinfeatureabovethesouthernloopresemblesand
ward.Inaddition,abrightstreamerstartedtoappearatthe is along the same direction as the streamer/current sheet
northeastlimbonJanuary10andoverlappedwiththiscur- featureseeninLASCOC2.Sincethedensityinsidethecusp
rent sheet feature on the plane of thesky (see Fig. 6, lower isexpectedtobehigherthanthatintheambientcorona(see
panels).However,asweexaminetheC2runningdifference x1),the‘‘loop’’seenintheMK4imagemayactuallybethe
images,wefindthatthiscurrentsheetfeaturepreservedits white-lightcounterpartoftheX-rayloops/cusp.However,
general morphology despite this January 10 CME, and it it is not easy to definitely associate the north loop system
appeared to be a separate feature, distinct from the bright seeninMK4andEITwithstructuresseeninC2,althoughit
streamertoitsnorth(Fig.8). ispossiblethatitisassociatedwiththetwothin‘‘spikelike’’
streamers just north of the long current sheet of interest.
This‘‘fan’’ofspikesseeninC2(cf.alsoFigs.5and6)seems
2.2. Post-CMELoops
to be analogous to the ‘‘fan of spikelike rays’’ seen above
Thepost-CMEloopsstartedtoappearinEITataround theflareloopsbyYohkohSXT(McKenzie&Hudson1999).
P:A:(cid:4)90(cid:5) at 21:17 UT, January 8, as indicated by the Thelatterwasfoundtohaveaone-to-oneassociationwith
arrowinthetopleftpanelofFigure9.Afterthat,aseriesof the X-ray loops in the flare arcade and was interpreted as
loops appeared at the limb and successively appeared at thecurrentsheet,consistentwiththestandardreconnection
higher altitude (see Fig. 9). They gradually faded on Janu- picture.
ary 10. This indicates continuous reconnection along the Figure11plotstheheightversustimeoftheouteredgeof
current sheet as described in x 1 to power the flare loops. the two post-CME loop systems measured from the EIT
Theseloopsappeartobeintwogroupsascanbeseenfrom images.Themeasurementsendwhenthetopoftheloopsis
Figure9(e.g.,indicatedbythetwoarrowsinthemiddleleft out ofthe EIT FOV (in the case of thenorthloops)or the
and bottom left panels). Figure 10 shows the MLSO MK4 loops are too faint to be visible (in the case of the south
white-lightimageon17:52UTofJanuary9(leftpanel)and loops).Theseloopsystemsseemtoshowbehaviorsofboth
a composite image of EIT (at 17:36 UT), MK4 (at 17:52 the postflare loops (the south loops, upward motion with
UT), and LASCO C2 (at 17:54 UT) (right panel). It is decreasingspeed)andthesoftX-raygiantarches(thenorth
No. 2, 2003 POST–CORONAL MASS EJECTION CURRENT SHEET 1073
Fig.8.—LASCOC2runningdifferenceimagesshowingthecurrentsheet
featureatP:A:(cid:4)78(cid:5).Left:at16:54UT,January10;middle:at23:13UT,
January10;right:at02:30UT,January11.Thecurrentsheetpersistedeven
afteraCMEeruptionat12:06UT,January10,althoughitwaspushedabit
tothenorth.ItgraduallydissipatedonJanuary11.
(lessthan5%when(cid:6)1daytothelimbasinthiscase)com-
paredwiththegrowthoftherisingloops.TheMK4images
maybeusedtoroughlyestimatetheheightofthe‘‘loops.’’
Wefindthatat17:52UTofJanuary9,thenorthloop(see
Fig.10)isat(cid:4)1.7R andthesouthloopisat(cid:4)1.4R .Note
(cid:2) (cid:2)
thatifthewhite-lightloopsindeedcorrespondtotheX-ray
Fig. 7.—Six consecutive LASCO C2 running difference images from
03:06UTto04:54UTofJanuary9showingcontinuouslyappearingblobs
flowingalongthecurrentsheetfeatureatP:A:(cid:4)89(cid:5).
loops,upwardmotionwithconstantspeed,oracceleration)
as defined by Sˇvestka (1996). The former group of loops Fig.9.—EIT195A˚ imagesshowingamoredetailedviewoftheevolu-
approachedtheheightof0.3R abovethesolarlimbovera tionofthepost-CMEloops.TheseloopsfirstappearedintheEITimageat
(cid:2)
21:17UT,January8(indicatedbythearrowinthetopleftpanel),andseem
periodof40hr,andthelattergroupofloopsroseabove0.4
tobeintwogroups(indicatedbythearrowsinthemiddleleftandbottom
R abovethelimboveraperiodof(cid:4)20hrandappearedto
(cid:2) left panels). The subsequent appearance of loops toward higher height
continue growing. Note that the projection effect is small impliescontinuousmagneticreconnection.
1074 KO ET AL. Vol. 594
942.2–1097.7A˚ fortheOviprimarychannel(471–549A˚ in
second order). Several other lines of interest can also be
obtained from the O vi redundant channel. They are Ly(cid:1),
Mgx(cid:3)609,[Fexii](cid:3)1242,Nv(cid:3)1238,and[Sx](cid:3)1196.For
a detailed description of the UVCS instruments, see Kohl
et al. (1995). The raw data have been wavelength and flux
calibrated. The uncertainty in the flux calibration is about
20%forthefirst-order lines (Gardneretal. 2000) and 50%
forthesecond-orderlines(Gardneretal.2002).Straylight
is negligible in the data, so no stray light correction is per-
formed. Figure 12 plots the spatial distribution of several
majorlinesandtheintensityratiooftheOvi(cid:3)1032/(cid:3)1037
doublet (Fig. 12j). We separate the data into four spatial
zones,whichareindicatedontheplots.Figure13plotsthe
spatiallyaveragedspectraforzones2,3,and4takenatone
ofthe grating positions where we measure most ofthe line
fluxes.Wecanseethatzone3containsstrongemissionfrom
highly ionized ions (such as [Fe xviii] (cid:3)974 and [Ca xiv]
Fig.10.—Left:MLSOMK4white-lightimageat17:52UT,January9,
(cid:3)943)thatarerarelyseeninthecoronaatthisheight.This
showing two loop/cusp systems after the CME. The inset shows our
high-temperature emission is concentrated around
perception of what the two loop/cusp systems might look like. Right:
CompositeofEIT195A˚ (at17:36UT),MK4(at17:52UT),andLASCO P:A:¼81=3(cid:6)3=7(halfoftheFWHM),whichisalongthe
C2(at17:54UT)imagesofJanuary9showingthatthecurrentsheetfeature same direction asthe thincurrent sheet feature seen inC2.
seeninC2correspondstothesouthloopsysteminMK4andEIT.[Seethe Zone 2 contains enhanced emission from low-ionization
electroniceditionoftheJournalforacolorversionofthisfigure.] ionssuchas[Fex](cid:3)1028and[Siviii](cid:3)944.Wesuspectthat
itisrelatedtotheeruptedprominencematerialinthelegof
loops, we would expect them to be above the EUV loops the CME (that occurred starting 12:06 UT, January 10, in
(seex1),whichisconsistentwithFigure11. C2asmentionedbefore)justsouthofthecurrentsheet(see
Fig.6,lowerrightpanel,andFig.8).Zone4wouldcontain
materialfromthebrightstreamernorthofthecurrentsheet.
2.3. High-TemperatureEmission Table 1 lists the line intensities for all four zones. In the
SOHO/UVCS observed this current sheet feature from followinganalysisoftheUVCSdata,wewillconcentrateon
17:48 UT, January 10, to 03:47 UT, January 11. The data zones 2, 3, and 4 only. The fluxes measured for zone 1 are
reported here were taken from the O vi channel. The slit listedbutarenotrelevantforthepresentstudy.
widthwassetto100lm,whichcorrespondstoaninstanta- Figure 12j plots the O vi (cid:3)1032/(cid:3)1037 intensity ratio,
neous FOV of 400(cid:3)2800. The slit angle was at P:A:¼78(cid:5) which can be used as an indicator of the outflow speed of
and a projected heliocentric height of 1.53 R . The data the ions because of the Doppler dimming effect (Hyder &
(cid:2)
wereobtainedwithaspatialbinningof10pixels(7000)anda Lites 1970; Noci, Kohl, & Withbroe 1987; Strachan et al.
spectral binning of 2 pixels (0.198 A˚). Using four grating 2000).Whentheoutflowislessthan100kms(cid:1)1,the(cid:3)1032/
positions, UVCS data have extended spectral coverage of (cid:3)1037ratioisbetween2and4dependingmainlyontheout-
flowspeedandtheelectrondensity(andweaklyontheelec-
tron temperature)—the higher the speed or the larger the
density,thesmallertheratio.Whentheoutflowexceeds100
km s(cid:1)1, the ratio goes below 2 because of the pumping of
Ovi (cid:3)1037.6 by Cii (cid:3)1037.0 (Noci et al. 1987). Figure 12j
shows that the ratios are larger than 2 in all four zones,
whichimpliesthattheoutflowspeedislessthan100kms(cid:1)1.
Zone3hasthelowestvaluesoftheratio,whichimpliesthat
eithertheelectrondensityortheoutflowspeed(orthecom-
binedeffectofthetwo)isthehighestinzone3.Sincethere
areprobablybothcurrentsheetandbackgroundcorona(as
inzone4;seex3.2)overlappinginthelineofsight,itispos-
sible that the outflow speed in the current sheet is higher
than 100 km s(cid:1)1 (which should give the O vi doublet ratio
below2)butthe‘‘average’’Ovidoubletratioisstillabove
2asintegratedalongthelineofsightwithmostcontribution
from a much lower outflow speed in the background
corona.
SOHO/CDSmadeobservationsfrom07:00to15:30UT
on January 10. The observation sequence consists of six
rasters (ARDIAG_2), each with FOV of 40(cid:3)40. Twenty
spectral windows include a wide selection of transition
regionandcoronallines,providingtemperaturediagnostics
Fig.11.—Height-timemeasurementofthetwopost-CMEloopsystems
as well as relative abundances of Fe, Si, Ca, Ne, Mg, and
showingthemrisingabovethelimb.Sincetheposition(longitude)ofthese
loopsshouldbewithin1dayofthelimb,theprojectioneffectissmall. O. Figure 14 shows the images in Mg ix 368.07 A˚ (106 K
No. 2, 2003 POST–CORONAL MASS EJECTION CURRENT SHEET 1075
Fig.12.—SpatialdistributionofseveralmajorlinesfromUVCSdataandtheintensityratiooftheOvi(cid:3)1032/(cid:3)1037doublet(Fig.12j)
formation temperature) and Fe xvi 360.8 A˚ (2:5(cid:3)106 K) observed by UVCS (see Fig. 12). Even though these loops
composed from the six rasters. Both images show multiple maybemixedwithotherpartsoftheactiveregion,theouter
loop structure extending up to the edge of the FOV. For boundaryoftheloopsisprobablyassociatedwiththepost-
example,theoutertipoftheMgixloopsappearedtobeat CME reconnection loops like those seen by EIT (Fig. 9;
around X ¼(cid:1)140000, Y ¼þ20000, which corresponds to also cf. Fig. 2). However, the two loop systems seen by
(cid:4)1.46 R at P:A:¼82(cid:5). The Fe xvi loops appeared to be EIT on January 9 were not clearly distinguishable in the
(cid:2)
alongthesamedirectionbutevenhigherup.Thisisconsis- CDS images that were taken on January 10. In fact, the
tent with the P.A. of the high-temperature emission two loop systems have also become too faint to be clearly
1076 KO ET AL.
Fig.13.—UVCSspectraforzones2,3,and4takenatgratingposition236900,whichcoversthewavelengthrangeof942.2–1042.5A˚ (471–521A˚ insecond
order).Low-ionizationlines([Siviii](cid:3)944and[Fex](cid:3)1028)areparticularlystronginzone2,andhigh-ionizationlines([Fexviii](cid:3)974and[Caxiv](cid:3)943)are
particularlystronginzone3.
distinguishable in EIT on January 10 (see bottom right loops were projected in the plane of the sky. These loops
panel of Fig. 9). In Figure 14, the lower edge of the Mg ix seenbyMK4areprobablyunrelatedtothebrightstreamer
loops (probably associated with the south loop system) is seenbyC2atthenorthofthecurrentsheetbecausethesize
plottedonbothpanels.Itshowsthatthelowertemperature of the loops in MK4 is much less than that of the bright
emission loops (i.e., Mg ix) were located inside the higher streamerinC2.Wecanseethatthehigh-temperatureemis-
temperature loops (Fe xvi). This is consistent with the sionseenbyUVCSandCDSisalongthesamedirectionof
expectation of the flare loop models (e.g., Forbes & Acton thecurrentsheetfeatureinMK4andC2,asexpectedfrom
1996;seex1). the models, although the center of the high-temperature
Figure15showstheMK4imageat19:28UT,January10 emission (seen by UVCS and CDS) is at P:A:(cid:4)81(cid:5) and
(leftpanel),andthecompositeimagefromEIT195A˚ (19:36 that of the current sheet seen by C2 is at P:A:(cid:4)78(cid:5). This
UT), CDS Fe xvi, MK4 (19:28 UT), and C2 (20:26 UT) maybeduetothefactthatthefieldlinesassociatedwiththe
images (right panel; all on January 10). Superposed on the current sheet are nonradial, especially when it was dis-
imageattherightpanelistheprojectionoftheUVCSsliton turbed by the CME earlier that day (see Fig. 8). Note that
the plane of the sky (note that the width is not to scale)
showingthespatialdistributionofthe[Fexviii](cid:3)974emis-
sion. The two ‘‘loops’’ (see the inset in the left panel) and
thecurrentsheetfeature(thinstreamer)abovetheloopsin
theMK4imageseemtobestillpreservedatthistime.How-
ever, it is less certain that the current sheet corresponds to
thesouthloopsystem,probablybecauseofthewaythetwo
Fig.15.—Left:MLSOMK4white-lightimageat19:28UT,January10,
showingthatthetwoloop/cuspsystemsstillexist2daysaftertheCME.
Theinsetshowsourperceptionofwhatthetwoloop/cuspsystemsmight
looklike.Right:CompositeimagefromEIT195A˚ (19:36UT),CDSFexvi
Fig.14.—ImagesofMgix(cid:3)368.07(formationtemperatureof(cid:4)106K, (smallrectangleatthelimb),MK4(19:28UT),andC2(20:26UT)images
withasmallcontributionofMgvii(cid:3)367.68)andFexvi(cid:3)360.8(formation ofJanuary10.SuperposedontheimageistheprojectionoftheUVCSslit
temperatureof(cid:4)2:5(cid:3)106K)composedfromsixrastersoftheCDSobser- ontheplaneofthesky(thewidthisnottoscale)showingthespatialdistri-
vationsonJanuary10.ThelinemarksthelowerboundaryoftheMgix bution of the [Fe xviii] (cid:3)974 emission (cf. Fig. 12k). It is clear that the
loops,whichindicatesthattheMgixloopsareinsidetheFexviloops. high-temperatureemissionisalongthedirectionofthecurrentsheet.
n
o
alibratiaError 22222113311111111111111
C
nt
e
m
sureError(%) <20<10<10<20<10<20<20<30<20<20b<10<10<10<10<10<10<10<10<10<10<10<20<10
a
e
M
1(cid:1))
r
s
1(cid:1)
s
2(cid:1)
Zone4otonscm 19.92634.01580.532.9491.6...3.6...16.413.85.44.875.837.78.61176.1417.1398.34.554.828171.549.0122.9
h
p
70
1
(cid:3)
(
1(cid:1))
r
s
1(cid:1)
s
2(cid:1)
Zone3otonscm 102.09668.56419.6148.92705.745.319.212.027.554.2201.518.8287.7205.627.13086.91598.71171.421.1176.761149.590.1540.8
h
p
70
1
(cid:3)
(
s
e
E1nsiti 1(cid:1)sr)
TABLLineInte Zone221(cid:1)(cid:1)otonscms ...7414.31488.246.4649.6...68.539.976.5...7.19.7121.762.276.83597.91438.11259.0...156.643703.5111.5308.9
h
p
70
1
(cid:3)
(
1(cid:1))
r
s
1(cid:1)
Zone172(cid:1)Identification(10photonscms(cid:3) xv121spPpD...Fe33312x22sSpPMg221988.4==1232xii22sSpPSi22378.7==1232xiii22335pPspSs33322.1Fe]322xii22sSpPSi22116.1==1212xiv3432pSpD...[Ca]22==3232viii3432SpPp212.7[Si]2==3232viii3432pSpP[Si]227.6==3212ix2321pPpS(cid:4)[Si]22,Ly18.210(cid:5)...Lyxviii5252pPpP...[Fe]22==3212xii3432pSpD[Ar]222.4==3252(cid:6)Ly32.0(cid:6)Lycollisionalcomponent10.3x44dDdF[Fe]3314.7==7272vi22sSpPO221041.2==1232viOcollisionalcomponent331.9vi22sSpPO22343.3==1212xii3432SpDp...2[Ar]2==3232x3432pSpD[S]2237.2==3252(cid:1)Ly15785.5v22sSpPN2240.1==1232xiic3432pSpP[Fe]3364.9==3232 <der(20%);2=secondorder(50%);3=blend.(cid:4)only;zones2and4haveuncertaintiesoflessthan30%.v(cid:3)hasbeensubtractedusinghalfoftheN1238flux.
ore342
(cid:3)ID˚(A) 1.45........9.79........9.37........0.05........0.66........3.61........4.38........9.22........0.15........2.54........4.86........18.60......25.72......25.72......28.04......31.91......31.91......37.61......54.90......96.25......15.67......38.82......42.03...... a1=firstbForzonvc(cid:3)N12
80912444577000000001222
46455999999111111111111